This document provides an overview of broadcast fundamentals, covering: 1. The history of television and development of color television standards over time. 2. How the human eye perceives images and color, and concepts like primary/secondary colors, hue, saturation, and luminosity. 3. The basic television signal, including interlaced scanning, synchronization, bandwidth considerations, and the development of monochrome and color NTSC/PAL standards. 4. Components of a video camera like lenses, filters, and early/modern image sensors like CCD, discussing their operation and role in image capture. 5. Magnetic recording principles of video tape recorders and developments in analog/digital recording

1. Broadcast Fundamentals
Sony Training Services
Version 4
14 February 2006

3. Broadcast Fundamentals
Part 1 Table of Contents
Part 1 Table of Contents i
Part 2 The history of television 1
Multimedia timeline 1
Part 3 Image perception & colour 29
The human eye 29
The concept of primary colours 35
Secondary and tertiary colours 36
Hue saturation and luminosity 37
The CIE space 38
Part 4 The basic television signal 40
The problem of getting a picture from A to B 40
Interlaced raster scanning 41
Half lines 42
Synchronisation 44
The oscilloscope 45
Part 5 The monochrome NTSC signal 46
The 405 line system 46
The 525 line monochrome system 46
Frame rate and structure 46
Line rate and structure 47
Bandwidth considerations 47
Part 6 Colour and television 50
Using additive primary colours 50
Ensuring compatability 50
Adding colour 51
Combining R-Y & B-Y 54
Video signal spectra 55
Combining monochrome and colour 56
Using composite video 57
Part 7 Colour NTSC television 58
Similarity to monochrome 58
Choice of subcarrier frequency. 58
Adding colour 59
The vectorscope 61
The gamut 61
The gamut detector 61
Vertical interval structure 62
Part 8 PAL television 64
i Sony Broadcast & Professional Europe

4. Part 1 –Table of contents
What is PAL? 64
The PAL signal 65
The PAL chroma signal 65
Choice of subcarrier frequency 67
Bruch blanking 68
The disadvantages of PAL 71
Part 9 SECAM television 72
The video camera 73
Types of video camera 73
System cameras 73
Parts of a video camera 74
Video camera specifications 76
Lenses 79
Refraction 79
The block of glass 80
The prism 80
The convex lens 81
The concave lens 83
Chromatic aberration 84
Spherical aberration 86
Properties of the lens 86
The concave and convex mirrors 88
Lens types 89
Extenders and adaptors 90
Filters 91
Part 10 Early image sensors 92
Selenium detectors 92
The Ionoscope 92
The Orthicon tube 93
The Image Orthicon tube 93
The Vidicon tube 94
Variations on the Vidicon design 95
Part 11 Dichroic blocks 96
The purpose of a dichroic block 96
Mirrors and filters 96
Optical requirements of a dichroic block 98
Variation on a theme 98
Using dichroic blocks in projectors 98
Part 12 CCD sensors 100
Advantages of CCD image sensors 100
Sony Training Services ii

5. Broadcast Fundamentals
The basics of a CCD 101
Using the CCD as a delay line 102
Using CCD’s as image sensors 106
Back lit sensors 109
Problems with CCD image sensors 110
CCD image sensors with stores 111
HAD technology 115
HyperHAD 117
SuperHAD sensors 117
PowerHAD sensors 117
PowerHAD EX (Eagle) sensors 118
EX View HAD sensors 119
Single chip CCD designs 119
Noise reduction 123
The future of CCD sensors 127
Part 13 The video tape recorder 128
A short history 128
The present day 130
Magnetic recording principles 133
The essentials of helical scan 135
Modern video recorder mechadeck design 140
Variation in tape path designs 147
Definition of a good tape path 148
The servo system 149
Analogue video tape recorder signal processing 150
Popular analogue video recording formats 154
Digital video tape recorders 157
Popular digital video tape formats 159
Part 14 Betacam and varieties 163
Part 15 The video disk recorder 170
History 170
Present day 170
RAID technology 172
Realising RAID systems 179
Part 16 Television receivers & monitors 187
The basic principle 187
Input signals 187
Part 17 Timecode 189
A short history 189
Timecode 190
Timecode’s basic structure 190
iii Sony Broadcast & Professional Europe

6. Part 1 –Table of contents
Longitudinal timecode 194
Bi-phase mark coding 199
Adjusting the LTC head 199
Vertical Interval Timecode 202
Drop frame timecode 207
Which timecode am I using ? 207
Timecode use in video recorders 208
Typical VTR timecode controls 208
The future 210
Part 18 SDI (serial digital interface) 211
Parallel digital television 211
Serial digital television 220
Serial digital audio 221
SDI 226
Video index 226
Part 19 Video compression 227
Traditional analogue signals 227
Analogue to digital conversion 227
Compressing digital signals 227
Digital errors in transmission 228
Compensating for digital errors 228
The advantage of digital compression 228
Entropy and redundancy 228
The purpose of any compression scheme 230
Lossless and lossy compression 230
Inter-frame and Intra-frame 231
What is DCT? 232
The church organ 232
The Fourier transform 233
The Discrete Fourier Transform (DFT) 235
Discrete Cosine Transform (DCT) solution to judder 237
What does the result of DCT look like? 238
DCT in video 238
The mathematics of DCT as used for video 241
DCT in audio 243
Basis pictures 243
Why bother? 243
Huffman’s three step process 244
The principle behind variable length codes 248
The results of discrete cosine transforms 248
Using bell curves for variable length coding 248
Decoding variable length codes 250
Sony Training Services iv

7. Broadcast Fundamentals
Disadvantages of variable length codes 250
The television station 253
The studio 253
The post production studio 253
The edit suite 254
The news studio 254
The outside broadcast vehicle 254
Part 20 CCTV, security & surveillance 256
What is CCTV? 256
CCTV privacy & evidence 256
CCTV use 257
CCTV terminology 259
The typical CCTV chain 263
CCTV cameras 266
Reading CCTV camera specifications 271
CCTV lenses 277
CCTV switchers and control stations 283
CCTV over IP 286
Character and shape recognition 286
Part 21 Numbers & equations 287
Decibels 287
Part 22 Things to do 289
v Sony Broadcast & Professional Europe

9. Broadcast Fundamentals
Part 2 The history of television
Multimedia timeline
Prehistoric
BC
45,000 Neanderthal carvings on Wooly Mammoth tooth, discovered near
Tata, Hungary.
30,000 Ivory horse, oldest known animal carving, from mammoth ivory,
discovered near Vogelherd, Germany.
28,000 Cro-Magnon notation, possibly of phases of the moon, carved
onto bone, discovered at Blanchard, France.
@ 10,000 Engraved antler baton, with seal, salmon and plants
portrayed, discovered at Montgaudier, France.
8,000 - 3100 In Mesopotamia, tokens used for accounting and record-
keeping
3500 In Sumer, pictographs (cuneiforms) of accounts written on clay
tablets.
3400 - 3100 Inscription on Mesopotamian tokens overlap with
pictography
2600 Scribes employed in Egypt.
2400 In India, engraved seals identify the writer.
2200 Date of oldest existing document written on papyrus.
1500 Phoenician alphabet.
1400 Oldest record of writing in China, on bones.
1270 Syrian scholar compiles an encyclopedia.
900 China has an organized postal service for government use.
775 Greeks develop a phonetic alphabet, written from left to right.
530 In Greece, a library.
500 Greek telegraph: trumpets, drums, shouting, beacon fires, smoke
signals, mirrors.
500 Persia has a form of pony express.
500 Chinese scholars write on bamboo with reeds dipped in pigment.
400 Chinese write on silk as well as wood, bamboo.
@ 300 Alexandria library founded by Ptolomy. At its peak the library at
Alexandria had about 700000 manuscripts and books and was a magnet
for scolary from all over the world.
200 Books written on parchment and vellum.
200 Tipao gazettes are circulated to Chinese officials.
1 Sony Broadcast & Professional Europe

10. Part 2 – The history of television
59 Julius Caesar orders postings of Acta Diurna.
48 Alexandria library burnt during Julius Caeser’s siege of
Alexandria.
AD
100 Roman couriers carry government mail across the empire.
105 T'sai Lun invents paper.
175 Chinese classics are carved in stone which will later be used for
rubbings.
180 In China, an elementary zoetrope.
250 Paper use spreads to central Asia.
350 In Egypt, parchment book of Psalms bound in wood covers.
450 Ink on seals is stamped on paper in China. This is true printing.
600 Books printed in China.
700 Sizing agents are used to improve paper quality.
751 Paper manufactured outside of China, in Samarkand by Chinese
captured in war.
765 Picture books printed in Japan.
868 The Diamond Sutra, a block-printed book in China.
875 Amazed travelers to China see toilet paper.
950 Paper use spreads west to Spain.
950 Folded books appear in China in place of rolls.
950 Bored women in a Chinese harem invent playing cards.
1000-1499
1000 Mayas in Yucatan, Mexico, make writing paper from tree bark.
1035 Japanese use waste paper to make new paper.
1049 Pi Sheng fabricates movable type, using clay.
1116 Chinese sew pages to make stitched books.
1140 In Egypt, cloth is stripped from mummies to make paper.
1147 Crusader taken prisoner returns with papermaking art, according
to a legend.
1200 European monasteries communicate by letter system.
1200 University of Paris starts messenger service.
1241 In Korea, metal type.
1282 In Italy, watermarks are added to paper.
1298 Marco Polo describes use of paper money in China.
1300 Wooden type found in central Asia.
Sony Training Services 2

11. Broadcast Fundamentals
1305 Taxis family begins private postal service in Europe.
1309 Paper is used in England.
1392 Koreans have a type foundry to produce bronze characters.
1423 Europeans begin Chinese method of block printing.
1450 A few newsletters begin circulating in Europe.
1451 Johnannes Gutenberg uses a press to print an old German
poem.
1452 Metal plates are used in printing.
1453 Gutenberg prints the 42-line Bible.
1464 King of France establishes postal system.
1490 Printing of books on paper becomes more common in Europe.
1495 A paper mill is established in England.
1500 – 1599
1500 Arithmetic + and - symbols are used in Europe.
1510 By now approximately 35,000 books have been printed, some 10
million copies.
1520 Spectacles balance on the noses of Europe's educated.
1533 A postmaster in England.
1545 Garamond designs his typeface.
1550 Wallpaper brought to Europe from China by traders.
1560 In Italy, the portable camera obscura allows precise tracing of an
image.
1560 Legalized, regulated private postal systems grow in Europe.
1556 The pencil.
1600 – 1699
1609 First regularly published newspaper appears in Germany.
1627 France introduces registered mail.
1631 A French newspaper carries classified ads.
1639 In Boston, someone is appointed to deal with foreign mail.
1639 First printing press in the American colonies.
1640 Kirchner, a German Jesuit, builds a magic lantern.
1650 Leipzig has a daily newspaper.
1653 Parisians can put their postage-paid letters in mail boxes.
1659 Londoners get the penny post.
1661 Postal service within the colony of Virginia.
1673 Mail is delivered on a route between New York and Boston.
3 Sony Broadcast & Professional Europe

12. Part 2 – The history of television
1689 Newspapers are printed, at first as unfolded "broadsides."
1696 By now England has 100 paper mills.
1698 Public library opens in Charleston, S.C.
1700 - 1799
1704 A newspaper in Boston prints advertising.
1710 German engraver Le Blon develops three-color printing.
1714 Henry Mill receives patent in England for a typewriter.
1719 Reaumur proposes using wood to make paper.
1725 Scottish printer develops stereotyping system.
1727 Schulze begins science of photochemistry.
1732 In Philadelphia, Ben Franklin starts a circulating library.
1755 Regular mail ship runs between England and the colonies.
1770 The eraser.
1774 Swedish chemist invents a future paper whitener.
1775 Continental Congress authorizes Post Office; Ben Franklin first
Postmaster General.
1780 Steel pen points begin to replace quill feathers.
1784 French book is made without rags, from vegetation.
1785 Stagecoaches carry the mail between towns in U.S.
1790 In England the hydraulic press is invented.
1792 Mechanical semaphore signaler built in France.
1792 In Britain, postal money orders.
1792 Postal Act gives mail regularity throughout U.S.
1794 First letter carriers appear on American city streets.
1794 Panorama, forerunner of movie theaters, opens.
1794 Signaling system connects Paris and Lille.
1798 Senefelder in Germany invents lithography.
1799 Robert in France invents a paper-making machine.
1800 - 1899
1800 Paper can be made from vegetable fibers instead of rags.
1800 Letter takes 20 days to reach Savannah from Portland, Maine.
1801 Semaphore system built along the coast of France.
1801 Joseph-Marie Jacquard invents a loom using punch cards.
1803 Fourdrinier continuous web paper-making machine.
1804 In Germany, lithography is invented.
1806 Carbon paper.
Sony Training Services 4

13. Broadcast Fundamentals
1807 Camera lucida improves image tracing.
1808 Turri of Italy builds a typewriter for a blind contessa.
1817 Jons Berzelius discovered selenium, an element shown in later
years to have photo-voltaic effects. The material was a bi-products of
chemical processes carried out in a Swedish factory. At first he though
the material was tellurium “earth”, but later found it to be a new element
and named it selenium from the Greek word “selene” meaning “moon”.
1831 Michael Faraday in Britain and Joseph Henry in the United
States experiment with electromagnetism, providing the basis for
research into electrical communication.
1844 Samuel Morse publicly demonstrates the telegraph for the first
time.
1862 Italian physicist, Abbe Giovanni Caselli, is the first to send fixed
images over a long distance, using a system he calls the "pantelegraph".
1873 Two English telegraph engineers, May and Smith, experiment
with selenium and light, giving inventors a way of transforming images
into electrical signals.
1880 George Carey builds a rudimentary system using dozens of tiny
light-sensitive selenium cells.
1884 In Germany, Paul Nipkow patents the first mechanical television
scanning system, consisting of a disc with a spiral of holes. As the disc
spins, the eye blurs all the points together to re-create the full picture.
1895 Italian physicist Guglielmo Marconi develops radio telegraphy
and transmits Morse code by wireless for the first time.
1897 Karl Ferdinand Braun, a German physicist, invents the first
cathode-ray tube, the basis of all modern television cameras and
receivers.
1900 – 1909
1900
Kodak Brownie makes photography cheaper and simpler.
Pupin's loading coil reduces telephone voice distortion.
1901
Sale of phonograph disc made of hard resinous shellac
First electric typewriter, the Blickensderfer.
Marconi sends a radio signal across the Atlantic.
1902
Germany's Zeiss invents the four-element Tessar camera lens.
Etched zinc engravings start to replace hand-cut wood blocks.
5 Sony Broadcast & Professional Europe

14. Part 2 – The history of television
U.S. Navy installs radio telephones aboard ships.
Photoelectric scanning can send and receive a picture.
Trans-Pacific telephone cable connects Canada and Australia.
1903
Technical improvements in radio, telegraph, phonograph, movies and
printing.
London Daily Mirror illustrates only with photographs.
A telephone answering machine is invented.
Fleming invents the diode to improve radio communication.
Offset lithography becomes a commercial reality.
A photograph is transmitted by wire in Germany.
Hine photographs America's underclass.
The Great Train Robbery creates demand for fiction movies.
The comic book.
The double-sided phonograph disc.
1905
In Pittsburgh the first nickelodeon opens.
Photography, printing, and post combine in the year's craze, picture
postcards.
In France, Pathe colors black and white films by machine.
In New Zealand, the postage meter is introduced.
The Yellow Pages.
The juke box; 24 choices.
1906
A program of voice and music is broadcast in the U.S.
Lee de Forest invents the three-element vacuum tube.
Dunwoody and Pickard build a crystal-and-cat's-whisker radio.
An animated cartoon film is produced.
Fessenden plays violin for startled ship wireless operators.
An experimental sound-on-film motion picture.
Strowger invents automatic dial telephone switching.
1907
Bell and Howell develop a film projection system.
Lumiere brothers invent still color photography process.
Sony Training Services 6

15. Broadcast Fundamentals
DeForest begins regular radio music broadcasts.
In Russia, Boris Rosing develops theory of television and transmits
black-and-white silhouettes of simple shapes, using a mechanical mirror-
drum apparatus as a camera and a cathode-ray tube as a receiver.
1908
Campbell-Swinton, a Scottish electrical engineer, publishes proposals
about an all-electronic television system that uses a cathode-ray tube for
both receiver and camera.
In U.S., Smith introduces true color motion pictures.
1909
Radio distress signal saves 1,700 lives after ships collide.
First broadcast talk; the subject: women's suffrage.
1910-1919
1910
Sweden's Elkstrom invents "flying spot" camera light beam.
1911
Efforts are made to bring sound to motion pictures.
Rotogravure aids magazine production of photos.
"Postal savings system" inaugurated.
1912
U.S. passes law to control radio stations.
Motorized movie cameras replace hand cranks.
Feedback and heterodyne systems usher in modern radio.
First mail carried by airplane.
1913
The portable phonograph is manufactured.
Type composing machines roll out of the factory.
1914
A better triode vacuum tube improves radio reception.
Radio message is sent to an airplane.
In Germany, the 35mm still camera, a Leica.
In the U.S., Goddard begins rocket experiments.
7 Sony Broadcast & Professional Europe

16. Part 2 – The history of television
First transcontinental telephone call.
1915
Wireless radio service connects U.S. and Japan.
Radio-telephone carries speech across the Atlantic.
Birth of a Nation sets new movie standards.
The electric loudspeaker.
1916
David Sarnoff envisions radio as "a household utility."
Cameras get optical rangefinders.
Radios get tuners.
1917
Photocomposition begins.
Frank Conrad builds a radio station, later KDKA.
Condenser microphone aids broadcasting, recording.
1918
First regular airmail service: Washington, D.C. to New York.
1919
The Radio Corporation of America (RCA) is formed.
People can now dial telephone numbers themselves.
Shortwave radio is invented.
Flip-flop circuit invented; will help computers to count.
1920-1929
1920
The first broadcasting stations are opened.
First cross-country airmail flight in the U.S.
Sound recording is done electrically.
Post Office accepts the postage meter.
KDKA in Pittsburgh broadcasts first scheduled programs.
1921
Quartz crystals keep radio signals from wandering.
The word "robot" enters the language.
Western Union begins wirephoto service.
Sony Training Services 8

17. Broadcast Fundamentals
1922
A commercial is broadcast, $100 for ten minutes.
Technicolor introduces two-color process for movies.
Germany's UFA produces a film with an optical sound track.
First 3-D movie, requires spectacles with one red and one green lens.
Singers desert phonograph horn mouths for acoustic studios.
Nanook of the North, the first documentary.
1923
Vladimir Zworykin patents the "Iconoscope", an electronic camera tube.
By the end of the year he has also produced a picture display tube, the
"Kinescope".
People on one ship can talk to people on another.
Ribbon microphone becomes the studio standard.
A picture, broken into dots, is sent by wire.
16 mm nonflammable film makes its debut.
Kodak introduces home movie equipment.
Neon advertising signs.
The A.C. Nielsen Company is founded. Nielsen's market research is
soon being used by companies deciding where to advertise on radio.
1924
John Logie Baird is the first to transmit a moving silhouette image, using
a mechanical system based on Paul Nipkow's model.
Low tech achievement: notebooks get spiral bindings.
The Eveready Hour is the first sponsored radio program.
At KDKA, Conrad sets up a short-wave radio transmitter.
Daily coast-to-coast air mail service.
Two and a half million radio sets in the U.S.
1925
John Logie Baird obtains the first actual television picture.
Vladimir Zworykin takes out the first patent for colour television.
The Leica 35 mm camera sets a new standard.
Commercial picture facsimile radio service across the U.S.
All-electric phonograph is built.
A moving image, the blades of a model windmill, is telecast.
From France, a wide-screen film.
9 Sony Broadcast & Professional Europe

18. Part 2 – The history of television
1926
John Logie Baird gives the first successful public demonstration of
mechanical television at his laboratory in London.
The National Broadcasting Company (NBC) is formed by Westinghouse,
General Electric and RCA.
Commercial picture facsimile radio service across the Atlantic.
Some radios get automatic volume control, a mixed blessing.
The Book-of-the-Month Club.
In U.S., first 16mm movie is shot.
Goddard launches liquid-fuel rocket.
Permanent radio network, NBC, is formed.
Bell Telephone Labs transmit film by television.
1927
The British Broadcasting Corporation is founded.
Columbia Phonographic Broadcasting System, later CBS, is formed
Pictures of Herbert Hoover, U.S. Secretary of Commerce, are
transmitted 200 miles from Washington D.C. to New York, in the world's
first televised speech and first long-distance television transmission.
NBC begins two radio networks.
Farnsworth assembles a complete electronic TV system.
Jolson's "The Jazz Singer" is the first popular "talkie."
Movietone offers newsreels in sound.
U.S. Radio Act declares public ownership of the airwaves.
Technicolor.
Negative feedback makes hi-fi possible.
1928
Station W2XBS, RCA's first television station, is established in New York
City, creating television's first star, Felix the Cat — the original model of
which is featured in Watching TV
Later in the year, the world's first television drama, The Queen's
Messenger, is broadcast, using mechanical scanning
John Logie Baird transmits images of London to New York via
shortwave.
The teletype machine makes its debut.
Television sets are put in three homes, programming begins.
Baird invents a video disc to record television.
In an experiment, television crosses the Atlantic.
Sony Training Services 10

19. Broadcast Fundamentals
In Schenectady, N.Y., the first scheduled television broadcasts.
Steamboat Willie introduces Mickey Mouse.
A motion picture is shown in color.
Times Square gets moving headlines in electric lights.
IBM adopts the 80-column punched card.
1929
In London, John Logie Baird opens the world's first television studio, but
is still able to produce only crude and jerky images. However, because
Baird's television pictures carry so little visual information, it is possible
to broadcast them from ordinary medium-wave radio transmitters.
Experiments begin on electronic color television.
Telegraph ticker sends 500 characters per minute.
Ship passengers can phone relatives ashore.
Brokers watch stock prices on an automated electric board.
Something else new: the car radio.
In Germany, magnetic sound recording on plastic tape.
Air mail flown from Miami to South America.
Bell Lab transmits stills in color by mechanical scanning.
Zworykin demonstrates cathode-ray tube "kinescope" receiver, 60 scan
lines.
1930-1939
1930
The first commercial is televised by Charles Jenkins, who is fined by the
U.S. Federal Radio Commission.
The BBC begins regular television transmissions.
Photo flashbulbs replace dangerous flash powder.
"Golden Age" of radio begins in U.S.
Lowell Thomas begins first regular network newscast.
TVs based on British mechanical system roll off factory line.
Bush's differential analyzer introduces the computer.
AT&T tries the picture telephone.
1931
Owned jointly by CKAC and La Presse, Canada's first television station,
VE9EC, starts broadcasting in Montreal. Ted Rogers, Sr. receives a
licence to broadcast experimental television from his Toronto radio
station. Also this year, RCA begins experimental electronic
transmissions from the Empire State Building.
11 Sony Broadcast & Professional Europe

20. Part 2 – The history of television
Commercial teletype service.
Electronic TV broadcasts in Los Angeles and Moscow.
Exposures meters go on sale to photographers.
NBC experimentally doubles transmission to 120-line screen.
1932
Parliament creates the Canadian Radio Broadcasting Commission,
superseded by the CBC in 1936.
Disney adopts a three-color Technicolor process for cartoons.
Kodak introduces 8 mm film for home movies.
The "Times" of London uses its new Times Roman typeface.
Stereophonic sound in a motion picture, "Napoleon."
Zoom lens is invented, but a practical model is 21 years off.
The light meter.
NBC and CBS allow prices to be mentioned in commercials.
1933
Western Television Limited's mechanical television system is toured and
demonstrated at Eaton's stores in Toronto, Montreal and Winnipeg.
Armstrong invents FM, but its real future is 20 years off.
Multiple-flash sports photography.
Singing telegrams.
Phonograph records go stereo.
1934
Drive-in movie theater opens in New Jersey.
Associated Press starts wirephoto service.
In Germany, a mobile television truck roams the streets.
In Scotland, teletypesetting sets type by phone line.
Three-color Technicolor used in live action film.
Communications Act of 1934 creates FCC.
Half of the homes in the U.S. have radios.
Mutual Radio Network begins operations.
1935
William Hoyt Peck of Peck Television of Canada uses a transmitter in
Montreal during five weeks of experimental mechanical broadcasts.
Germany opens the world's first three-day-a-week filmed television
service. France begins broadcasting its first regular transmissions from
the top of the Eiffel Tower.
Sony Training Services 12

21. Broadcast Fundamentals
German single lens reflex roll film camera synchronized for flash bulbs.
Also in Germany, audio tape recorders go on sale.
IBM's electric typewriter comes off the assembly line.
The Penguin paperback book sells for the price of 10 cigarettes.
All-electronic VHF television comes out of the lab.
Eastman-Kodak develops Kodachrome color film.
Nielsen's Audimeter tracks radio audiences.
1936
There are about 2,000 television sets in use around the world. The BBC
starts the world's first public high-definition/electronic television service
in London.
Berlin Olympics are televised closed circuit.
Bell Labs invents a voice recognition machine.
Kodachrome film sharpens color photography.
Co-axial cable connects New York to Philadelphia.
Alan Turing's "On Computable Numbers" describes a general purpose
computer.
1937
Stibitz of Bell Labs invents the electrical digital calculator.
Pulse Code Modulation points the way to digital transmission.
NBC sends mobile TV truck onto New York streets.
A recording, the Hindenburg crash, is broadcast coast to coast.
Carlson invents the photocopier.
Snow White is the first feature-length cartoon.
1938
Allen B. DuMont forms the DuMont television network to compete with
RCA. Also this year, DuMont manufactures the first all-electronic
television set for sale to the North American public. One of these early
DuMont television sets is featured in Watching TV.
Strobe lighting.
Baird demonstrates live TV in color.
Broadcasts can be taped and edited.
Two brothers named Biro invent the ballpoint pen in Argentina.
CBS "World News Roundup" ushers in modern newscasting.
DuMont markets electronic television receiver for the home.
Radio drama, War of the Worlds," causes national panic.
13 Sony Broadcast & Professional Europe

22. Part 2 – The history of television
1939
Because of the outbreak of war, the BBC abruptly stops broadcasting in
the middle of a Mickey Mouse cartoon on September 1, resuming at that
same point when peace returns in 1945. The first major display of
electronic television in Canada takes place at the Canadian National
Exhibition in Toronto. Baseball is televised for the first time.
Mechanical scanning system abandoned.
New York World's Fair shows television to public.
Regular TV broadcasts begin in USA.
Air mail service across the Atlantic.
Many firsts: sports coverage, variety show, feature film, etc.
1940-1949
1940
Dr. Peter Goldmark of CBS introduces a 343-line colour television
system for daily transmission, using a disc of three filters (red, green and
blue), rotated in front of the camera tube.
Fantasia introduces stereo sound to American public.
1941
North America's current 525-line/30-pictures-a-second standard, known
as the NTSC (National Television Standards Committee) standard, is
adopted.
Stereo is installed in a Moscow movie theater.
FCC sets U.S. TV standards.
CBS and NBC start commercial transmission; WW II intervenes.
Goldmark at CBS experiments with electronic color TV.
Microwave transmission.
Zuse's Z3 is the first computer controlled by software.
1942
Atanasoff, Berry build the first electronic digital computer.
Kodacolor process produces the color print.
1943
Repeaters on phone lines quiet long distance call noise.
1944
Harvard's Mark I, first digital computer, put in service.
IBM offers a typewriter with proportional spacing.
NBC presents first U.S. network newscast, a curiosity.
Sony Training Services 14

23. Broadcast Fundamentals
1945
BBC returns regular transmission of television, at the exact same time of
day at exactly the same point in the programme.
Clarke envisions geo-synchronous communication satellites.
It is estimated that 14,000 products are made from paper.
1946
NBC and CBS demonstrate rival colour systems. The world's first
television broadcast via coaxial cable is transmitted from New York to
Washington D.C.
Jukeboxes go into mass production.
Pennsylvania's ENIAC heralds the modern electronic computer.
Automobile radio telephones connect to telephone network.
French engineers build a phototypesetting machine.
1947
A permanent network linking four eastern U.S. stations is established by
NBC. On June 3, Canadian General Electric engineers in Windsor
receive the first official electronic television broadcast in Canada,
transmitted from the new U.S. station WWDT in Detroit. This year also
sees the development of the transistor, on which solid-state electronics
are based.
Hungarian engineer in England invents holography.
The transistor is invented, will replace vacuum tubes.
The zoom lens covers baseball's world series for TV.
1948
Television manufacturing begins in Canada. The television audience
increases by 4,000 percent this year, due to a jump in the number of
cities with television stations and to the fact that one million homes in the
U.S. now have television sets. The U.S. Federal Communications
Commission puts a freeze on new television channel allocations until the
problem of station-to-station interference is resolved.
The LP record arrives on a viny disk.
Shannon and Weaver of Bell Labs propound information theory.
Land's Polaroid camera prints pictures in a minute.
Hollywood switches to nonflammable film.
Public clamor for television begins; FCC freezes new licenses.
Airplane re-broadcasts TV signal across nine states.
15 Sony Broadcast & Professional Europe

24. Part 2 – The history of television
1949
The first Emmy Awards are presented, and the Canadian government
establishes an interim policy for television, announcing loans for CBC
television development.
An RCA research team in the U.S. develops the Shadow Mask picture
tube, permitting a fully electronic colour display.
Network TV in U.S.
RCA offers the 45 rpm record.
Community Antenna Television, forerunner to cable.
Whirlwind at MIT is the first real time computer.
Magnetic core computer memory is invented.
1950-1959
1950
Cable TV begins in the U.S., and warnings begin to be issued on the
impact of violent programming on children.
European broadcasters fix a common picture standard of 625 lines. (By
the 1970s, virtually all nations in the world used 625-line service, except
for the U.S., Japan, and some others which adopted the 525-line U.S.
standard.)
Over 100 television stations are in operation in the U.S.
Regular USA color television transmission.
Vidicon camera tube improves television picture.
Changeable typewriter typefaces in use.
A.C. Nielsen's Audimeters track viewer watching.
1951
The first colour television transmissions begin in the U.S. this year.
Unfortunately, for technical reasons, the several million existing black-
and-white receivers in America cannot pick up the colour programmes,
even in black-and-white, and colour sets go blank during television's
many hours of black-and-white broadcasting. The experiment is a failure
and colour transmissions are stopped.
The U.S. sees its first coast-to-coast transmission in a broadcast of the
Japanese Peace Conference in San Francisco.
One and a half million TV sets in U.S., a tenfold jump in one year.
Cinerama will briefly dazzle with a wide, curved screen and three
projectors.
Computers are sold commercially.
Still camera get built-in flash units.
Coaxial cable reaches coast to coast.
Sony Training Services 16

25. Broadcast Fundamentals
1952
Cable TV systems begin in Canada. On September 6, CBC Television
broadcasts from its Montreal station; on September 8, CBC broadcasts
from the Toronto station.
The first political ads appear on U.S. television networks, when
Democrats buy a half-hour slot for Adlai Stevenson. Stevenson is
bombarded with hate mail for interfering with a broadcast of I Love Lucy.
Eisenhower, Stevenson's political opponent, buys only 20-second
commercial spots, and wins the election.
3-D movies offer thrills to the audience.
Bing Crosby's company, Crosby Enterprises, tests video recording.
Wide-screen Cinerama appears; other systems soon follow.
Sony offers a miniature transistor radio.
EDVAC takes computer technology a giant leap forward.
Univac projects the winner of the presidential election on CBS.
Telephone area codes.
Zenith proposes pay-TV system using punched cards.
Sony offers a miniature transistor radio.
1953
A microwave network connects CBC television stations in Montreal,
Ottawa and Toronto.
The first private television stations begin operation in Sudbury and
London.
Queen Elizabeth's coronation is televised.
CBC beats U.S. competitors to the punch by flying footage across the
Atlantic.
In the USA TV Guide is launched.
NTSC colour standard adopted and the USA begins colour transmission
again, this time successfully.
Japanese television goes on the air for the first time.
CATV system uses microwave to bring in distant signals.
1954
Magazines now routinely offer the homemaker tips on arranging living-
room furniture for optimal television-viewing pleasure.
U.S.S.R. launches Sputnik.
Radio sets in the world now outnumber newspapers printed daily.
Regular colour TV broadcasts established.
Sporting events are broadcast live in colour.
17 Sony Broadcast & Professional Europe

26. Part 2 – The history of television
Radio sets in the world now outnumber daily newspapers.
Transistor radios are sold.
1955
Tests begin to communicate via fiber optics.
Music is recorded on tape in stereo.
1956
Ampex Corporation demonstrates videotape recording, initially used only
by television stations.
Henri de France develops the SECAM (sequential colour with memory)
procedure. It is adopted in France, and the first SECAM colour
transmission between Paris and London takes place in 1960.
Several Louisiana congressmen promote a bill to ban all television
programmes that portray blacks and whites together in a sympathetic
light.
Bell tests the picture phone.
First transatlantic telephone calls by cable.
1957
The Soviet Union launches the world's first Earth satellite, Sputnik.
Soviet Union's Sputnik sends signals from space.
FORTRAN becomes the first high-level language.
A surgical operation is televised.
First book to be entirely phototypeset is offset printed.
1958
The CBC's microwave network is extended from Victoria, B.C. to Halifax
and Sydney, Nova Scotia, to become the longest television network in
the world.
Pope Pius XII declares Saint Clare of Assisi the patron saint of
television. Her placement on the television set is said to guarantee good
reception.
Videotape delivers colour pictures.
Stereo recording is introduced.
Data moves over regular phone circuits.
Broadcast bounced off rocket, pre-satellite communication.
The laser is introduced.
Cable carries FM radio stations.
Sony Training Services 18

27. Broadcast Fundamentals
1959
CBC Radio-Canada Montreal producers go on strike.
Bonanza debuts, starring Canadian actor Lorne Greene.
Local announcements, weather data and local ads go on cable.
The microchip is invented.
Xerox manufactures a plain paper copier.
Bell Labs experiments with artificial intelligence.
French SECAM and German PAL systems introduced.
1960-1969
1960
The Nixon-Kennedy debates are televised, marking the first network use
of the split screen. Kennedy performs better on television than Nixon,
and it is believed that television helps Kennedy win the election.
Sony develops the first all-transistor television receiver, making
televisions lighter and more portable.
Ninety percent of American homes now own television sets, and
America becomes the world's first "television society". There are now
about 100 million television sets in operation worldwide.
Echo I, a U.S. balloon in orbit, reflects radio signals to Earth.
In Rhode Island, an electronic, automated post office.
A movie gets Smell-O-Vision, but the public just sniffs.
Zenith tests subscription TV; unsuccessful.
1961
The Canadian Television Network (CTV), a privately owned network,
begins operations.
The beginning of the Dodd hearings in the U.S., which examined the
television industry's "rampant and opportunistic use of violence".
Boxing match test shows potential of pay-TV.
FCC approves FM stereo broadcasting; spurs FM development.
Bell Labs tests communication by light waves.
IBM introduces the "golf ball" typewriter.
Letraset makes headlines simple.
The time-sharing computer is developed.
1962
The Telstar television satellite is launched by the U.S., and starts
relaying transatlantic television shortly after its launch. The first
programme shows scenes of Paris.
19 Sony Broadcast & Professional Europe

28. Part 2 – The history of television
A survey indicates that 90 percent of American households have
television sets; 13 percent have more than one.
Cable companies import distant signals.
FCC requires UHF tuners on television sets.
The minicomputer arrives.
Comsat created to launch, operate global system.
1963
From Holland comes the audio cassette.
Zip codes introduced.
CBS and NBC TV newscasts expand to 30 minutes in color.
PDP-8 becomes the first popular minicomputer.
Polaroid camera instant photography adds color.
Communications satellite is placed in geo-synchronous orbit.
On November 22, regular television programming is suspended
following news of the Kennedy assassination.
On November 24, live on television, Jack Ruby murders Lee Harvey
Oswald, Kennedy's suspected assassin. Kennedy's funeral is televised
the following day. 96 per cent of all American television sets are on for
an average 31 hours out of 72 during this period — watching, many say,
simply to share in the crisis.
1964
The Beatles appear for the first time on Ed Sullivan Show.
Procter and Gamble, the largest American advertiser, refuses to
advertise on any show that gives "offense, either directly or by inference,
to any organized minority group, lodge or other organizations,
institutions, residents of any State or section of the country or a
commercial organization."
Olympic Games in Tokyo telecast live globally by satellite.
Touch Tone telephones and Picturephone service.
From Japan, the videotape recorder for home use.
Russian scientists bounce a signal off Jupiter.
Intelsat, international satellite organization, is formed.
1965
The Vietnam War becomes the first war to be televised, coinciding with
CBS's first colour transmissions and the first Asia-America satellite link.
Protesters against the war adopt the television-age slogan, The whole
world is watching.
Sony Training Services 20

29. Broadcast Fundamentals
Sony introduces Betamax, a small home videorecorder.
Electronic phone exchange gives customers extra services.
Satellites begin domestic TV distribution in Soviet Union.
Computer time-sharing becomes popular.
Color news film.
Communications satellite Early Bird (Intelsat I) orbits above the Atlantic.
Kodak offers Super 8 film for home movies.
Cartridge audio tapes go on sale for a few years.
Most television broadcasts in the USA are in colour.
FCC rules bring structure to cable television.
Solid-state equipment spreads through the cable industry.
1966
Colour television signals are transmitted by Canadian stations for the
first time.
Linotron can produce 1,000 characters per second.
Fiber optic cable multiplies communication channels.
Xerox sells the Telecopier, a fax machine.
1967
Sony introduces the first lightweight, portable and cheap video recorder,
known as the "portapak". The portapak is almost as easy to operate as a
tape-recorder and leads to an explosion in "do-it-yourself" television,
revolutionizing the medium.
Also this year, the FCC orders that cigarette ads on television, on radio
and in print, carry warnings about the health dangers of smoking.
Dolby introduces a system that eliminates audio hiss.
Computers get the light pen.
Pre-recorded movies on videotape sold for home TV sets.
Cordless telephones get some calls.
Approx. 200 million telephones in the world, half in U.S.
1968
Sony develops the Trinitron tube, revolutionizing the picture quality of
colour television.
World television ownership nears 200 million, with 78 million sets in the
U.S. alone. The U.S. television industry now has annual revenues of
about $2 billion and derives heavy support from tobacco advertisers.
FCC approves non-Bell equipment attached to phone system.
The RAM microchip reaches the market.
21 Sony Broadcast & Professional Europe

30. Part 2 – The history of television
1969
On July 20, 1969, the first television transmission from the moon is
viewed by 600 million television viewers around the world.
Sesame Street debuts on American Public Television, and begins to
revolutionize adult attitudes about what children are capable of learning.
Astronauts send live photographs from the moon.
1970-1979
1970
Postal Reform Bill makes U.S. Postal Service a government corporation.
In Germany, a videodisc is demonstrated.
U.S. Post Office and Western Union offer Mailgrams.
The computer floppy disc is an instant success.
1971
Canada's Anik I, the first domestic geo-synchronous communications
satellite, is launched, capable of relaying 12 television programmes
simultaneously.
India has a single television station in New Delhi, able to reach only 20
miles outside the city.
South Africa has no television at all.
Intel builds the microprocessor, "a computer on a chip."
Wang 1200 is the first word processor.
1972
The Munich Olympics are broadcast live, drawing an estimated 450
million viewers worldwide. When Israeli athletes are kidnapped by
Palestinian terrorists during the games, coverage of the games cuts
back and forth between shots of the terrorists and footage of Olympic
events.
The American-conceived Intelsat system is launched this year,
becoming a network and controlling body for the world's communications
satellite system.
HBO starts pay-TV service for cable.
Sony introduces 3/4 inch "U-Matic" cassette VCR.
New FCC rules lead to community access channels.
Polaroid camera can focus by itself.
Digital television comes out of the lab.
The BBC offers "Ceefax," two-way cable information system.
"Open Skies": Any U.S. firm can have communication satellites.
Landsat I, "eye-in-the-sky" satellite, is launched.
Sony Training Services 22

31. Broadcast Fundamentals
"Pong" starts the video game craze.
1973
Ninety-six countries now have regular television service.
Watergate unfolds on the air in the U.S. and ends the following year with
Nixon's resignation.
U.S. producers sell nearly $200 million dollars worth of programmes
overseas, more than the rest of the world combined.
The microcomputer is born in France.
IBM's Selectric typewriter is now "self-correcting."
The term Electronic News Gathering, or ENG is introduced.
"Teacher-in-the-Sky" satellite begins educational mission.
1975
A study indicates that the average American child during this decade will
have spent 10,800 hours in school by the time he or she is 18, but will
have seen an average 20,000 hours of television. Studies also estimate
that, by the time he/she is 75, the average American male will have
spent nine entire years of his life watching television; the average British
male will have spent eight years watching.
The microcomputer, in kit form, reaches the U.S. home market.
Sony's Betamax and JVC's VHS battle for public acceptance.
"Thrilla' from Manila"; substantial original cable programming.
1976
The Olympics, broadcast from Montreal, draw an estimated 1 billion
viewers worldwide.
Apple I compter introduced.
Ted Turner delivers programming nationwide by satellite.
Still cameras are controlled by microprocessors.
1977
South Africans see television for the first time on May 10, as test
transmissions begin from the state-backed South Africa Broadcast Co.
The Pretoria government has yielded to public pressure after years of
banning television as being morally corrupting. Half the broadcasts are
in English, half in Afrikaans.
Columbus, Ohio, residents try 2-way cable experiment, QUBE.
1978
Ninety-eight percent of American households have television sets, up
from nine percent in 1950. Seventy-eight percent have colour
televisions, up from 3.1 percent in 1964.
23 Sony Broadcast & Professional Europe

32. Part 2 – The history of television
From Konica, the point-and-shoot camera.
PBS goes to satellite for delivery, abandoning telephone lines.
Electronic typewriters go on sale.
1979
There are now 300 million television sets in operation worldwide.
Flat-screen pocket televisions, with liquid crystal display screens, are
patented by the Japanese firm Matsushita. The pocket television is no
bigger than a paperback book.
Speech recognition machine has a vocabulary of 1,000 words.
From Holland comes the digital videodisc read by laser.
In Japan, first cellular phone network.
Computerized laser printing is a boon to Chinese printers.
1980-1989
1980
During the 1980s, in the U.S. and Germany, laws and policies are
enacted to preserve a person's right to television in the event of financial
setback. Later in the year, the U.S. Cable News Network (CNN) goes on
the air in the U.S.
India launches its national television network.
Sony Walkman tape player starts a fad.
In France, a holographic film shows a gull flying.
Phototypesetting can be done by laser.
Intelsat V relays 12,000 phone calls, 2 color TV channels.
Public international electronic fax service, Intelpost, begins.
Atlanta gets first fiber optics system.
CNN 24-hour news channel started.
Addressable converters pinpoint individual homes.
1981
450,000 transistors fit on a silicon chip 1/4-inch square.
Hologram technology improves, now in video games.
The IBM PC.
The laptop computer is introduced.
The first mouse pointing device.
1982
From Japan, a camera with electronic picture storage, no film.
Sony Training Services 24

33. Broadcast Fundamentals
USA Today type set in regional plants by satellite command.
Kodak camera uses film on a disc cassette.
1983
Cellular phone network starts in U.S.
Lasers and plastics improve newspaper production.
Computer chip holds 288,000 bits of memory.
Time names the computer as "Man of the Year."
ZIP + 4, expanded 9-digit ZIP code is introduced.
AT&T forced to break up; 7 Baby Bells are born.
American videotext service starts; fails in three years.
1984
Trucks used for SNG transmission.
Experimental machine can translate Japanese into English.
Portable compact disc player arrives.
National Geographic puts a hologram on its cover.
A television set can be worn on the wrist.
Japanese introduce high quality facsmile.
Camera and tape deck combine in the camcorder.
Apple Macintosh, IBM PC AT.
The 32-bit microprocessor.
The one megabyte memory chip.
Conus relays news feeds for stations on Ku-Band satellites.
1985
Digital image processing for editing stills bit by bit.
CD-ROM can put 270,000 papers of text on a CD record.
Cellular telephones go into cars.
Synthetic text-to-speech computer pronounces 20,000 words.
Picture, broken into dots, can be transmitted and recreated.
USA TV networks begin satellite distribution to affiliates.
At Expo, a Sony TV screen measures 40x25 meters.
Sony builds a radio the size of a credit card.
In Japan, 3-D television; no spectacles needed.
Pay-per-view channels open for business.
25 Sony Broadcast & Professional Europe

34. Part 2 – The history of television
1986
HBO scrambles its signals.
Cable shopping networks.
1987
Half of all U.S. homes with TV are on cable.
American government deregulates cable industry.
1988
Government brochure mailed to 107 million addresses.
1989
Tiananmen Square demonstrates power of media to inform the world.
Pacific Link fiber optic cable opens, can carry 40,000 phone calls.
1990- 2000
1990
Flyaway SNG aids foreign reportage.
IBM sells Selectric, a sign of the typewriter's passing.
Most 2-inch videotape machines are also gone.
Videodisc returns in a new laser form.
1991
During the Gulf War, CNN coverage of the conflict is so extensive and
wide-ranging that it is commonly remarked, only half in jest, that Saddam
Hussein is watching CNN for his military intelligence, instead of relying
on his own information-gathering methods.
Beauty and the Beast, a cartoon, Oscar nominee as best picture.
Denver viewers can order movies at home from list of more than 1,000
titles.
Moviegoers astonished by computer morphing in Terminator 2.
Baby Bells get government permission to offer information services.
Collapse of Soviet anti-Gorbachev plot aided by global system called the
Internet.
More than 4 billion cassette tape rentals in U.S. alone.
3 out of 4 U.S. homes own VCRs; fastest selling domestic appliance in
history.
Sony Training Services 26

35. Broadcast Fundamentals
1992
Cable TV revenues reach $22 billion.
At least 50 U.S. cities have competing cable services.
After President Bush speaks, 25 million viewers try to phone in their
opinions.
1993
A TV Guide poll finds that one in four Americans would not give up
television even for a million dollars.
Dinosaurs roam the earth in Jurassic Park.
Unfounded rumors fly that cellphones cause brain cancer.
Demand begins for "V-chip" to block out violent television programs.
1 in 3 Americans does some work at home instead of driving to work.
1994
After 25 years, U.S. government privatizes Internet management.
Rolling Stones concert goes to 200 workstations worldwide on Internet
"MBone."
To reduce Western influence, a dozen nations ban or restrict satellite
dishes.
Prodigy bulletin board fields 12,000 messages in one after L.A. quake.
1995
CD-ROM disk can carry a full-length feature film. (CD-Video)
Sony demonstrates flat TV set.
DBS feeds are offered nationwide.
Denmark announces plan to put much of the nation on-line within 5
years.
Major U.S. dailies create national on-line newspaper network.
Lamar Alexander chooses the Internet to announce presidential
candidacy.
There are over a billion television sets in operation around the world.
2002
Bibliotheca Alexandrina is due to open on April 23. This is intended as
the modern equivalent to the ancient Alexandria Library which burnt
down about 1600 years ago with great loss of information and human
understanding.
27 Sony Broadcast & Professional Europe

36. Part 2 – The history of television
Sony Training Services 28

37. Broadcast Fundamentals
Part 3 Image perception & colour
The human eye
Evolutionary advantage
The human eye is a marvel of evolution and selective breeding. Mapping
the evolutionary history of the eye is difficult but almost certainly started
as in some ancient creature that possessed a group of especially light
sensitive cells on the surface of its skin.
The advantage of being able to sense possible attack, the presence of
possible food and a possible mate, must have been a very big
advantage. The eye must have evolved quickly from one generation of
creature to another.
It is perhaps easy to see how the light sensitive cells became better, and
how the ability to see colour must have given creatures a clear
advantage over those that could not. Even the ability to see a wide
spectrum of colours must have helped creatures.
Exactly how the lens evolved is less clear. However the lens started its
evolution is obviously gave those creatures that possessed them the
ability to see with greater clarity. It is also not clear why certain
evolutionary paths favoured the multi-lens compound eye, and why
others favoured the single lens design.
Evolution has not been entirely favourable, especially to humans. The
human eye is not perfect. It has a few drawbacks, most of which we
have adapted to. Some of these shortcomings actually make it easier to
design television, as we will see later.
What is the eye
Most of us have two working eyes. Sight in humans is more important
than any of our other senses. If either or both of our eyes fails to work, it
is one of the most disabling disabilities humans can have.
The eye grows from rudimentary skin cells before we are born. Neural
connections are made directly to the brain early on in development, and
what results is one of the most complex and wonderful structures in the
human body.
The eye’s structure
As far as broadcast video is concerned most of the complexity of the
human eye is irrelevant. However there are a few features and facts
about the eye that are interesting.
The human eye approximates to a sphere. In fact for somebody with
perfect sight the back of the eye is very close to a perfect sphere.
The eye is filled with a jelly like fluid called the vitreous humor. This fluid
keeps the eyeball in shape. The fact that its is clear means that light can
pass through it from front to back.
29 Sony Broadcast & Professional Europe

38. Part 3 - Colour
The front of the eye is covered with a clear protective film called the
conjunctiva. Behind this is another protective film called the cornea.
Just behind this is the iris, a muscular ring that allows the amount of light
entering the eye to be regulated. In bright light the iris closes. The iris is
tinted. There appears to be no reason why this is so, but this is what
gives the eye its ‘colour’.
Between the cornea and the iris is a watery fluid called the aqueous
humor. This keeps the front of the eye in shape.
Behind the iris is the lens. A marvel of evolution this organic structure
focuses light to the back of the eyeball. The amazing thing about this
lens is that its shape can be altered to change the focal length. The
cillary muscle, a small muscle surrounding the lens, squashed it and
allows the eye to focus on closer objects. When the muscle relaxes the
eye focuses to infinity.
Figure 1 The human eye
(Lens optics is discussed in a later chapter.)
As mentioned the back of the eye is almost spherical. It comprises a
large structure called the retina.
The retina
The retina is a structure that senses light and colour, and sends this
information to the brain. It is between 200 and 250 microns thick and
comprises various layers.
The outermost layer is a pigment layer. This acts as the outer wall to the
retina and as a light stop.
Inside this is the receptor cells. There are two types of receptor cells.
One type are rod shaped, and the other fatter and are cone like. For this
Sony Training Services 30

39. Broadcast Fundamentals
reason they are commonly referred to as rod and cones. Light hitting
these cells starts a protein electro-chemical reaction in a material called
rhodopsin. This reaction quickly passes along the length of the cell’s
axion. The end of the axion connected to the axion of a nerve cell,
called a bipolar cell, via a structure called a synapse. A synapse is not
actually a connection, but a small gap across which a protein electro-
chemical transfer takes place.
Figure 2 The human retina
Once the transfer has taken place another protein electro-chemical
reaction travels the length of the bipolar cell's axion to its body, and then
out along another axion to another synapse.
This second synapse connects to another nerve cell called the ganglion
cell. The signal passes down the ganglion’s axion using the same
reaction mechanism. The ganglion cell’s axions pass across the inner
surface of the eyeball and out through the nerve bundle, out of the eye.
The bundle passes back into the head and directly to the brain.
Light therefore has to pass through the whole thickness of the retina
before hitting the rods and cones.
31 Sony Broadcast & Professional Europe

40. Part 3 - Colour
Rods
Rod receptor cells have a broad sensitivity range. They are most
sensitive to green, which is nearer the centre of the optical electro-
magnetic spectrum.
Figure 3 Rod and cone cells
Rod cells measure the brightness of the image, or put another way the
black and white parts of the image.
Cones
Cone receptor cells have a narrow sensitivity range. There are three
types of cone cell. The first is sensitive to about 440nm wavelength light
(blue), the second to about 530 nm (green), and the third to about
560nm (red)
Cone cells are therefore responsible for seeing colour. Every colour is a
mix of blue, green and red.
Receptor density across the retina
There are about 120,000,000 rod cells in the retina and about 7,000,000
cone cells.
About 64% of the cone cells are sensitive to red light, about 32% to
green light and just 2% to blue light.
Most of the retina is the same, with an even concentration of rod and
cone cells. However there are two areas of the retina where this even
concentration is different, the fovea and the blind spot.
Sony Training Services 32

41. Broadcast Fundamentals
The fovea
The lens focuses the centre of the image to a point on the retina called
the fovea. This area of the retina has a very dense concentration of
receptor cells. Furthermore, all these cells are cones. There are no rod
cells in the fovea. The fovea allows the eye to study the centre of an
image or scene in great colour detail.
The blind spot
Because all the ganglion cell axions are on the inside of the retina they
need to pass out of the eyeball at some point. It stands to reason
therefore that wherever this point is there can be no receptor cells at all.
This area is therefore known as the blind spot.
Interesting facts about the eye
The eye is far from perfect
Although the eye is a marvel of biological engineering it has a number of
design flaws. The cornea, lens and vitreous humor are not absolutely
clear. They all reduce the amount of light hitting the retina and colour it
slightly.
The eye’s image is bent out of shape and upside down
The image falling on the retina is reasonably well proportioned near to
the fovea. However the nearer you get to the outer edge the more
compressed and distorted the image becomes.
The lens also focuses the image upside down and back to front on the
retina.
The brain corrects for imperfections
The brain corrects the image to remove colour casting from the cornea,
lens and vitreous humor. It also corrects edge distortion giving us the
impression of a flat correctly proportioned image.
Having two eyes allows us to measure distance
When focusing of close object, not only do the lenses squash to focus,
but also the eyeballs turn towards each other. This can be used by the
brain the measure how far away an object is.
You can see this happening by getting a friend to hold their finger up at
arm’s length, and focus on it. Then ask them to keep focussing on the
finger while slowly moving it closer to their face.
The eye gets board easily
The eye is very good as seeing change. If you stare at something long
enough it will disappear. The brain eventually cancels the image out
altogether. Thus the eye works best if it continually moves, scanning
across edges and shapes continually updating what the brain receives.
33 Sony Broadcast & Professional Europe

42. Part 3 - Colour
Images can ‘burn in’ to the retina
Linked to the last interesting fact, if you stare at something long enough
it will appear to disappear but the image is ‘burnt in’. If you then look at
something else the original image will appear in negative for a while.
The eye remembers
The protein electro-chemical reactions in the eye’s cells that sense light
and pass the signals back to the brain take a certain time to react and
stop.
A flash is therefore ‘stretched’ so that the eye effectively sees it for
longer than it actually occurs. This effect is known as persistence of
vision. Film and television rely heavily on persistence of vision to turn
what is actually many still images flashing one after another, into what
appears to be a constantly changing image.
The eye is good at seeing patterns
The eye can pick out patterns very well. This is a problem for television
and digital imagery because lines and pixels tend to stand out.
For instance a digital photograph will appear to be not as good an image
when compared to a conventional photograph of exactly the same
resolution, because the digital photograph pixels have a pattern and the
conventional photograph grains are random.
The eye is very sensitive to green
A third of the cone cells are sensitive to green. The rod cells, although
intended for seeing the overall brightness of an image. are more
sensitive to green.
This makes the eye sensitive to green and more sensitive to changes in
the green part of the spectrum.
This has an important impact of the design of colour television.
The fovea is not good for dark vision
Rod receptor cells are more sensitive that cone receptor cells. Thus in
dark conditions things appear to turn black and white.
It is best not to look at something directly in low light, but to look just to
the side or above it. This will put the object on the retina where there are
plenty of rod cells and you will be able to see it. (Incidentally, it may not
be a good idea to look just below an object in dark conditions as you
may put it into the blind spot, where you can’t see it at all.)
Sony Training Services 34

43. Broadcast Fundamentals
The concept of primary colours
Any colour can be described as a combination of 3 primary colours.
Children are often taught that the three primary colours are Red, Yellow
and Blue. This is a perfectly reasonable assumption when learning
painting and art. Mixing these colours allows children to make almost
any colour they want.
Figure 4 Children’s primary colours
Subtractive colour mixing
This concept is called subtractive colour mixing, because the overall
colour gets darker the more paint you add to the mix.
Figure 5 Subtractive primary colours
In reality Red Yellow and Blue are not the correct primary colours for
subtractive colour mixing. The reason for this is that mixing Red , Blue
and Yellow does not give Black, it makes Brown. True subtractive
primaries should remove all colour and brightness when mixed together,
i.e. Black.
35 Sony Broadcast & Professional Europe

44. Part 3 - Colour
The true subtractive primaries are Magenta, Yellow and Cyan. While
these three colours might appear close to Red, Yellow and Blue as far
as children are concerned, they are sufficiently different to go to Black
when mixed together in equal proportions.
Additive colour mixing
The opposite of subtractive colour mixing is additive colour mixing.
Additive primary colours are relevant to light. If three additive primary
coloured lights are mixed in equal proportions the result is White light.
The three additive primary colours are Red, Blue and Green.
Figure 6 Additive primary colours
Secondary and tertiary colours
Each set of primary colour has a set of secondary colours. If you mix any
two of the primary colours in equal proportions you will get a secondary
colour.
In fact the three subtractive primary colours are the secondary colours of
the additive primary colours, and visa versa.
A tertiary colour is found by mixing equal proportions of three primary
colours. There are only two tertiary colours, White and Black.
Sony Training Services 36

45. Broadcast Fundamentals
Hue saturation and luminosity
Colour can be described as a 3 dimensional shape. At the top is white.
Half way down is a circle of all the colours at their full intensity. You can
see all six primaries, both additive and subtractive around the edge of
the circle. At the bottom is Black.
W h ite W h ite
B lu e Red Y e llo w C yan
M a g e n ta G re e n
B la c k B la c k
Figure 7 Colour 3D shape (front & back)
This is a 3 dimensional space, therefore it is possible to pick any point at
have any colour you want. The line running down the centre runs from
White to Black through Grey.
G re e n G re e n
Y e llo w Y e llo w
C yan C yan
W h ite B la c k
B lu e Red
Red B lu e
M a g e n ta M a g e n ta
Figure 8 Colour 3D shape (top & bottom)
37 Sony Broadcast & Professional Europe

46. Part 3 - Colour
If you look at the shape from the top you will get a circle with all the
colours around the edge and White in the middle. Look at the shape
from the bottom and you will see Black in the middle.
Figure 9 Hue saturation & luminosity
Hue
Hue is the colour. You can change the hue by rotating around the centre
of the circle.
Saturation
Saturation can be called colour intensity. It is a measure of how far from
the centre you are. Zero saturation is White, Grey or Black. Full
saturation is somewhere on the edge of the circle.
Luminosity
Luminosity is how far up or down the shape you are. If you take any
colour and force its luminosity up it will tend towards White, and visa-
versa down to Black.
The CIE space
The common method for describing colour is the CIE colour space. This
2 dimensional representation is used for additive colour fields to define
the ability of a video system to capture and display colour. As you can
Sony Training Services 38
Figure 10 Hue, saturation & luminosity

47. Broadcast Fundamentals
see the NTSC and PSL gamuts are well within the total range of natural
colours.
Each corner of the gamut triangles for NTSC and PAL specify the
primary colours. They are different for each standard.
Television cameras and displays have a long way to go before they are
able to capture and display every colour available in nature.
Figure 11 CIE colour space
39 Sony Broadcast & Professional Europe

48. Part 4 – The basic television signal
Part 4 The basic television signal
The problem of getting a picture from A to B
A picture is a 2 dimensional object. It has height and width. A moving
picture adds a third dimension, time, to the other two.
If we are to send a moving image from one place to another we need to
change the image content into a serial signal.
Film frames
Film conveys a moving image as a series of frames. These are like 2
dimensional chunks of data appearing at once, one after another, so
rapidly that it appears to be smooth.
The raster scan
A raster scan scans an image and turns it into a serial stream of data.
By combining film’s method of conveying frames with the raster scan
method we could convey a moving image as a serial signal.
The basic raster frame
The normal raster scan, and the method used by all broadcast television
standards, scans each line from left to right, and the each successive
line from top to bottom. This is called a frame.
The definition of the signal itself is simple. The brighter the image is at
that point on the line the higher the signal’s voltage.
Figure 12 The raster scan
Sony Training Services 40

49. Broadcast Fundamentals
Lines and frame rate
We need to decide on a frame rate and the number of lines. We want
the highest quality possible so it would be better to have as many lines
as possible, and as many frames per second as possible. We would also
want to ensure that each line had the highest quality (bandwidth)
possible.
However the overall bandwidth of the signal is strictly limited by
broadcast standards authorities, so we have to find a reasonable
compromise between the number of frames per second, the number of
lines, and the quality of each line.
An increase in either the number of lines, the number of frames per
second, or the bandwidth of each line, will increase the signal’s overall
bandwidth.
The blanking intervals
Horizontal blanking
Each raster line is normally referred to as the active line. This is where
the line is traced out on the image. There is a short interval between one
active line and the next. The scanning system uses this time to fly back
to the beginning of the next line. The signal is ‘cut’ for this period of time
to prevent the flyback appearing on the television set. In the raster
scanning system this interval is referred to as the horizontal flyback.
The interval is also called the line blanking interval or horizontal
blanking interval.
Thus every video line consists of the active line period and the horizontal
blanking interval which is used as a flyback period.
Vertical blanking
There is a longer interval between one entire scan and the next. During
this time the scanning system moves back from the bottom right corner
to the top left corner. Just as with the horizontal flyback interval the
signal is ‘cut’, to prevent it appearing on the television screen. This is
referred to as vertical flyback.
This is normally also called the vertical blanking interval.
Interlaced raster scanning
If a frame is raster scanned and the frame rate is the same as that of film
i.e. 24 frames per second, there is a severe amount of picture flicker.
This is because every point in the image will have faded before the
scanning mechanism can go back around to refresh it.
41 Sony Broadcast & Professional Europe

50. Part 4 – The basic television signal
H o r iz o n ta l fly b a c k
(o n ly o n e s h o w n )
E v e n lin e s
(fie ld 1 ) O d d lin e s
(fie ld 2 )
V e r tic a l fly b a c k
fr o m fie ld 2 to 1
V e r tic a l fly b a c k
fr o m fie ld 1 to 2
Figure 13 The interlaced raster scan
Televisions could be designed to reduce flicker by increasing the
persistence of the screen. However this would mean any rapid
movement on the screen would be seen as blurring and streaking.
You could increase the frame rate but this would increase bandwidth.
The solution is to interlace the raster scan. Interlaced scans scan the
odd numbered lines first, from top to bottom. Then the raster scan starts
from the top again and scans the even lines from top to bottom.
This method of scanning reduces flicker by effectively writing an image
at twice the frame rate.
Each of the scans is called a field, and two interlaced field make up a
frame.
Half lines
Modern video standards also take into account that each line in the
raster scan is not exactly horizontal. In fact the raster scan is
progressing slowly from the top of the image to the bottom at a constant
rate. The left side of each line is actually slightly lower than the right
side.
Therefore video standards have an odd number of lines per frame. The
first field of each frame begins with a whole line and ends with a half
line. The second field begins with a half line and ends with a whole line.
This system gives a more rectangular raster scanned image.
Sony Training Services 42

51. 43
H o r iz o n t a l A c tiv e
b la n k in g v id e o
Figure 14
H o r iz o n t a l
syncs
Broadcast Fundamentals
V e r t ic a l f ly b a c k H a lf lin e V e r t ic a l b la n k in g
( V e r t ic a l b la n k in g )
A c t iv e v id e o lin e
A c tiv e v id e o
H o r iz o n t a l s y n c s H o r iz o n t a l b la n k in g
H o r iz o n ta l fly b a c k
( H o r iz o n t a l b la n k in g )
V e r t ic a l b la n k in g
H o r iz o n ta l b la n k in g
E q u a lis in g p u ls e s
L in e s y n c
B r o a d p u ls e s
V e r t ic a l E q u a lis in g p u ls e s L in e s y n c
b la n k in g
lin e
V e r t ic a l b la n k in g
Basic horizontal and vertical detail
Sony Broadcast & Professional Europe

52. Part 4 – The basic television signal
Synchronisation
The basic principle
Synchronisation is the principle of making sure two pieces of equipment,
that both have some kind of regular clock or rhythm to run at the same
rate. The two pieces of equipment are said to be ‘locked’ together.
Synchronisation is often done with some form of synchronisation signal,
generally simply called a sync signal.
How does television sync?
All television equipment contains some form of clock or oscillator. This
will have a natural frequency which is close to the correct frequency.
Somewhere in the television transmission station will be a master sync
pulse generator containing a precision master oscillator. Its frequency is
correctly set to within 1 cycle in several million.
All the equipment in the transmission station is locked to this master
sync pulse generator. This is easy because the equipment’s own clock is
running as about the same rate. The sync signal ‘pulls’ the equipment’s
own oscillator to exactly the correct frequency.
The transmission station will send out a television signal that contains
sync pulses. All equipment from the transmission station to the television
at home contains similar oscillators which are ‘pulled’ to exactly the
correct frequency by the sync pulses.
Line, or horizontal, sync pulses
Line sync pulses are parts of the video signal that define the beginning
of each video line. They occur at a certain time during the horizontal
blanking interval.
Line sync pulses are short intervals of time where the video signal drops
below the voltage specified for black (the blanking level).
Line sync pulses have a particular shape, because they are bandwidth
limited. The beginning and end of the pulses are sloped. The beginning
of the video line is specified as the mid-point of the slope at the
beginning of the sync pulse.
These pulses are placed some time during the horizontal interval. Their
position relative to the beginning of the active line is set and known, so
once the position of the pulse is found the beginning of the active line is
known.
Vertical sync pulses
The vertical blanking interval is more complex, and is relatively longer,
than the horizontal blanking interval. The time interval is the same as
many video lines. It contains a complex series of pulses that define the
beginning of each field and each frame.
Sony Training Services 44

53. Broadcast Fundamentals
Blanked vertical lines
The vertical interval starts and finishes with a few blanked video lines.
These are simply video lines with their respective horizontal sync pulses,
but with the active line period blanked as well.
Equalising pulses
The vertical interval contains a number of equalising pulses near to the
end of one field and the start of the next field. Equalising pulses are
shorter than line sync pulses and occur every half line.
The reason for this is so that there is a same pattern of equalising pulses
for every field, even though the transition between the first field and the
second is half way through one line.
Broad pulses
Broad pulses are placed in between the two groups of equalisation
pulses. These are very wide pulses, in fact, so wide that only a small
portion of time is spent not in a broad pulse.
Definition of the start of the field
The definition of the start of each field is the beginning of the first broad
pulse.
The oscilloscope
An oscilloscope is an instrument that allows engineers to view video
signals, not as a picture, but as a constantly changing signal. It
shows the kind of signals show on page 43 as a bright trace across
the display.
All oscilloscopes have more than one input, so that various signals
can be compared to one another. They also often allow for complex
triggering so that complex or intermittent signals can be caught and
studied.
Most oscilloscopes use tubes, similar to monochrome televisions.
The most modern ones use digital flat screen colour technology.
Engineers use the oscilloscope to check the levels and timings of
video signals.
45 Sony Broadcast & Professional Europe

54. Part 5 – The monochrome NTSC signal
Part 5 The monochrome NTSC signal
The 405 line system
The first important monochrome video signal was the 405 line
monochrome system adopted by many countries around the world.
Although an important video standard in its time, the 405 line standard is
now obsolete. Furthermore it is different from any of the modern video
standards.
Therefore we will look at the 525 line monochrome standard as the first
important and relevant standard.
The 525 line monochrome system
The 525 line monochrome standard was proposed by the American
NTSC (National Television Standards Committee) and quickly became
popular.
This standard formed a strong basis for the existing 525 line colour
system used by many countries around the world, and so it seems
sensible to study it first.
The 525 line monochrome system has 525 lines per video frame, with
262.5 lines per field.
Frame rate and structure
The chosen frame for NTSC was 30 frames per second or 60 fields per
second. It is commonly thought that this was so that NTSC televisions
could be locked to the mains power. This is only half true. Mains power
alternating frequency is not accurate enough to provide a reliable
synchronisation signal for television receivers. Television equipment
does not use mains as a locking signal. However if television equipment
is not somehow linked to mains, the resulting beating and aliasing
frequencies can cause undesirable effects on the screen. Making the
frame rate the same as the mains power frequency at least makes these
undesirable effects stand still.
Field 1 starts on line 1. There are 6 equalisation pulses, then 6 broad
pulses, then 6 more equalisation pulses. Normal horizontal syncs starts
on line 10.
The first active video line is line 22, and the last is line 262. Half of line
263 is active.
Field 1 starts half way through line 263 with 6 equalisation pulses, 6
broad pulses and 6 more equalisation pulses. Normal horizontal syncs
start at the beginning of line 273.
The first active video line is line 285 and the last is line 525.
Sony Training Services 46

55. Broadcast Fundamentals
Field start displacement
The trigger point for the start of a field is normally the first broad pulse.
This is the point television receivers use to start the next field. However
the ‘official’ start point of each field is the beginning of the first broad
pulse.
This gives a discrepancy between the technical start of each field and
the line numbers. The first broad pulse is at the start of line 4, and half
way through line 266.
Line rate and structure
The line rate is simply the frame rate multiplied by 525, or 15.75kHz.
Disregarding the vertical interval, all NTSC lines have the same basic
structure.
Bandwidth considerations
The video signal can have energy that can stretch to 10Mhz and
beyond. However, because of the highly repetitive nature of video, with
each video line similar to the one before and after, and each video field
and frame similar to the one before and after, most of the energy is
centred around harmonics of line, field and frame rate. This makes the
bandwidth look like a series of spikes, with very little between them.
V id e o s ig n a l b a n d w id th
A m p litu d e
H a r m o n ic s to 1 0 M H z o r h ig h e r
DC F re q u e n c y
1 lin e 1 fra m e
15750H z 30H z
Figure 15 Video signal bandwidth
The television signal is modulated onto a radio frequency carrier before
being sent to the transmitter mast and out to the home. The harmonics
may possibly spread out either side of the carrier to 10MHz or more,
47 Sony Broadcast & Professional Europe

56. Part 5 – The monochrome NTSC signal
giving a possible total bandwidth of over 20MHz. These spreads either
side of the carrier are called the upper and lower sidebands.
The regulatory authorities assigned a 6MHz bandwidth to each television
channel. The designers of the original television standard therefore had
to devise a scheme for restricting the video signal down to a 6MHz limit.
L o w e r s id e b a n d U p p e r s id e b a n d
A m p litu d e
V H F c a r r ie r F re q u e n c y
DC fre q u e n c y
L o w e r s id e b a n d
filte r e d le a v in g U p p e r s id e b a n d
v e s tig a l s id e b a n d filt e r e d to 4 .2 M H z
V H F c a r r ie r
fre q u e n c y A u d io c a r r ie r
- 1 .2 5 0 4 .2 4 .5 4 .7 5
6 M H z t o ta l c h a n n e l b a n d w id th
T e le v is io n c h a n n e ls
o n th e r a d io s p e c tr u m
A u d io V id e o A u d io V id e o A u d io V id e
c a r r ie r c a r r ie r c a r r ie r c a r r ie r c a r r ie r c a rri
C h a n n e l b a n d w id th C h a n n e l b a n d w id th
(6 M H z ) (6 M H z )
Figure 16 Video channel bandwidth
Filters are used to cut off as much of the lower sideband as possible. It
is not possible to cut everything off, so the filter restrains the lower
sidebands to just 1.25MHz. What is left is commonly called the vestigal
sideband.
Sony Training Services 48

57. Broadcast Fundamentals
The upper sideband is filtered and restricted to about 4.2MHz. Filters
cannot create a sharp clean cut-off at 4.2MHz, but rather a smooth roll-
off that disappears to zero just below 4.5MHz.
A simple audio carrier is placed at 4.5MHz, clear of the video signal. Its
sidebands do not extend very far and there is nothing left at 4.75MHz.
Thus the total bandwidth including the video and audio signals is
constrained to 6MHz.
Quality considerations
The low frequency detail of the video signal is centred around the carrier
frequency and the low order sidebands. Fine detail is centred around the
high frequency sidebands above 4MHz. It is worth remembering that
random noise will also tend to be centred around the high frequency
sidebands.
Most home television receivers cannot show much detail above 4MHz. It
is therefore pointless trying to transmit this level of detail to the home.
Radio spectrum and television channels
The regulatory authorities specified a series of analogue television
channels 6MHz apart. Each video carrier is 1.25MHz from the bottom of
the channel, and each audio carrier is 5.75MHz from the bottom of the
carrier (4.5MHz above the video carrier).
Television companies have the responsibility to ensure that each
channel they transmit has carriers at exactly the correct allocated
frequency, and that the bandwidth is properly filtered to constrain it to
within the 6MHz limit.
49 Sony Broadcast & Professional Europe

58. Part 6 – Colour and television
Part 6 Colour and television
Using additive primary colours
The additive primary colour principle has particular relevance to
television because it uses light to detect the image in the camera and to
display it in the television set at the other end.
The colour television camera splits the image into three separate
images, one for the Red part of the image, one for Green and one for
Blue.
The colour television set has three sets of phosphor dots, one type that
shine Red, one type that shine Green and the last type that shine Blue.
The television set also has three electron guns, each one targeting one
set of phosphor dots.
Original plans
The original idea was to use Red, Green and Blue throughout the whole
transmission system, from camera to television set.
Both RCA and CBS (amongst others) developed systems that
sequentially send Red, Green and Blue parts of the original image over
a normal monochrome transmission system.
RCA chose to send ‘dots’ of colour, Red Green Blue in a rotating
sequence. CBS chose to filter successive video fields through a rotating
Red Green Blue filter wheel.
Neither system worked too well. What was needed was a system that
added some colour element to the existing monochrome signal, allowing
those with monochrome television sets to watch television with the same
quality as before, but allowing those with new colour televisions to see
that same basic quality of image in colour.
Ensuring compatability
The original ideas for a colour television system were not popular
because they were not compatible with the existing monochrome
standard.
A compatible monochrome signal
The colour video camera produces three separate Red, Green and Blue
images. In theory simply mixing these three in equal proportions should
give a perfect White.
The human eye is more sensitive to different Green than to Red or Blue.
Therefore any miss-calculations in generating the Green primary colour
in the television set would be more obvious that for Red or Blue.
The proportions of Red, Green and Blue was therefore adjusted to
match human eye characteristics, standard luminosity curves, and make
up for the non-linear nature of the light/voltage characteristics of a
Sony Training Services 50

59. Broadcast Fundamentals
standard video camera and voltage/light characteristics of the standard
television set. The equation for White (Y) is therefore :-
Y = 0.299R + 0.587G + 0.114B
This provided for a perfectly balanced monochrome signal that could be
used to generate a standard monochrome signal compatible with
existing monochrome television sets.
Maintaining compatible channel bandwidth
The regulatory authorities are constantly being pressured to provide
space on the limited radio spectrum for all kinds of radio services,
including commercial radio and television stations, airline radio
communications, ambulance, police and other emergency services,
radio control model enthusiasts, citizen band radio and HAM radio.
They were therefore not prepared to allocate more of the precious radio
bandwidth to television companies wanting to switch from monochrome
to colour.
Designers had therefore to somehow fit the colour television signal into
the existing 6MHz allocated to them for monochrome television.
Adding colour
Colour difference signals
There are three theoretical colour difference signals, each one being the
difference on a primary colour from White. The colour difference signals
are therefore (R-Y), (G-Y) and (B-Y).
It is possible to generate any hue, saturation of luminance using any
three of the four signals, Y, (R-Y), (G-Y) and (B-Y). It is also possible to
generate the Y signal or any of the three primary colours from any three
of these four signals.
The Y signal was an essential requirement of any compatible colour
system. As already mentioned this is the same as a standard
monochrome signal.
A decision had to be made as to which two of the three available colour
difference signals would be used.
The Green signal is a much higher proportion of Y than either Blue or
Red. Therefore any miscalculation in Green will not be so obvious as it
would be for either Red or Blue.
It was therefore decided to use the Red and Blue colour difference
signals (R-Y) and (B-Y).
Generating the (R-Y) and (B-Y) colour difference signals
Generating the colour difference signals is a simple piece of
mathematics. The relationship between the Y signal and the three
primary colour signals has already been established. Thus the (R-Y)
colour difference signal is simply :-
R-Y = R - (0.299R + 0.587G + 0.114B)
51 Sony Broadcast & Professional Europe

60. Part 6 – Colour and television
= R-0.299R – 0.587G – 0.114B
= 0.701R – 0.587G – 0.114B
Likewise the (B-Y) signal can be derived in the same way.
B-Y = B - (0.299R + 0.587G + 0.114B)
= –0.299R – 0.587G + B – 0.114B
= –0.299R – 0.587G + 0.886B
Component colour video signals
Component colour video signals can either be in the original R, G, B
form, or more commonly, are defined as the Y, (R-Y) and (B-Y) signals.
Their relationship to the original primary colour signals is as previously
mentioned, i.e. :-
Y = 0.299R + 0.587G + 0.114B
(R-Y) = 0.701R – 0.587G – 0.114B
(B-Y) = –0.299R – 0.587G + 0.886B
Sony Training Services 52

61. Broadcast Fundamentals
Using component video
The advantage of component video is that the three signals are kept
apart. This ensures that the quality is kept as high as possible.
Analogue component video equipment has completely separate
electronics for the three signals. Connections between different
pieces of equipment if always done using three separate cables and
connectors, or using one cable with three cable cores.
Having three separate sets of electronics, and three separate
connections is obviously more expensive, compared to the needs of
monochrome television, therefore analogue component video is
something normally retained for broadcast and professional use.
Component connections
Analogue component is normally connected between different
pieces of equipment using three 75 ohm coaxial cables and three
BNC connectors at each end. The cables should be about the same
length although this only becomes a problem where the difference
between the cables is greater than several metres.
BNC cables are often cut to the same length, tied together as a
triple, and made up with colour coded BNC connectors at both end.
Conventionally red, green and blue colour coded connectors are
used. If R,G,B component video is being used the connectors are
used as is. Conventionally, if there is a sync signal provided, it is
normally on the green signal, hence the phrase “Sync on Green”.
With Y,(R-Y),(B-Y) component video the red colour coded
connectors are conventionally used for the (R-Y) connection, the
blue connectors for the (B-Y) connection, and the green ones for
the Y connection.
.
53 Sony Broadcast & Professional Europe

62. Part 6 – Colour and television
Figure 17 Basic component video signal
Combining R-Y & B-Y
Colour television requires that there be just one colour signal. This must
therefore be a combination of the R-Y and B-Y contributions. The
designers of the first popular colour television standard decided to
combine the two colour difference signals onto a special carrier called a
subcarrier, because its frequency is under the main RF carrier used to
transmit the video signal.
Quadrature amplitude modulation
The designers devised an ingenious way of modulating the two colour
signals onto one carrier by using quadrature amplitude modulation.
Amplitude modulation is simple to achieve and to understand. The
amplitude of the carrier is simply the level of the signal being modulated.
What results is a steady frequency sine wave with varying amplitude.
Sony Training Services 54

63. Broadcast Fundamentals
S u b c a r r ie r
S u b c a r r ie r v e c to r
Q u a d r a tu r e c a r r ie r Q u a d r a t u r e c a r r ie r v e c to r
Figure 18 Quadrature vector representative
This sine wave can be thought of as a rotating vector. The vector is
rotating in an anticlockwise direction and its length defines the amplitude
of the signal.
It becomes easy to see how two signals can be modulated onto one
subcarrier when you consider them as vectors. One of the signals can
be modulated on a carrier that is delayed by 1/4 cycle (90 degrees).
When the two modulated signals are combined they will not interfere
with each other because they are 90 degrees apart.
The subcarrier is sent with the video signal. This makes decoding the
two colour difference signals easy. You simply look at the amplitude of
the signal in phase with the subcarrier, and the amplitude of the signal
90 degrees out of phase.
Video signal spectra
The monochrome signal
As explained on page , the monochrome video signal is highly repetitive,
and the signal’s spectrum is not a smooth spread of energy from DC to
high frequency. There are very definite energy peaks at harmonics of
line rate, field rate and frame rate. There is very little energy between
these peaks.
When the monochrome signal is modulated onto the radio carrier,
sidebands would normally extend above and below the carrier with
energy peaks at harmonics of line, field and frame rate. However the
monochrome signal is filtered, and the lower sidebands are cut. The
upper sidebands are filtered to about 4.2MHz.
55 Sony Broadcast & Professional Europe

64. Part 6 – Colour and television
The colour signal
The basic colour signal has the same bandwidth as the monochrome
signal. When modulated onto the subcarrier it has the same basic
bandwidth extending either side of the subcarrier frequency, and its
energy is centred around harmonics of line rate, field rate and frame
rate.
If this signal is added to the monochrome signal it would be swamp it.
The upper sidebands would also extend way beyond the total 6MHz
channel bandwidth allowed by the regulatory authorities. The colour
signal is therefore attenuated so that its total bandwidth is very much
smaller. This prevents it from interfering with the monochrome signal,
and constrains it to within the total channel bandwidth.
Combining monochrome and colour
C o lo u r s u b c a r r ie r
C o lo u r s id e b a n d s A u d io c a r r ie r
Y s ig n a l
1 lin e 1 fra m e
1 5 7 3 4 .2 5 H z 30H z
C o lo u r s id e b a n d s C o lo u r s u b c a r r ie r Y s ig n a l
Figure 19 Mixing colour and monochrome together
Sony Training Services 56

65. Broadcast Fundamentals
The subcarrier signal is not only a useful way of combining the two
colour difference signals into one signal, by using quadrature
modulation. It is also a neat way of combining the colour signal with the
monochrome signal.
Both the monochrome and colour signals have spectra with energy
centred around the harmonics of line rate, with gaps between each
harmonic. If the frequency of the subcarrier is carefully chosen the
colour harmonics can be made to sit exactly between the monochrome
harmonics.
The subcarrier is also chosen to be as high as possible, but ensuring
that the whole colour signal is within the 4.5MHz overall bandwidth of the
complete video signal.
Placing the subcarrier as high as possible also ensures that, if there is
any interference between the monochrome and colour parts of the signal
it will only affect the high frequency fine detail of the picture.
The final video signal, with Y, (R-Y) and (B-Y) mixed together is called
composite video.
Using composite video
Composite video is a very popular method of connecting analogue
video between pieces of video equipment and only requires a single
cable.
However, although analogue component video requires three
cables to connect pieces of video equipment together it has higher
quality compared to analogue composite video.
P in a s s ig n m e n t
P lu g S ocket
C o re
C o n n e c tio n F u n c tio n I/ P o r
O /P
C o re S ig n a l I& 0
S h ie ld G ro u n d -
S h ie ld
P a n e l c o n n e c to r C a b le c o n n e c to r
‘T ’ p ie c e Te rm in a to r
P lu g
P lu g
S ocket
57 Sony Broadcast & Professional Europe

66. Part 7 – Colour NTSC television
Part 7 Colour NTSC television
Similarity to monochrome
The colour NTSC television signal is based on the monochrome NTSC
signal. It has exactly the same number of lines per frame and per field.
The active video region is the same, as is the structure of the vertical
blanking region, the vertical sync pulses, the horizontal blanking region
and the horizontal sync pulses.
Choice of subcarrier frequency.
The subcarrier frequency was originally chosen to be between the 227th
and the 228th harmonic of line rate. This would make it 3,583,125Hz.
I s ig n a l A u d io c a r r ie r
Y s ig n a l Q s ig n a l
S u b c a r r ie r
1 lin e 1 fra m e
1 5 7 3 4 .2 5 H z 30H z
Figure 20 Colour NTSC bandwidth
Sony Training Services 58

67. Broadcast Fundamentals
However this frequency produces interference with the audio carrier
signal in the final television signal. The frame rate was therefore altered
slightly, from 30 frames per second to 29.97 frames per second. The
subcarrier was moved to 3,579,545Hz.
In hindsight this was probably not a good idea. It may have been better
to have moved the audio carrier slightly instead. The fact that NTSC is
not now an integer number of frames per second causes many problems
with standardisation and editing. (See page 207)
Adding colour
As mentioned the colour difference signal must be filtered before they
are modulated, to restrict their bandwidth compared to the monochrome
signal, to prevent them from interfering with it too much.
What is more, with a subcarrier frequency of about 3.58MHz, it would
appear that the colour bandwidth capacity above this frequency is only
about 0.5MHz before it starts to interfere with the audio signal carrier
itself.
The solution is not to modulate the R-Y and B-Y signals directly but two
other signals called I and Q.
It was found that the human eye was more sensitive to colour around the
orange/cyan axis, compared to white, than to colours 90 degrees from
this around the magenta/green axis, compared to white.
Thus two signals were generated, one called the I (in-phase) signal
which was modulated with the subcarrier on the orange vector, and the
other called the Q (quadrature) signal which was modulated with a
subcarrier signal phase shifted to a delay of 90 degrees.
The I signal
The I signal is found by the equation :-
I = 0.877(R-Y) cos 33deg – 0.493(B-Y) sin 33deg
= 0.74(R-Y) – 0.27(B-Y)
In terms of the original R, G and B signal, I can be described as :-
I = 0.60R – 0.28G – 0.32B
The I signal is asymmetrically filtered to a bandwidth of +0.5MHz and –
1.5MHz. This allows relatively high definition for the I signal.
The Q signal
The Q signal is found by the equation :-
Q = 0.877(R-Y)sin 33deg + 0.493(B-Y)cos 33deg
= 0.48(R-Y) + 0.41(B-Y)
In terms of the original R, G and B signal, Q can be described as :-
Q = 0.21R – 0.52G + 0.31B
59 Sony Broadcast & Professional Europe

68. Part 7 – Colour NTSC television
The Q signal is symmetrically filtered to a bandwidth of 0.5MHz. This
allows relatively low definition for the for the Q signal.
Burst
Between 8 and 10 cycles of subcarrier are sent during the horizontal
blanking interval of every line, except during part of the vertical blanking
region. The television receiver uses this to lock its own internal oscillator
to the correct frequency and phase so that it can determine colour
correctly.
For NTSC colour video, burst is defined as having a phase of 180
degrees.
The colour NTSC vector display
9 0 (R -Y )
100
R ed
103
90
8 8 IR E
M a g e n ta
80 61
8 2 IR E
70
60
I com ponent
50 33
40
30
Y e llo w
167 20
6 2 IR E
10
180 0 (B -Y )
B u rs t
180
4 0 IR E B lu e
347
6 2 IR E
G re e n
241
8 2 IR E
Q com ponent
C yan 303
283
8 8 IR E
270
Figure 21 Colour NTSC vector space
A good way of showing the colour elements of an NTSC signal is to
show the signal on a vector display. This is also sometimes called a
polar display.
The vector display shows the amplitude and phase of the chroma
(colour) signal. The display is circular. The centre represents zero
amplitude. An increase in amplitude is represented by a move towards
Sony Training Services 60

69. Broadcast Fundamentals
the outer edge of the display. The position around the display represents
the phase of the signal with respect to the sub-carrier.
Thus no colour will be seen as a bright dot in the centre of the display.
Fully saturated colour will be seen as a bright dot near the edge of the
display. The position of the dot will indicate the colour, or hue.
Burst appears as a bright dot to the left of the centre of the display (at
180 degrees).
The vectorscope
A vectorscope is a special kind of oscilloscope for measuring the
colour content of a composite video signal. It shows a vector display
of the signal and has a graticule with markings for the vertical and
horizontal axes, concentric circles for colour saturation and target
boxes for all the important colours and for burst.
Engineers use a vectorscope to check that composite video has the
correct colour phase and saturation and that there is no distortion
on the colour signal.
The gamut
Gamut is the limit the colour content of a composite signal can attain on
a vector display.
The NTSC gamut is not circular but more shaped like a rugby ball, i.e. it
is tall and thin.
No part of the signal can extent beyond the gamut as this will produce
illegal colours.
The gamut detector
A Gamut detector is an instrument that is connected between any
piece of video equipment and the composite monitor it is playing
into. It checks that the video signal is within gamut at every point,
and shows any area of the picture that has illegal colours as an
enclosed warning area on the monitor.
61 Sony Broadcast & Professional Europe

70. Part 7 – Colour NTSC television
Vertical interval structure
The vertical interval for NTSC is similar to that of monochrome NTSC.
The equalisation and broad pulses have the same basic construction.
4 field structure
Colour NTSC differs from monochrome NTSC in that it contains a
subcarrier signal and subcarrier burst elements at the beginning of each
video line, except during the vertical interval.
The phase relationship between subcarrier and horizontal syncs does
not repeat every frame, but every 2 frames, (4 fields). This relationship is
called the sc/h (subcarrier to horizontal sunc) relationship. The
sequence of fields and frames is called the colour frame sequence.
This relationship becomes important when editing. If the final video
sequence is to maintain a steady sc/h relationship after it has been
edited, edits must be made to the correct field in the 4 field sequence.
Sony Training Services 62

71. E n d o f f ie ld 4 S ta r t o f fie ld 1
63
V s y n c tr ig g e r
A c t iv e v id e o lin e s V e r t ic a l b la n k in g in t e r v a l A c t iv e v id e o lin e s
Figure 22
6 e q u a lis a tio n p u ls e s 6 b r o a d p u ls e s 6 e q u a lis a tio n p u ls e s
523 524 525 1 2 3 4 5 6 7 8 9 10 11 12 21 22 23
Broadcast Fundamentals
E n d o f f ie ld 1 S ta r t o f fie ld 2
H a lf lin e H a lf lin e
261 262 263 264 265 266 267 268 269 270 271 272 273 274 284 285 286
E n d o f f ie ld 2 S ta r t o f fie ld 3
523 524 525 1 2 3 4 5 6 7 8 9 10 11 12 21 22 23
E n d o f f ie ld 3 S ta r t o f fie ld 4
H a lf lin e H a lf lin e
261 262 263 264 265 266 267 268 269 270 271 272 273 274 284 285 286
Colour NTSC vertical interval (showing 4 field sequence)
Sony Broadcast & Professional Europe

72. Part 8 – PAL television
Part 8 PAL television
What is PAL?
PAL stands for Phase Alternate Line. It is a description of the way colour
video information is encoded and presented.
The disadvantages of NTSC
The NTSC colour television system is based on the original 525 line per
frame monochrome signal. Colour was added to the monochrome signal
without increasing the overall channel bandwidth by modulating the
colour signal onto a subcarrier, and hiding it within the upper harmonics
of the monochrome signal. A burst of subcarrier just before every line
ensured that the television receiver could lock to the subcarrier phase
and decode the colour properly.
After NTSC was introduced and went into common use it was found that
the transmitters suffered from non-linearity problems, i.e. the phase of
the chroma shifted as the level of the chroma signal changed. The
phase was locked to burst, so was correct at burst level. Any colour at
levels significantly different from burst came out wrong. Hence NTSC
television receivers have a Hue control to alter the colours and try to
make them look better, and NTSC gained the dubious title “never the
same colour”.
Hue controls never worked completely. You could correct one colour and
others would go wrong.
At the same time, it was always felt that it was a pity that the change
from monochrome NTSC to colour could not have also been used to
increase the number of lines per frame, and improve the vertical
resolution of the image.
Two solutions were introduced in later years to overcome the colour
problem, and increase the number of lines per frame. These were PAL
and SECAM. Both took a slightly different approach to the problem of
transmitter non-linearity, but both increased the vertical resolution of
NTSC in the same way.
The PAL solution
The PAL system was introduced later than the NTSC system and was
able to correct the main disadvantages of the NTSC system. It aimed to
eradicate the colour phase shifting problem by alternating the phase of
the (R-Y) part of the colour signal on each successive video line. The
phase switched from positive to negative, negative to positive, for each
new line.
The PAL receiver was able to use this alternating phase to detect if the
overall colour phase had shifted, at any level of chroma signal, and pull it
back to where it should be, resulting in true colour on home television
screens.
PAL also has 625 lines per frame. This improved the vertical resolution
of PAL video pictures giving a better picture.
Sony Training Services 64

73. Broadcast Fundamentals
The PAL signal
The PAL video line
There a several forms of PAL video line depending on its position in the
overall video frame. Most of these are the active video lines.
Each active PAL video line consists of a horizontal sync pulse, burst and
the video information for that line. The rest of the line is blanked.
The total line duration is 64uS. Blanking extends for 12us and the active
line for 52us.
Start of the line
The start of the line is defined as the half transition point of the leading
edge of the horizontal sync pulse. This is 1.5us from the end of the last
active video region, and 10.5us from the beginning of the next active
video region.
The frame
The PAL frame consists 625 video lines. 576 of these are active, leaving
49 lines for the vertical blanking interval.
The PAL frame is divided into 2 fields. Both fields have 312.5 lines each.
Field 1 has 287.5 active lines, field 2 has 288.5 active lines.
Vertical blanking parameters
Broad and equalisation pulses
PAL has 6 equalisation pulses, followed by 6 broad pulses, followed by 6
more equalisation pulses. All broad pulses and equalisation pulses have
a ½ line duty cycle, i.e. repeat every ½ line.
Start of the field and frame
The start of field 1 is defined as the start of the half transition point of the
first broad pulse, i.e. at the beginning of line 1, and for field 2, half way
through line 313.
The start of the frame is the same as the start of field 1.
Video blanking
PAL vertical video blanking extends from lines 311 to 336 between fields
1 and 2, and between lines 623.5 and 23.5 between fields 2 and 1.
The PAL chroma signal
The PAL component (R-Y) and (B-Y) signals are attenuated to reduce
their bandwidth, but they are not rematrixed to two signals at different
phase to (R-Y) and (B-Y) as they are in NTSC with the I and Q signals.
The (B-Y) signal is attenuated to a signal called U, and the (R-Y) signal
to a signal called V according to the equations :-
65 Sony Broadcast & Professional Europe

74. Part 8 – PAL television
U = 0.492 (B-Y)
V = 0.877 (R-Y)
The U signal is modulated onto the subcarrier and the V signal onto a
quatrature signal 90 degrees advanced from subcarrier.
+R ed -C y a n
103 100 77
9 5 IR E 9 5 IR E
90
-G re e n 80 + M a g e n ta
120 61
8 9 IR E 70 8 9 IR E
60
50
40
30
+ Y e llo w + B u rs t - B lu e
167 20 45 13
6 7 IR E 2 1 .5 IR E 6 7 IR E
10
0
180 (U c o m p o n e n t)
(B -Y )
- Y e llo w -B u rs t + B lu e
193 315 347
6 7 IR E 2 1 .5 IR E 6 7 IR E
+ G re e n
241
8 9 IR E
- M a g e n ta
300
8 9 IR E
-R e d +C yan
257 283
9 5 IR E 9 5 IR E
270
Figure 23 PAL vector space
V switching
The V component of the chroma signal is switched negative/positive
positive/negative every line. Line 1 of field 1 is positive. The receiver
uses this to determine difference in chroma phase at any level of
chroma.
Burst phase and swinging burst
Burst is chosen to be at 135 degrees. Burst also switches
negative/positive positive/negative every line in accordance with the V
component of the chroma signal. This is called the swinging burst, and
the television receiver uses this to determine if the V component of the
chroma signal is negative or positive on each video line.
Sony Training Services 66

75. Broadcast Fundamentals
PAL vector display
The PAL vector display has twice as many box targets and the NTSC
display. Half of these are similar to NTSC, the other half are negative
versions for those lines when the V component and burst and negative.
Thus, for instance positive red is at 103 degrees and negative red at 257
degrees. Positive cyan is at 283 degrees while negative cyan is at 77
degrees.
Choice of subcarrier frequency
As mentioned on page video is highly repetitive and the bandwidth
spectra of the monochrome and colour signals have energy centred
around the harmonics of line, field and frame rate. In NTSC, the colour
subcarrier is chosen so that the colour harmonics sit in the energy gaps
of the Y signal.
V switching component problem
However in PAL the V component switches every video line. This means
that the colour signal is similar, not every line, but every other line. The
energy of the PAL chroma signal is not based on harmonics of line rate
but of half line rate.
Placing the subcarrier exactly between two Y harmonics will mean half
the colour harmonics sit exactly on the Y harmonics, just what we are
trying to avoid!
So the original designers of PAL chose to place the subcarrier between
line harmonics 283 and 284, offset slightly from the centre between
these two Y harmonics. This is called a ¼ line offset and two colour
harmonics now sit between each pair of Y harmonics.
Dot crawl problem
It is happy chance that the NTSC subcarrier phase between subsequent
line is exactly opposite, and between subsequent fields is also exactly
opposite. This helps to cancel out any patterning effect due to subcarrier
itself.
In PAL, without the ¼ line offset, V component switching causes the
exact opposite to happen. The phase of each subsequent line, and field,
is the same, causing fine vertical stripes in the picture.
With the ¼ line offset this turns into a crawling dot pattern across the
image. Thus the designers of PAL added a further 25Hz to the
subcarrier to ‘spoil’ this dot pattering effect and make it much less
noticeable. This is called picture frequency shift because it is the same
as frame rate.
Subcarrier frequency calculation
Thus the final calculation for the PAL subcarrier calculation is :-
fsc = (( N – ¼ ) x L x fv ) + fv
67 Sony Broadcast & Professional Europe

76. Part 8 – PAL television
fsc = subcarrier freq.
N = chosen harmonic
L = lines per frame
Fv = frames per
second
= ( (284 – ¼) x 625 x 25 ) + 25
= ( 283.75 x 625 x 25 ) + 25
= 4433593.75 + 25
= 443361875 Hz
= 4.43361875 MHz
Bruch blanking
During the development of PAL it was found that the swinging burst
caused problems with some reference generators. If the same pattern of
vertical blanking is us in PAL as in NTSC, the first and last burst for each
field could have either a positive or negative V component.
Bruch blanking is a method of blanking the burst during the vertical
interval so that the first and burst of every field always has a positive V
component.
The Bruch blanking pattern extends over 4 fields. The table below shows
the first and last lines to have burst for all 8 fields.
Field First line with burst Last line with burst
1 6 310
2 320 622
3 7 309
4 319 621
5 6 310
6 320 622
7 7 310
8 319 621
Sony Training Services 68

77. 69
Figure 24
0 .7 V
Y
Broadcast Fundamentals
0 .3 V
0 .3 5 V
R -Y
&
B -Y
0 .3 5 V
F ro n t p o rc h B a c k p o rc h
(1 .5 5 u S ) L in e s y n c (5 .8 u S )
(4 .7 u S )
L in e b la n k in g A c t iv e lin e
(1 2 .0 5 u S ) (5 2 u S )
Component PAL video line timings
Sony Broadcast & Professional Europe
L in e 6 4 u S

78. Part 8 – PAL television
E n d o f fie ld 4 S ta r t o f fie ld 1
A c tiv e v id e o lin e s V e r tic a l b la n k in g in te r v a l A c tiv e v id e o lin e s
H a lf lin e H a lf lin e
5 e q u a lis a tio n p u ls e s 5 b r o a d p u ls e s 5 e q u a lis a tio n p u ls e s
620 621 622 623 624 625 1 2 3 4 5 6 7 8 22 23 24 25
E n d o f fie ld 1 S ta r t o f fie ld 2
308 309 310 3 11 312 313 314 315 316 317 318 319 320 321 335 336 337
H a lf lin e H a lf lin e
E n d o f f ie ld 2 S ta r t o f fie ld 3
620 621 622 623 624 625 1 2 3 4 5 6 7 8 22 23 24 25
H a lf lin e
E n d o f f ie ld 3 S ta r t o f fie ld 4
308 309 310 3 11 312 313 314 315 316 317 318 319 320 321 335 336 337
H a lf lin e H a lf lin e
E n d o f f ie ld 4 S ta r t o f fie ld 5
620 621 622 623 624 625 1 2 3 4 5 6 7 8 22 23 24 25
E n d o f fie ld 5 S ta r t o f fie ld 6
308 309 310 3 11 312 313 314 315 316 317 318 319 320 321 335 336 337
H a lf lin e H a lf lin e
E n d o f fie ld 6 S ta r t o f f ie ld 7
619 620 621 623 624 625 1 2 3 4 5 6 7 8 22 23 24 25
E n d o f fie ld 8 S ta r t o f fie ld 1
308 309 310 3 11 312 313 314 315 316 317 318 319 320 321 335 336 337
Figure 25
Sony Training Services 70

79. Broadcast Fundamentals
Different types of PAL
There are in fact 8 different types of PAL. The differences are small. Line
durations and subcarrier frequencies are different between the different
PAL types.
The ITU has given a letter designation to each of the PAL types. These
are ‘B’, ‘D’, ‘G’, ‘H’, ‘I’, ‘M’, ‘N’ and ‘Combination N’. Great Britain uses
PAL I.
The disadvantages of PAL
PAL is more complex than NTSC (which is more complex than
monochrome television).
Monochrome television has a 2 field repeating relationship, i.e. 2 fields
make one complete frame. NTSC television has a subcarrier to
horizontal sync (SC/H) relationship that repeats every 4 fields.
The SC/H relationship of PAL however is even more complicated and
results in an 8 field relationship. This has an impact on editing systems
and special effects. Good editing with PAL signals can only be done at
the correct 8 field editing point. Decoding is also more complex. It is
more difficult to separate the colour and monochrome components.
71 Sony Broadcast & Professional Europe

80. Part 9 – SECAMtelevision
Part 9 SECAM television
SECAM is another approach to solving the inherent problems of NTSC.
However SECAM is not popular in the studio. Although clever, SECAM
is very much more complex than PAL.
SECAM is not covered in great detail her because it is not used very
much within the studio.
SECAM is similar to PAL, with 625 lines per frame, with 312.5 lines per
field, and 50 fields per second. However that is where the similarity
ends.
SECAM transmits Y on every video line and each colour difference
signal on each successive video line, i.e. R-Y then B-Y then R-Y, as so
on. Thus all SECAM receivers have a line memory so the the colour
difference signal from one line can be used in the decoding of the next
line as well.
SECAM also used a form of low frequency preemphasis on the colour
difference signals.
Sony Training Services 72

81. Broadcast Fundamentals
The video camera
Types of video camera
The camera
Although all the cameras mentioned here may be described as video
cameras the true video camera is a unit that simply converts a moving
image into an electrical signal.
All surveillance and CCTV cameras tend to fall into this description.
Some professional and broadcast cameras also fall into this bracket as
well although the more professional ones tend to be dockable (see
below).
The video camcorder
The video camcorder differs from the camera in that it can record the
image it is looking at onto some storage medium it is holding within itself.
The most common camcorders today record to tape. An increasing
number of camcorders are using disk or solid-state technology instead.
Much of the design for the initial part of a camcorder is exactly the same
as it is for a camera. This is called the camcorder front end.
It is when we reach the signal processing part of the camcorder that
things start to look a little different.
The dockable camera
The obvious next step is to split the camera part of the camcorder from
the recorder part. This has been done to a number of broadcast and
professional camcorder designs, and the technology is called ‘dockable’.
The term ‘dockable’ is also used for some broadcast system cameras.
These cameras have no tape recorder section, they are purely a
camera. However the front end can be split from the back end, and form
part of a much larger system. Dockable system cameras are described
in the next section.
Dockable units allow you to ‘pick and mix’ the back and front halves of a
camera depending on the requirements of the shoot, technical reasons,
or on financial constraints.
System cameras
System cameras are used in television studios and outside broadcast
trucks. Their whole philosophy is that the camera forms part of a
complete environment that is operated and controlled by a team of
people, rather than just one person.
The beginning of the system is the lens. As with many professional and
broadcast cameras this will be removable, and will be chosen to match
the application for which it is being used.
The system camera itself is of the dockable variety. The front end has
the optical block and the circuitry for processing the signals from the
73 Sony Broadcast & Professional Europe

82. Part 10 – The video camera
optical sensors. The back end handles the conversion into a standard
video signal. This may be a standard analogue composite or component
connection, a digital connection, maybe compressed, or something a
little more professional like a triax connector. Indeed triax connectors
have a true system approach as they allow for very high quality output
from the camera, as well as allowing for power and control signals to the
camera, all in one cable.
Connections from the camera are sent into camera control units. These
units allow the camera to be controlled remotely as well as allowing for
adjustments like colour correction.
This whole approach is to allow the cameraman to simply frame the
shot. Someone else back in the control room will see feeds from all the
cameras and will be able to ensure that there is a good balance between
them. Yet another person can take care of colour, ensuring that whites
look white and skin tone looks correct, for instance.
Parts of a video camera
The lens
Every camera will use a lens to focus the image. In some cameras the
lens is fixed, i.e. it is not removable. In other cameras the lens can be
removed and replaced with another with different characteristics and/or
quality.
All removable lens used to be a screw fix. The screw fix lens is not as
popular now as it used to be and is generally only used on cameras
where it is unusual to change the lens.
The popular method of changing lenses on most modern cameras if the
bayonet or breach mount. Rather than screwing the lens into place
bayonet and breach lenses are removed and fitted by a simple twist
through about 90 degrees. The action is far quicker, and far more
positive than the screw fixing.
Lens electrical connections
Modern cameras generally require electrical connections between the
camera and the lens, for three possible controls, focus, zoom and
aperture. Some camera lenses may have controls for one, two, or all
three controls.
These electrical connections allow the camera operator to control the
lens from the camera grips, rather than reaching forward to the lens
itself. This helps the cameraman balance the camera and keep it steady.
Alternatively the electrical connections can be fed into the camera
electronics to allow for automatic iris control, or focus and zoom from a
remote camera control box.
In some cases the electrical connection is made through the bayonet
mount itself. This is useful because it is a good positive action, and does
not involve any cables.
In other cases a separate connection may have to be made after the
lens is fitted.
Sony Training Services 74

83. Broadcast Fundamentals
The sensor
Light from the lens passes into the camera itself and into a sensor.
There are various designs of sensor, but they all change the image into
an electrical signal.
Colour camera considerations
Colour cameras need to split the image into three primary colours. This
can be done using a specially designed colour sensor, or can be done
by first splitting the image into three separate images, one for each
primary colour, in a special piece of optics called a diachroic splitter
block. Each output from the block is sent to a separate sensor, and is
really just a normal monochrome sensor.
Diachroic blocks and multiple sensors add size, weight and cost to the
camera design, but produce a better image. Therefore cheaper colour
cameras, and small colour cameras, use colour sensors. Professional
and broadcast cameras use diachroic blocks and three, or maybe four,
normal sensors.
However there are signs of a radical change in colour camera design
allowing for high quality colour cameras with no diachroic block and only
one sensor. This is explained in the Parts on Image Sensors.
Signal processing
Signals from the image sensors are passed into the camera’s
electronics. This electronics buffers, amplifies and converts the signals
into a form that can be used outside the camera. This could be a
composite of component signal, digital or analogue, baseband or
compressed. The available outputs will depend of the camera’s
application, cost, and sometimes size.
The camera’s electronics also allow the signals to be modified by the
operator. Many cameras have controls for brightness, and maybe some
sort of colour control. Professional and broadcast cameras often have
complex controls for colour balance, white and black level adjustments,
and adjustments like latitude and knee controls.
Camcorder signal processing
Camcorders generally have similar signal processing as cameras, with
the same controls, and the same outputs. However camcorder signal
processing also turns the signal from the sensors into some kind of
signal that can be recorded onto the internal medium.
In the case of a tape this would be a serial signal with some form of
channel coding. Channel coding is where the signal is modified in some
way to allow it to be recorded on tape effectively and without loss. Digital
camcorders often also use some form of error correction.
Disk storage also requires its own type of channel coding and error
correction.
Solid state storage does not require and channel coding and may not
require any error correction.
75 Sony Broadcast & Professional Europe

84. Part 10 – The video camera
Video camera specifications
Resolution
Resolution is a measure of the resolving power of the camera.
All cameras, colour or monochrome, are the single sensor type. The
sensor pixels in colour cameras are divided between the three primary
colours. Thus, for the same sensor density, there is a difference in
resolution between monochrome and colour cameras. Monochrome
cameras will therefore tend to have a higher resolution than colour
cameras.
Still cameras often use the number of pixels in the sensor as a measure
of resolution. However it is not a common method of defining resolution
in video cameras. Sensor resolution will give a basic figure for the
sensor itself. In many cases only a proportion of the pixels are actually
used in the picture. If specifications mention ‘active pixels’ or ‘effective
pixels’ rather than simply ‘pixels’, this will give greater assurance that all
these pixels are part of the picture.
The camera’s circuitry will also affect the sensor’s resolution. Badly built
circuitry will have a poor bandwidth that will reduce the resolution
provided by the sensor by the time the signal reaches the output. Having
a good sensor and bad circuitry is a waste. CCTV camera resolution
figures should always be related to the final output signal.
Resolution figures are sometimes given as vertical resolution. This is the
number of active lines in the picture. All PAL based CCTV cameras are
built around the PAL television system with 625 lines per frame. Of this
576 lines are active. All PAL based CCTV cameras should be able to
achieve a vertical resolution of 576 lines.
Resolution figures are normally given as horizontal resolution. This is a
measure of the number of individual pixels per line the camera is able to
resolve, and is measured in vertical lines. Horizontal resolution can
never be higher than the sensor’s horizontal resolution, and is often
lower, due to bandwidth limitations of the circuitry.
Horizontal resolution and bandwidth are related by the equation :-
 1 
Bandwidth =  
 Period 
Each horizontal line lasts about 50uS long (exactly 52uS). The pixels, or
vertical lines, are divided up into this 50uS. The period is one clock
cycle, producing two vertical lines, one black, one white.
Therefore :-
50 × 10 −6
Period =
 Lines 
 
 2 
Sony Training Services 76

85. Broadcast Fundamentals
1 × 10 −4
=
Lines
Therefore the bandwidth can be found by combining these two
equations :-
 
 
 1 
Bandwidth =  −4 
  1 × 10  
 
  Lines  
 
= Lines × 10000
These equations boil down to a very simple rule. If the number of lines or
pixels is measured in hundreds, and the bandwidth in MHz, the two are
equal, i.e. 400 vertical lines = 4MHz bandwidth, 600 lines = 6MHz
bandwidth.
Bandwidth, probably more than any other parameter, is the figure that is
more difficult to achieve. Bandwidth costs money and separates the
good cameras from the bad ones. For square pixels the horizontal
resolution would need to be 768 vertical lines, or pixels, which gives
almost 8MHz bandwidth! No CCTV camera can achieve this. Cameras
achieving 600 vertical lines are considered good quality.
Sensitivity
Sensitivity is a measurement of how much signal the camera produces
for a certain amount of light.
Sensitivity can be measured as the minimum amount of light that will
give a recognisable picture, and is sometimes called ‘minimum
illumination’. Figures of below 10 lux should be possible for standard
CCTV cameras. However although this method provides an easy guide
to CCTV planners and installers, it is a highly subjective measurement.
What is a recognisable picture to one person may be unrecognisable to
another.
Professional and broadcast cameras use a different, more quantifiable
method for measuring sensitivity. The camera is pointed towards a
known light source. This is often a 2000 lux source at 3200K light
temperature (colour). The iris is then closed until the output is exactly
700mV.
Thus a reasonable sensitive camera may be f11 at 2000lux, whereas a
less sensitive camera may be f8 at 2000lux.
CCTV camera specifications are often not so consistent. Different lux
levels are specified. In the case of low light and night cameras normal
colour temperatures are meaningless because the camera is not
designed to be lit with standard 3200K light! These cameras often
77 Sony Broadcast & Professional Europe

86. Part 10 – The video camera
specify the minimum illumination sensitivity, and should quote figures
very much less than 1.
Dome camera manufacturers specify sensitivity with the dome removed,
because the figure is better than with it fitted, Some give figures with the
dome fitted as well. The camera would normally be used with the dome
fitted. This factor needs to be remembered. Dome cameras need to be
more sensitive than other cameras, if they are to overcome the losses
through the dome itself.
Signal to noise ratio (SNR)
A camera’s SNR is found by comparing the amount of video signal to the
amount of noise, in decibels, with the equation :-
 video 
20 log dB
 noise 
As a guide, an SNR of about 20dB is poor and is probably not viewable.
30dB will give a barely distinguishable image. 50dB is acceptable and
60dB good.
As a ratio of video signal to noise, 20dB is 10:1, and 60dB is 1000:1.
Gain
CCTV cameras with automatic gain control (AGC) add another
complication to the specifications. Manufacturers will quote sensitivity
figures with AGC switched on. However they will generally quote SNR
figures with the AGC switched off. The reasons for this are obvious. It
makes the figures look better!
Output formats
CCTV cameras use many different video output formats, from the simple
analogue composite output fitted to most cameras, through the analogue
Y-C output format, digital formats of one kind or another, and direct
computer network outputs used by some of the latest cameras.
Specifications always show the SNR, sensitivity, etc. from the best
output. The most common output connection people use is the analogue
composite output. Most cameras have it fitted and it is a simple
connection. However it is also the worst quality output.
Sony Training Services 78

87. Broadcast Fundamentals
Lenses
A lens is a transparent curved object capable of bending light. Most
lenses are made from glass. However any clear material will make a
lens. Different materials have different optical and physical
characteristics, some of which are better than those of glass.
A lens is based on some basic properties any transparent material have,
with respect to light. The most important is their ability to bend light.
Refraction
If a light ray passes from one transparent material to another it is bent
according to the relative refractive indices of the two materials.
Snell’s law
Snell’s law defines the behaviour of the light ray. It states :-
n1 sin i = n2 sin r
Where n1 is the refractive index of one material and n2 is the refractive
index of the other. i is the angle if incident (approach angle) and r is the
angle or refraction (leaving angle).
Every material has a different refractive index.
The refractive index of air is 1. Therefore the refractive index of any
transparent material can be found by rearranging the equation above,
thus :-
n2 = n1 sin i / sin r however n1 = 1 therefore
n2 = sin i / sin r
The coin in the tank of water
Figure 26 The coin in the tank
Light passing from water to air, or visa-versa, is bent because water has
a refractive index that is different to air.
Imagine a tank of water with a coin sitting at the bottom of it. The rays of
light coming from the coin pass upwards through the water and out into
79 Sony Broadcast & Professional Europe

88. Part 11 – Lenses
the air. As they pass from water to air they are bent by an angle relative
to the angle you are looking at from vertical.
Thus if you are looking at the coin at any other angle that from directly
above the coin itself it will appear to be in a different position that it
actually is.
The block of glass
The next step is to imagine a block or thick sheet of any clear material,
like glass.
Figure 27 Refraction through a block of glass
Light passing through the glass at an angle is bent as it passes from air
to glass and out again from glass to air. The angle the light ray bends as
it passes from air to glass is exactly the same but opposite to the angle
as it passes from glass to air.
The ray of light on the incoming side of the glass is parallel but displaced
to the outgoing side. This displacement can be found by the following
equation :-
d = t sin i (1-1/n)
Where d is the displacement, t is the thickness of the glass, i is the
incident angle and n is the refractive index of the glass.
Notice that the refracted r angle is not part of the equation. The entrance
and exit rays are parallel and therefore irrelevant.
The prism
A prism is a little like the block of glass we have just looked at where the
two sides of the glass are not parallel with one another.
The most common prism is a block of glass with a triangular section.
The sides of the triangular section can be at any angle to one another
Sony Training Services 80

89. Broadcast Fundamentals
although most prisms have something approaching an equalateral
triangular section.
Light bending properties of a prism
Light entering one side of the prism is bent and leaves the prism at a
different direction. The angle of bend is called the deviation and can be
found by the equation :-
D = A (n-1)
Where D is the deviation, A is the angle of the prism, and n is the
refractive index of the glass (or whatever material the prism is made of).
Figure 28 Refraction through a prism
Colour splitting properties of a prism
White light is made up of many different colours. Each colour has a
different wavelength.
Figure 29 Splitting light through a prism
When a ray of white light passes from one transparent material to
another the different wavelengths are refracted by different angles. This
has the effect of splitting the white light into its constituent colours.
The convex lens
The convex lens is a little like a series of prisms placed next to each
other, all with slightly different angles between their two side.
81 Sony Broadcast & Professional Europe

90. Part 11 – Lenses
Figure 30 The convex lens as prisms
In fact if you increase the number of prisms, making them smaller and
smaller you will eventually have a perfect convex lens.
Convex lenses have a focal point. If parallel light enters one side of the
lens it is focused to a single point. This is the basis for all lens designs. If
the lens did this perfectly all lens designs would be just one convex
element. However, as we will see the convex lens is not perfect, and
certain things have to be done to eliminate these imperfections.
Figure 31 The convex lens
The sides of most convex lenses are made from part of a sphere. While
this is easy to produce, and perfectly good enough for most lenses, it
can present problems for certain lenses.
Figure 32 The convex lens as part of two spheres
Sony Training Services 82

91. Broadcast Fundamentals
The concave lens
The concave lens is effectively the opposite of the convex lens. In the
same way it can be seen as an infinite arrangement of prisms and its
two sides are based on parts of a sphere.
Figure 33 The convex lens as prisms
Figure 34 The concave lens focus
Concave lenses have a focal point. However the concave lens focal
point is a ‘virtual’ focal point on the approach side the lens. The focal
point has no practical purpose in the same way as it does with convex
lenses, but is used mathematically to calculate the properties of lens
designs that use concave lens elements.
Figure 35 The concave lens as spheres
83 Sony Broadcast & Professional Europe

92. Part 11 – Lenses
Chromatic aberration
There are two types of chromatic aberration, axial (sometimes called
longitudinal) chromatic aberration and lateral (sometimes called
transversal) chromatic aberration.
Axial chromatic aberration
The basic prism showed how it is possible to split white light into its
constituent colours. Any refractive surface will bend different coloured
light by different degrees. This effect is called dispersion.
Figure 36 Axial chromatic abberation
A lens is really an infinite number of small prisms laid out in a particular
way. Therefore it stand to reason that a lens will split white light into its
constituent colours.
This effect is called axial, or longitudinal, chromatic aberration and
presents problems for lens designers.
Looking at a basic convex lens when parallel rays of white light enter at
one side of the lens it is split into its constituent colours, with each colour
having a different focal point depending on its wavelength. Shorter
wavelength colours, at the ultra-violet end of the spectrum, are refracted
more and have a shorter focal point.
Correcting axial chromatic aberration
In order to correct axial chromatic aberration cause by one lens element
you need to add another lens element with the opposite error.
The required overall effect of most lens designs is to produce a perfect
convex lens. However a perfect convex lens does not exist. By sticking a
concave lens onto the convex lens you can eliminate the axial chromatic
aberration of the basic convex lens.
This is why lens designers stick lens element together.
However simply sticking a concave lens with the opposite effect to a
convex lens also eliminates the focal effect and the lens ends up looking
like a flat piece of glass!
The trick is to use a convex lens and concave lens with different
refractive indices. Thus although chromatic aberration is eliminated. The
two lenses together still focus to a point.
Sony Training Services 84

93. Broadcast Fundamentals
Figure 37 Lens doublet
This design is called an achromatic doublet.
Figure 38 Various lens doublet designs
Achromatic doublets come in all different forms, depending on the
particular use for which they are intended. The important thing is that the
convex element is fatter in the middle that at the edge and the concave
lens is thinner in the middle than at the edge.
Lateral chromatic aberration
Lateral chromatic aberration is a less obvious problem than axial
chromatic aberration. It arises from the same limitation of lens elements
but effect the image laterally. It causes fringing near the outer edge of
images, where the different colours have been split apart.
Lateral chromatic errors affect lens with very long or very short focal
lengths, i.e. long telephoto lenses and fish-eye lenses.
Correcting for lateral chromatic aberration
Lateral chromatic aberration in telephoto lens designs can be reduced
by not using refractive elements in the design. Mirror lenses use curved
mirrors instead of lenses. Thus no refraction and not dispersion.
The other method is to use low dispersion material such as fluorite.
However this material is difficult and expensive to work, and is affected
by normal air. Flourite lens elements can only therefore be used as an
internal element where they can be protected by a normal glass
element.
85 Sony Broadcast & Professional Europe

94. Part 11 – Lenses
Spherical aberration
As shown before the two sides of most lenses are designed as parts of a
sphere. This makes manufacture easy and is perfectly good in most
cases.
However making lens sides as part of a sphere is not actually correct.
The focal point at the edge of the lens is actually at a different point
compared to the middle.
Most lens elements are small enough for this not to be a problem.
However lens designs with large elements, such as some television
camera lenses, lenses intended for dim lighting conditions and some
wide angle lenses, can suffer from spherical aberration.
One answer is to use lens doublets or triplets where the spherical
aberration of one element is eliminated by another.
Aspheric elements
Another answer is to use lenses where the sides are not part of a
sphere. The perfect lens is slightly flatter at the edge than in the middle,
making the refractive power of the lens greater nearer the middle of the
lens.
These so-called aspherical lenses are difficult to produce, especially in
quality, making lens designs with good spherical aberration
characteristics more expensive.
Figure 39 The aspherical lens
Coma
Coma is a distortion effect that shows up as a fuzziness at the edge of
the image. It is caused by spherical aberration but shows itself in rays of
light passing through the lens at a sharp angle.
Properties of the lens
The principal element
Although lenses are made up from a collection or convex and concave
lens elements, doublets and triplets, and even mirror elements, they can
all be thought of as a single perfect convex lens element. Lenses use
Sony Training Services 86

95. Broadcast Fundamentals
many different elements to correct aberrations, reduce size and allow
the lens to be controlled.
The theoretical single perfect convex lens element, is referred to as the
principal element. Its position is called the principal point.
Focal point
The focal point is where light from infinity (i.e. parallel light) is brought to
a signal point. Convex lenses and concave mirrors have a real focal
point. Concave lenses and convex mirrors have a virtual focal point.
Focal length
The focal length can be found from the formula:
1/f = 1/u + 1/v where f = focal length
u = object distance
v = image distance
This can be simplified when the object is at infinity. The equation
becomes f=v. So a simple definition of focal length is defined as the
length from the principal element to the focal point.
The focal length has an effect on the field of view of a lens, and its
magnification. A lens with a short focal length has a wide field of view,
and low magnification, and is called a wide-angle lens. A lens with a long
focal length has a small field of view, and high magnification, and is
called a telephoto lens.
Aperture
The aperture of a lens is a way of expressing the amount of light passing
through it. Its maximum value is limited by the lens’s pupil, or iris.
Aperture is controlled by a multi-bladed iris mechanisn. The larger the
aperture the more light is passed through the lens. The aperture is
normally expressed as the f number.
This f-number is expressed mathematically as:
N = f/d where f = focal length,
d = diameter of the entrance pupil
For an aperture of 2 this would normally be written as f/2.
The larger the aperture the smaller the f-number. When the f-number
doubles the light passing through the lens is reduced by a factor of 4.
The markings on a lens are therefore normally indicated as ratios of 1.4
i.e. 1.5, 2, 2.8, 4, 5.6, 8 and so on. Each step or ‘f-stop’ represents a
halving of the light.
Depth of field
The depth of field is the range of object distances for which the image is
within a permissible degree of sharpness. Only objects at the focal point
are perfectly in focus, objects closer and further away are slightly out of
focus. The depth of field is therefore not an absolute figure, but is
derived from the concept of a circle of confusion.
87 Sony Broadcast & Professional Europe

96. Part 11 – Lenses
In c o m in g lig h t
Lens
Focal
p o in t
C ones of
c o n fu s io n
Ir is
Figure 40 Depth of field
Depth of field is dependent upon focal length and the aperture of the
lens. A long focal length (telephoto) lens has a small depth of field. The
smaller the aperture (bigger the f-number) the larger the depth of field.
Lens
D e p th o f fie ld
Focal
p o in t A c c e p ta b le
c o n fu s io n
Ir is a t la r g e a p e r tu r e
Lens
D e p th o f fie ld
Focal
p o in t A c c e p ta b le
c o n fu s io n
Ir is a t s m a ll a p e r tu r e
Figure 41 Change in depth of field with aperture
The concave and convex mirrors
Concave and convex mirrors are used a lot in optics. They provide an
alternative to lenses without all the disadvantages associated with light
as it refracts through glass (or whatever the lens is made from).
Sony Training Services 88

97. Broadcast Fundamentals
Mirrors have the opposite effect on incoming light that lenses have.
Concave lenses disperse incoming light. Concave mirrors focus light to a
point. Convex lenses focus light to a point. Convex mirrors disperse
light.
Mirrors are very useful for long lenses. They allow a lens design to be
‘folded’, reducing the overall length of the lens.
Lens types
Normal lens
A normal lens is one which produces an image which is equivalent to the
image from the human eye. This is a little subjective as the image seen
by the human eye is greatly distorted, however the normal lens aims to
produce an image with nearly the same level of magnification, image
distortion and perspective at the center of view of the human eye.
As a rough guide the focal length of the normal lens is approximately the
same distance as the image diagonal. For a 35mm camera a normal
lens is one with a focal length of 50mm.
Normal lenses are also able to attain a lower aperture f number, partly
because the optics are better at this focal length but also because the
mathematics for calculating the aperture f number are dependant on the
focal length and are more favourable for the normal lens.
The telephoto lens
The telephoto lens is a lens with a focal length greater than the normal
lens, although the term “telephoto” is normally reserved for lenses wit a
focal length greater than twice that of the normal lens. For an image size
of 35mm, a telephoto lens is considered to be one with a focal length of
more than 100mm.
Telephoto lenses can magnify objects from a long distance. They have
minimal image distortion, and compress perspective.
Wide-angle lens
A wide-angle lens is one with a focal length smaller than the normal
lens. It has a wide field of view permitting a wide vista to be captured.
For a 35mm camera a wide-angle lens is one with a focal length of
smaller than 50mm although the term “wide angle” is normally reserved
for lenses with a focal length smaller than about 30mm.
Wide angle lenses make object appear unnaturally small, they distort the
image and stretch perspective.
The fisheye lens
A fisheye lens is an extreme wide-angle lens. As the focal length
becomes shorter it becomes increasingly difficult to maintain a
geometrically correct image, i.e. one that is square. When this design
target is abandoned the image becomes curved at the edges but a very
wide view becomes possible. At the extreme it is possible to have a fish
89 Sony Broadcast & Professional Europe

98. Part 11 – Lenses
eye lens with totally circular image, and an angle of view of more than
180 degrees.
The zoom lens
A zoom lens is a lens with a variable focal length. For a 35mm camera it
is common to use a lens with a 30-100mm zoom lens. It can therefore
change the field of view from wide angle to telephoto. For television
camera lenses it is common to have a zoom lens with a focal length
range of 8mm to 150mm. For specialist applications, such as sports
there are lenses that have zoom ratios of more than 40:1.
Prime lens
A prime lens is any lens that is not a zoom lens, i.e. any fixed focal
length lens, and are so called because of their superior quality. While
zoom lenses are very versatile, and their quality has reached remarkable
levels in the last decade or so, they are still a compromise. The best
quality prime lenses are always better quality than the best quality zoom
lenses.
Mirror lenses
Mirror lenses use a combination of normal lenses and mirrors. Telephoto
lenses are sometimes mirror lenses. Mirrors allow for compact designs
for telephoto lenses with very long focal lengths.
A characteristic of mirror lenses is that anything out of focus appears as
a donut shape, rather than a simple blur.
Extenders and adaptors
There are various extenders and adaptors that can be fitted between the
lens and the camera. These can be used to allow lenses with one
mounting scheme to be fitted to a camera with another mounting
scheme. They can also be used to alter the characteristics of the lens.
Mount adaptors
Mount adaptors allow lenses intended for one mount to be fitted to a
camera with another mount. Mount adaptors are popular in still cameras
where there are a lot of different mounts. Optics tend to suffer because
the lens is pushed away from the camera and the back flange to film
distance is not optimal.
All camera manufacturers have mechanical and electrical connections
between the lens and camera to allow the camera to control the lens.
These connections are very specific to the manufacturer. Adaptors
cannot guarantee to provide a match for these connections between the
lens and camera.
Some lens manufacturers offer lenses with no specific mount. These
lenses are designed slightly shorter than they should be. You select
which mount you want and the appropriate adaptor is fitted, building the
lens up to the correct length. Mechanical and electrical connections are
much more likely to work with this kind of mount adaptor.
Sony Training Services 90

99. Broadcast Fundamentals
2x, 3x etc. adaptors
This kind of adaptor increases the focal length of the lens. The simplest
of these is little more than a tube pushing the lens away from the camera
and boosting the focal length as a result. The better ones have lens
elements in them to improve the optics. No matter what the adaptor,
they are always a compromise. Fitting a 2x adaptor to a 25mm lens will
never attain the quality of a 50mm lens. However they provide a way of
effectively doubling the number of lenses you have, with only a marginal
reduction in quality.
Mechanical and electrical quality can vary just as with the optical quality.
Some adaptors are able to transfer the mechanical and electrical
connections between the lens and camera better than others.
Filters
Filters are sometimes used to correct something in the picture, to protect
the camera from damage, or to add some kind of special effect. A filter
can be placed in front of the lens or built in behind the lens
Built-in filters
Filters placed behind the lens are always built-in because they would
otherwise push the lens away from the camera and alter its optical
characteristics. Two types exist, camera built-in, and lens built-in.
Lens built-in filter are used if a filter cannot be put in front of the lens.
This is particularly true of ultra-wide angle and fish-eye lenses, because
the front lens element tends to protrude from the front of the lens. A slot
somewhere at the back of the lens allows glass or gelatin filters to be
slotted into the lens.
Camera built-in filters are common in video cameras. There are often
about 5 of these filters built into a wheel. By turning a small knob on the
camera the camera operator can turn the wheel and bring different filters
between the lens and the sensor. Neutral density filter are used to cut
the amount of light in bright conditions. Yellow tinted filters are used to
correct the colour temperature for daylight operation.
Front filters
There are a myriad of different filters that can be fitted in front of the
lens. Professional still camera users can screw filters directly to the front
of the lens. These screw-in filters are specifically designed for the
diameter of the front of the lens.
If a screw-in filter cannot be found an adaptor can be fitted to the front of
the lens. Once fitted, this allows a wide range of standard square filter
sheets to by placed in front of the lens, removing the need to find a
screw-in filter of the correct diameter.
This method is popular in movie cameras and video cameras. The
adaptor is generally called a matt box.
91 Sony Broadcast & Professional Europe

100. Part 12 – Early image sensors
Part 10 Early image sensors
Selenium detectors
Selenium was the first photoelectric material to be found, in 1873. It was
used in the first mechanical experiments on television, like the Nipkow
disk system.
Selenium is classed as a photoconductive material because its
resistance changes when exposed to light.
Photovoltaic materials produce a voltage potential across themselves
under the influence of light.
The Ionoscope
The Ionoscope was the first image sensor of any commercial
importance. It consisted of an evacuated glass enclosure with a tube
fixed to it enclosing an electron gun.
The main enclosure had a screen made from a sandwich of
photosensitive particles, called a mosaic, a thin mica insulation layer and
a conductive sheet backing. The plate acted like a capacitor.
Figure 42 The Ionoscope
Light from the lens could enter the enclosure through a window and land
on the mosaic releasing electrons which were attracted away towards
the anode. Thus a positive charge image built up on the surface of the
mosaic. The charge was proportional to the intensity of light.
The electron gun fired electrons in a raster scan at the mosaic. Any
positive charge was cancelled by absorption of electrons from the beam.
This absorption was detected by the conductive plate and output as a
signal at the signal electrode.
The rest of the electrons bounced off the mosaic to be picked up by the
anode and drawn away to the anode electrode.
Sony Training Services 92

101. Broadcast Fundamentals
The Orthicon tube
The Orthicon tube was invented by Iams and Roase at RCA in 1939.
The Ionoscope used a high velocity electron beam which gave rise to
secondary emission of electrons which effectively reduced the tube’s
ability to catch all the electrons released purely by photoemission.
The Orthicon tube had a much reduced anode voltage. This tended to
make the mosaic saturate with electrons when no light was present. Any
more electrons would not strike the surface. Thus no signal appears
when there is no light. This produces better Black recognition.
Any electrons not striking the surface of the mosaic return back down
the same path as the electron beam to soak away in a collector next to
the electron gun.
The low anode voltage and the resulting low velocity of the electron
beam means that the bean was subject to interference by stray electric
fields near the mosaic. The resulted in a loss in resolution compared to
the Ionoscope.
The beam focusing and deflection was better than for the Ionoscope. A
long focus coil was used, and either electrostatic or electromagnetic
deflection. The beam retained its helical nature. This aided focussing.
The beam was also deflected such that it always struck the target at a
perpendicular angle.
Figure 43 The image orthicon tube
The Image Orthicon tube
The Image Orthicon was an improvement over the Orthicon. In this
design light was focussed onto a photocathode plate. The released
electrons which were attracted by an accelerating grid back into the tube
towards a two sided glass target plate. Thus an image in electrons was
formed on the target.
93 Sony Broadcast & Professional Europe

102. Part 12 – Early image sensors
A thin mesh was placed in front of the target. The electrons from the
photocathode penetrated straight through the mesh and onto the target.
However any secondary electron emission was soaked up by the mesh.
The electron bean was a similar low velocity perpendicular design as the
Orthicon. It scanned a raster image on the back of the target. Any
electrons not being soaked up by the target were returned back down
the beam to be collected by the anode, next to the electron gun.
Thus the return beam was a raster scan of the charge, and thus of the
image.
The Vidicon tube
Figure 44 The Vidicon tube
The Vidocon tube was introduced by RCA in 1950. It used antimony
trisulphide target. This is a photo conductive material. The resistance
across the target changes when it is exposed to light. With certain limits
the change of resistance is proportional to the intensity of light.
Figure 45 The vidicon tube
The back of the target is scanned by an electron gun with the same
basic design features as the low velocity perpendicular design used in
the Orthicon tube.
Sony Training Services 94

103. Broadcast Fundamentals
The target is biased to the anode voltage. As the bean strikes the back
of the target current flow to the front is inversely proportional to the
resistance, which is inversely proportional to the light intensity.
Thus the anode bias voltage will alter as a raster scan of the image.
Variations on the Vidicon design
There were various improvements on the basic Vidicon tube design. The
Plumbicon was introduced by Philips in 1962. It used lead oxide as a
target material.
The Saticon was another design with a target made from arsenic,
selenium and tellurium.
The Diode Gun Plumbicon used a photo diode instead of an electron
gun
These later designs offered better resolution, greater contrast and better
colour balance than the basic Vidicon
.
95 Sony Broadcast & Professional Europe

104. Part 13 – Dichroic blocks
Part 11 Dichroic blocks
The purpose of a dichroic block
The purpose of a dichroic block is to split an incoming colour image into
its three primary colours. Most of the block is coated in black paint to
stop light getting in, except for a window to let the incoming colour image
in and three windows to let the outgoing primary images out.
They are fitted just behind the lens of a colour video camera. A sensor is
placed on each outgoing window, one for each primary. Each sensor
measures the brightness of each primary and sends out a video signal
for each primary.
Mirrors and filters
Various designs have been created over the years. Most designs are
now beginning to look very similar. Two basic design pattern are now in
use. The first is generally simply called a prism block or dichroic block.
The other is called a cross block or X block.
Conventional dichroic blocks
The conventional prism block consists of at least three prisms, glued
together with a transparent epoxy cement. Light enters the incoming
window at the front and passes through the first prism. The back surface
is angled, and is coated with a red dichroic mirror. The choice of
material, and its thickness define the colour that is reflected.
Manufactures can ‘tune’ the dichroic mirror by altering the coating
thickness. The red light is reflected once more off the front of the first
prism and out through the red outgoing window. There is a red filter to
trim the light before it strikes the red sensor.
R e d tr im filte r
R sensor
R e d lig h t B lu e d ic h r o ic m ir r o r
G r e e n lig h t
In c o m in g G r e e n tr im filte r
w h ite lig h t
G sensor
Lens
R e d d ic h r o ic m ir r o r B lu e lig h t B lu e tr im filte r
B sensor
C y a n lig h t
Figure 46 Dichroic block
Sony Training Services 96

105. Broadcast Fundamentals
The cyan light passes through the second prism. The back of this prism
is coated with a blue dichroic mirror that reflects blue light, letting
everything else through (green). The blue light passes out through the
blue outgoing window, through a blue trim filter and onto the blue
sensor. Likewise the remaining green light passes out through the green
outgoing window, through a green trim filter and onto the green sensor.
It is worth noting that the sensors are basically the same device. Some
manufacturers may carefully select sensors that have the best
performance for each colour, most will not.
Cross dichroic block
The cross dichroic block consists of four small triangular prisms, glued
together to make a small cube with two intersecting planes. Some faces
of each prism are coated with dichroic mirrors, or trim filters.
G sensor G r e e n lig h t
G r e e n t r im f ilt e r
R e d d ic h r o ic m ir r o r B lu e d ic h r o ic m ir r o r
R e d lig h t
B lu e lig h t
B sensor
R sensor
B lu e t r im f ilt e r R e d t r im f ilte r
Lens
I n c o m in g
w h it e lig h t
Figure 47 The cross dichroic block
Light enters the front of the block. One of the intersecting planes is a
blue dichroic mirror, the other a red dichroic mirror. Blue light reflects off
the blue dichroic mirror and out from the left side through a blue trim
filter. Red light reflects off the red dichroic mirror, passing out the right
side through a red trim filter. The remaining light is green, and passed
out the back of the block through a green trim filter.
97 Sony Broadcast & Professional Europe

106. Part 13 – Dichroic blocks
The cross dichroic block has become popular recently because of its
compact and simple design. However it has one major drawback. The
intersection between the four prisms causes a small vertical line on the
outputs. Although the light is out of focus as it passes the intersection,
and careful manufacture can make this intersection as tight as possible,
this is the main reason the cross block is not used on professional and
broadcast video cameras.
Optical requirements of a dichroic block
Every optical path, from the incoming window, to each outgoing window,
is identical. This is essential, because the lens will focus through the
dichroic block and onto the surface of the sensors behind. If one of the
optical paths is different, that particular primary colour will be out of
focus.
The position of the sensors is critical. All three sensors must be mounted
in exactly the same place relative to its own window. If there is any error,
that particular primary colour image will be in a different position to the
other two. Recombining the three primary images on the monitor will be
practically impossible.
Variation on a theme
Most dichroic block designs are now tending to look similar to the two
designs mentioned above. Some designs vary slightly.
Some designs swap the position of the red and blue dichroic mirrors.
Some designs have slight variations in the angles of the prisms and the
paths each primary colour will take.
Most designs have blue and red dichroic mirrors. Both of these mirrors
are relatively easy to make, because both have one cut off wavelength.
Green dichroic mirrors are more difficult to make because they have two
cut of wavelength designers have to worry about.
Some blocks have a forth or fifth outgoing window. This may be used for
a monochrome viewfinder output for the camera operator, or for some of
the camera’s internal functionality, like auto focusing or metering.
Some specialist video cameras do not have standard primary colours at
the outgoing windows. Security cameras may use infra-red for night
vision. Video cameras used in food processing and monitoring also use
infra-red to check the quality of food. These cameras may have one
window in the dichroic block dedicated to infra-red.
Using dichroic blocks in projectors
The increased popularity of low cost video projectors has lead to an
explosion in the need for cheap compact dichroic blocks. Quality is not
so much of an issue with projectors, and the cross dichroic block
therefore a very popular.
Dichroic blocks are used the opposite way round from video cameras.
Simple filters split light from the lamp into three primary beams. These
Sony Training Services 98

107. Broadcast Fundamentals
three beams are passed into the dichroic blocks through light valves,
where the sensors would be in a video camera. The light valves build up
an image for each primary by shutting light on or off for each pixel.
The dichroic block then combines the three primary images into one
colour image that is projected out to the screen.
99 Sony Broadcast & Professional Europe

108. Part 14 – CCD sensors
Part 12 CCD sensors
Advantages of CCD image sensors
When looking at the advantages of CCD image sensors, you have to
realise what alternatives there are and what was used before these
devices became available.
Before CCD image sensors became popular video and television
cameras used some form of tube sensor. Plumbicon tubes were very
popular for a while.
Bearing these devices in mind, let us consider the advantages of CCD
image sensors.
Compact design
The first and most obvious advantage of CCD image sensors are that
they are considerably smaller that tube sensors. They allow very
compact cameras to be made which can be used in discreet surveillance
and remote investigation in dangerous or confined places.
Light design
CCD image sensors are considerable lighter than tube sensors. They
can weigh only a few ounces. This allows them to be designed into
portable cameras without increasing the overall weight of the camera by
any undue amount.
High shock resistance
CCD image sensors have no moving parts. They also have a very light
duty mechanical construction that is highly resistant to acceleration and
deceleration damage.
Low power consumption
CCD image sensors use a lot less power than older tube sensors. This
makes them suitable for any battery powered device.
Good linearity
Linearity is important is measuring light levels accurately. Linearity
means that the output signal is proportional to the number of photons of
light entering the device.
Film and tube sensors are highly non-linear, partly because of their low
dynamic range. They give no output at all if the light level (number of
photons) is too low, and saturate if the light level is too high giving no
further output if the light intensity increases further.
CCD image sensors have good dynamic range, and good linearity over
this range.
Sony Training Services 100

109. Broadcast Fundamentals
Good dynamic range
CCD image sensors saturate in the same way as any light sensor, but
the light intensity required to saturate these devices is generally much
higher.
CCD image sensors have no effective minimum. Some specialised
devices can measure near total dark.
Typically photographic film has a dynamic range of about 100. CCD
image sensors achieve about 10,000.
High QE (quantum efficiency)
QE is the ratio of the number of photons of light detected to the number
of photons that enter the device.
Photographic film has a QE of about 5% to 20%. CCD image sensors
have a QE of between 50% and 90%. This makes them very efficient
and thus very useful for dark environment monitoring and studies of
deep space.
Low noise
In fact CCD image sensors can suffer from thermal thermal noise.
However this noise if predictable and can be controlled or reduced.
Cooling the image sensor using conventional cooling find and a fan can
keep thermal noise to very low levels.
For specialist scientific imaging Peltier effect heat pump and cooling by
liquid nitrogen can reduce thermal noise to virtually zero.
The basics of a CCD
A charge coupled device (CCD) is sometimes referred to as a bucket
brigade line. It consists of a series of cells. Each cell can store an
electric charge. The charge can then be transferred from one cell to the
next.
The line of buckets
A very good way of thinking of a CCD is to imagine a line of buckets. At
one end is a set of digital scales where you can measure the amount of
water you pour into the first bucket.
Figure 48 Single line of buckets
As soon as you pour the water from the digital scales into the first bucket
you will transfer the water from all the other buckets into the next bucket
101 Sony Broadcast & Professional Europe

110. Part 14 – CCD sensors
down the line. The water from the last bucket will be poured into the
digital scales at the other end, and measured.
Of course it would be impossible to move the water from one bucket to
the next at the same time. You would probably need another set of
buckets to store the water while you were transferring it.
Figure 49 The double line of buckets
The electronic reality
CCD’s use a line of metal oxide semiconductor (MOS) elements,
constructed on the same chip. Each element contains 2, 3 or 4
polysilicon regions sitting on top of a thin layer of silicon oxide.
Polysilicon can be used as a charge holder or a conductor. Although it is
not as good a conductor as other metals like copper or aluminium it is
easy to fabricate and is transparent, which is useful when CCD’s are
used in cameras.
Silicon oxide (glass) is a good insulator.
These elements are fabricated on a p type doped silicon subtrate.
At each end of this line is a region of n type doped silicon.
Connections are made to all the polysilicon regions and to the two n type
doped regions.
Using the CCD as a delay line
CCD’s have been very popular as a semiconductor delay line. They
were used in many electronic designs before semiconductor memory
became cheap and complex enough to be used instead.
Figure 50 The CCD delay line
CCD delay lines are essentially analogue. That is to say the charge they
carry is an analogue quantity. If a CCD delay line is to be used in a
Sony Training Services 102

111. Broadcast Fundamentals
digital environment there must be a digital to analogue converter fitted to
the input and an analogue to digital converter fitted to the output.
The transfer of charge is however digital, and the CCD will have a clock
input which is used to transfer the charge from one MOS element to the
next in the line.
How does the CCD delay line work?
The input signal is fed into the first polysilicon region. Using field effect
principles electrons are pulled from the n type doped region and collect
under the insulation layer.
The potential on the first region creates a potential ‘well’ that the
electrons effectively fall into.
Although there is a maximum charge that can be held in this potential
well, the amount of charge is proportional to the amount of time and the
potential applied to the first region.
The potential between the first and second regions are switched. This
effectively moves the potential well from just underneath the first region
to just underneath the second region. The first region becomes a
potential barrier. The charge is attracted to the second region.
The potential between the second and third regions are switched. The
charge is now attracted to the third region.
By switching the potential from one region to the next the charge can be
transferred from one region to the next, sitting just underneath the
insulation layer.
This leaves the first region clear. And the next charge packet can be
input to the line.
When the charge reaches the last region it transfers to the n type doped
region at the other end of the line and appears as a output signal.
2 region elements
CCD delay lines with 2 polysilicon regions per MOS element use the
second region in each element in the same way you might use the spare
buckets in the line of buckets.
The charge is transferred to the second region before being transferred
to the first region of the next element.
The disadvantage of the 2 region element is that the charge could flow
the wrong way. 2 region elements employ special gates in the element
and use a stepping transfer voltage to ensure the charge flows
correctly. This all adds to the complexity and cost of this type of CCD.
However 2 region elements offer higher density that 3 or 4 region
elements.
103 Sony Broadcast & Professional Europe

112. Part 14 – CCD sensors
Figure 51 The 2 region element CCD delay line
3 region elements
CCD delay lines with 3 polysilicon regions per MOS element are able to
ensure that the charge flows in the correct direction from one element to
the next.
Figure 52 The 3 region element CCD delay line
The charge is pulled from the left region to the centre region, then from
the centre region to the right region.
Sony Training Services 104

113. Broadcast Fundamentals
However clock phasing is more complex than both the 2 and 4 region
elements.
4 region elements
4 region elements have simpler clocking signal arrangements than 3
region designs and have better charge transfer capabilities, but it is
more difficult to achieve high density devices.
Figure 53 The 4 region element CCD delay line
105 Sony Broadcast & Professional Europe

114. Part 14 – CCD sensors
Using CCD’s as image sensors
The basic principles
Metal oxide semiconductors are sensitive to light. If light enters the
substrate of a MOS device, under certain conditions it excites electrons
in the silicon into the conduction band. Put simply, electrons are shaken
loose by light.
The elements used in image sensors are similar to those used on CCD
delay lines. They can be 2, 3 or 4 region elements.
Sensing light
When used as an image sensor a positive voltage is applied to the first
polysilicon region in each element. This develops a small potential well
just under the insulation layer.
Step A – Exposure
As light penetrates the p type substrate of the CCD it shakes electrons
loose. The loose electrons in the vicinity of the potential well fall in and
are trapped, forming a small collected charge.
The stronger the light level falling on that element, or the longer the time
allowed, the greater the number of loose electrons, and the greater the
stored charge.
Figure 54 The CCD delay line as an image sensor
Step B, C & D – Transfer
When the CCD sensor has been exposed to the image for the required
time the charges stored under each element have to be transfered to the
end of the row where they can be sensed and output.
In Step B the potential of region 2 is raised. Now the potential well
extends over two regions and the charge spreads to fill the space.
Sony Training Services 106

115. Broadcast Fundamentals
In Step C the potential of region 1 is lowered and region 3 is raised. The
potential well now occupies region 2 and 3. The charge is pulled across
so that it sits under regions 2 and 3.
In Step D the potential of region 2 is lowered and the potential of region
1 is raised. The potential well now occupies region 3 and 1 of the next
element. The charge is pulled across so that it sits under regions 3 and
1.
Steps B, C and D are repeated until the whole row has been transferred,
element by element to the output gate at the end of the row.
When this has been done and the whole row is empty of charge,
exposure can begin again.
The arrangement of MOS elements
CCD’s used as image sensors comprise a matrix of MOS elements. The
elements are laid out in columns. Each column is similar to a CCD delay
lines. There are many columns in the matrix.
The number of elements in each column and the number of column
defines the overall resolution of the device. Each element corresponds
to a single captured point from the image, otherwise called a picture
element or pixel.A system of channel stops is used to guard one column
from the next. These prevent charge from one row leaking into the next.
Reading columns
Steps B, C and D above explain how each column is read. This would
imply that CCD sensor would have an output gate at the end of each
row.
In fact CCD sensors have just one output. Therefore another CCD line is
placed at the end of the column, perpendicular to them all. This line is
called a read-out register. The charge from the element at the end of
each column is transferred to the elements in the read-out register.
Column clocking now stops. Clocking now transfers the charges in the
read-out register to the sense and output gate.
When the read-out register is empty, column clocking can start again
and clock the next charge from the columns into the read-out register.
This procedure carries on until the last charge in the columns has been
clocked into the read-out register and from there to the output.
Similarity to raster scans.
This method of reading one pixel from each row into a column line, then
transferring them one by one to the output is similar to a conventional
television raster scan.
CCD sensors therefore lend themselves very well as conventional
television camera sensors.
107 Sony Broadcast & Professional Europe

116. Part 14 – CCD sensors
Figure 55 Arrangement of MOS elements
Sony Training Services 108

117. Broadcast Fundamentals
Back lit sensors
Many sensors are now back lit. Rather than allowing light to enter the
front of the sensor, passing through the region gates and the insulation
layer, the whole sensor is turned over and light passes directly into the
substrate from below.
Figure 56 Back lit sensors
Substrate thickness
However the substrate is conventionally thick. This makes production
easier. Manufactures only work on the top surface. The thickness of the
sensor’s chip is irrelevant and therefore one less thing they have to
worry about.
Thick substrates also makes for a more robust sensor.
The problems with thick substrates is two fold. Firstly the electrons
loosened by the light are a long distance from the potential wells created
by the region gates. Secondly there is a risk that electrons loosened by
the light do not fall into the correct potential well.
Back thinning
Back lit sensors now tend to be only about 15um thick. This makes sure
that the area of substrate where electrons are loosened by incoming
light are close to the potential wells. There is a greater chance that all
the electrons will be caught, and that the electrons will fall into the
correct potential well.
The gate side of back thinned CCD optical sensors tend to be mounted
on a rigid surface to make the whole device more robust. This surface is
often reflective to make the sensor more efficient by driving any light that
leaks out the back into the substrate.
109 Sony Broadcast & Professional Europe

118. Part 14 – CCD sensors
Problems with CCD image sensors
CCD image sensors are not perfect. They can suffer from manufacturing
defects, and operational anomolies. A few of these are listed here.
Shorts
This is a manufacturing defect. Shorts can occur where silicon oxide
insulation breaks down or where any level in the MOS elements has not
been built properly.
Shorts result in the improper collection of charge, or charge loss. If the
collection of charge is damaged, individual pixels may be lost. If there is
a charge leakage there may be line smearing as charge is lost through
the short as the charge from each pixel is transferred down the line to
the output, past the short.
Traps
A trap is a manufacturing defect where charge is not able to transfer
sucessfully.
Thermal noise and dark current
As previously mentioned CCD image sensors have very noise
characteristics if they are kept cool. However if their temperature rises
thermal noise rises correspondingly.
This can give rise to a number of other problem, but overall will effect the
quality of the image capture process.
Electrons freed by thermal activity are attracted to the potential well
under each pixel. Thus charge develops even if there is no light falling
on the sensor. This gives rise to the term dark current.
CTE (charge transfer efficiency)
CCD sensors must transfer the charge from one element to the next in
the line as efficiently as possible.
Imagine a CCD image sensor with 1024 by 1024 pixels. Charge from the
far end of the furthest line of a device will be transferred 2048 times
before it reaches the output sense and gate.
If there is a 90% CTE in the device the charge will have dropped to
0.000000000000000000000000008 of its original value! This is clearly
not a good thing.
CCD image sensors generally have CTE’s better than 99.999%. With a
1024 by 1024 sensor this still means that the charge in the furthest pixel
has dropped by 0.02%. While this is significantly better than in the case
of a 90% CTE it is still a problem in accurate light measurement
situations.
As sensors increase in resolution so CTE ratings must be kept as close
to 100% as possible.
Sony Training Services 110

119. Broadcast Fundamentals
Chroma filtering and bad QE from front lit devices
Light passing into the sensor substrate passes through the region gates
and insulation layer.
Using polysilicon regions
Making the regions from polysilicon rather than from aluminium allows
light to pass through them. All front lit sensors use polysilicon region
gates.
Filtering effects
Light passing through the regions, even if they are made from
polysilicon, and the insulation layer are filtered. The filtering is non-
linear. Light at the blue end of the light spectrum is attenuated more than
at the red end.
Bad QE
Filtering effects not only make the sensor’s characteristics non-linear,
but they reduce its QE. This makes them less effective where accurate
light measurement is required of camera sensors
CCD image sensors with stores
The problem with the design mentioned is that you have to wait after the
sensor has been exposed and all the pixels charged, for the charges to
be read out. This takes a while.
FT sensors
In the FT (frame transfer) sensor design each column is twice as long.
Half of the columns are exposed. The other half acts as a temporary
store, and are covered by an aluminium mask.
After the image has been exposed the sensor the charges are
transferred quickly into the temporary store. The sensor can then start
exposing the next frame while the frame that was just exposed is output
through the read-out register.
111 Sony Broadcast & Professional Europe

120. Part 14 – CCD sensors
Figure 57 FT sensor
IT sensors
In the IT (interline transfer) sensor design all the pixel charges are
transferred into read-out gates, and from there into separate column
CCD lines. These are called vertical read-out registers. These registers
are masked.
The charges can be transferred into the vertical read-out registers very
quickly, leaving the sensor free to start exposure again.
The vertical read-out registers can then be transferred into the horizontal
read-out register in the normal way.
Sony Training Services 112

121. Broadcast Fundamentals
Overflow gate technology and shuttering
With the introduction of IT sensors came the introduction of an overflow
gate. This gate is placed on the opposite side of each sensor gate from
the vertical read-out register. It can be used in a number of ways.
Figure 58 IT sensor
113 Sony Broadcast & Professional Europe

122. Part 14 – CCD sensors
Using the overflow gate to eliminate flair and burnout
When the overflow gate is closed it will not draw any charge away from
the sensor gate. After exposure all this charge can be drawn away by
the vertical register.
However if light levels get too high the sensor gate will become flooded.
Any charge above a certain level will not be drawn away from the
vertical register and the device will peak causing ‘burnout’ in the image.
Furthermore the excess charge will leak out of the effected gate and into
the surrounding gates spreading the perceived brightness from the
actual bright area.
Therefore the overflow gate is never actually close altogether. In fact, in
its ‘off’ mode it will still draw charge from the sensor gate, but only if the
amount of charge becomes excessive. This prevents the gate from
peaking and stops the flood of charge leaking into any other gates.
Using the overflow gate for shutter opening and iris control
If the overflow gate is opened any charge building up underneath the
sensor gate as a result of light exposure will be immediately drawn away
into the overflow gate. This effectively switches the device off in the
same way as a mechanical shutter would do.
This can make the sensor behave a little like movie cameras with
variable shutter wheels. It is also used in applications like CCTV as an
electronic iris, assisting the mechanial iris in the lens itself.
Using the vertical register and overflow gate for shutter closing
To simulate the shutter closing the accumulated charge under the
sensor gate can be drawn into the vertical register. At the same time the
overflow gate is opened, so that any further charge is drawn away from
the sensor gate.
FIT sensors
One problem with IT designs are that the vertical read-out registers are
very close to the exposed regions of the device. If light levels are very
strong it is possible for charge to leak from the exposed region of the
sensor into the vertical register regions.
In the FIT (frame interline transfer) sensor design both FT and IT design
philosophies are used. The charges built up in the exposed areas are
transferred quickly to the vertical read-out registers, and then into an FT
type store. This takes them away from the vertical read out store so that
they cannot be corrupted if light levels are very high.
Sony Training Services 114

123. Broadcast Fundamentals
HAD technology
Figure 59 The HAD sensor
115 Sony Broadcast & Professional Europe

124. Part 14 – CCD sensors
Sony introduced a new technology for image sensors in 1984. This
technology was a real departure from the conventional designs up to
that date.
Rather than using the photo-excitation technology of older designs this
new design uses an embedded photo diode for each pixel. The photo
diode has a heavily doped p type region called a p++ region. p++ doped
regions have a high number of accumulated holes. Hence the name
Hole Accumulated Diode or simply HAD.
The HAD increases the number of electrons that are released as light
enters the device. These electrons flow down into the device’s substrate
and collect underneath the HAD.
HAD sensors also have the advantage that light does not have to pass
through polysilicon regions to reach the diode. This makes the sensor
more efficient and linear.
HAD sensor operation
HAD sensors comprise an array of HAD’s. An insulation layer of silicon
dioxide is laid on top of the HAD’s and a thin aluminium photo mask if
printed on top of the insulation. The photo mask prevents light from
getting into the sensor except where there is a HAD.
Light passes into the HAD and excites electron flow down into the n type
substrate. At a certain time, when the image is sampled, the voltage on
the polysilicon gate, to the right of the HAD, switches, creating a
potential well that attracts the accumulated charge away from
underneath the HAD.
The polysilicon region is part of a chain of polysilicon regions that form a
vertical register to transfer the accumulated charges out of the device.
Charge is therefore transferred out of the device in the same way as any
other IT or FIT device.
A region of p++ material is fabricated deeper into the device just to the
left of each HAD. This acts as a channel stop, preventing charge from
leaking from underneath each HAD into the vertical register to the left.
Problems with HAD devices
The first problem with HAD sensors is the increased manufacturing
complexity. However, in commercial terms this increased complexity and
its resulting higher cost is more than offset by the increase in
performance. Manufacturing techniques have also improved
considerably over the years making it easier to produce HAD devices
reliably.
The second problem with HAD devices is the same problem facing any
IT or FIT device. Ideally the whole of the front of the device should be
light sensitive so that light hitting anywhere on the surface of the device
is picked up and output as a signal. The amount of space taken up by
the vertical registers, channel stops, polysilicon regions, etc. detracts
from this light sensitive area. This problem is partially overcome in later
designs.
Sony Training Services 116

125. Broadcast Fundamentals
HyperHAD
HyperHAD, sometimes called microlenticular technology, improves on
the simple HAD device by fitting a small lens in front of each HAD. This
channels light from the area around the actual gate that would otherwise
be lost, increasing the effective area of each HAD outside of the HAD
itself.
HyperHAD sensors were introduced in 1989 and increased the
sensitivity and efficiency of HAD sensors.
Figure 60 The HyperHAD sensor
SuperHAD sensors
SuperHAD was introduced in 1997. It is basically similar to the
HyperHAD design, but the actual lenses are larger, and are therefore
able to capture more light, making SuperHAD sensors more sensitive
than HyperHAD sensors.
PowerHAD sensors
PowerHAD is a marginal improvement on SuperHAD. The microlens
structure is similar to that of SuperHAD but the capacitance of the
vertical registers is reduced.
117 Sony Broadcast & Professional Europe

126. Part 14 – CCD sensors
Figure 61 The SuperHAD & PowerHAD sensor
PowerHAD EX (Eagle) sensors
Previously simply called New Structure CCD and now sometimes called
Eagle sensors, PowerHAD EX sensors have another lens placed
between the on-chip microlens and the HAD. The microlenses are also
larger. Infact they are so large that they overlap, leaving no area on the
device that is not somehow concetrated into a HAD. This further
contentrates light capture increasing the efficiency of the sensor still
further.
Figure 62 The Eagle sensor
Sony Training Services 118

127. Broadcast Fundamentals
The insulation layer between the polysilicon gate and the potential well
in the substrate underneath is also thinner. This decreases the gate’s
capacitance and increases the ‘depth’ of the well, making it better able to
collect the HAD’s accumulated charge.
Figure 63 Lenticular designs
EX View HAD sensors
EX View HAD sensors are physically the same as any other HAD based
Figure 64 EX View HAD response
sensor. However the exact doping levels and construction of the HAD
makes it more sensitive to infra-red light. This makes EX View devices
very appropriate to security and low light level cameras.
Single chip CCD designs
Professional and broadcast camera systems normally process the image
they are looking at as three primary colours. (See Colour Perception.)
This is important for good colour matching, and is essential if the
camera’s outputs are to comply with broadcast signal standards. (See
Colour in Television.)
119 Sony Broadcast & Professional Europe

128. Part 14 – CCD sensors
The split from the original image to three images, one for each primary
colour, could be done in the camera’s electronics. However it is better to
do the split optically. Therefore all professional and broadcast cameras
have a three way dichroic splitter just behind the lens, and three CCD
sensors, one on each outputs from the dichroic splitter. Each sensor is
responsible for one of the primary colours. (See Dichroic Block Design.)
Figure 65 Single chip HAD design
However this is either too expensive or simply not possible for smaller
cameras and cameras intended for industrial and domestic use.
Small security cameras simply do not have enough space for a dichroic
block and three CCD sensors. The cost of the dichroic block and three
CCD sensors would make domestic cameras simply too expensive. In
any case the increase in quality would almost certainly not be
appreciated.
Therefore all these types of cameras have one CCD sensor. The split
from the original image to its primaries is still required and is still best
done optically. Therefore single chip CCD cameras have a filter screen
fitted over the sensor.
CCD filter screens
CCD filter screens consist of an array of small coloured squares. The
resolution of the CCD sensor and the filter squares is the same. Thus
each pixel in the sensor has one filter square.
Random filter screenThe human eye is very good at recognising
patterns. Thus it may seem a good idea to design a filter screen with a
random design of squares of the three primary colours.
Sony Training Services 120

129. Broadcast Fundamentals
However there is a chance that there will be discernable areas of one
colour.
Figure 66 The random screen
The Bayer filter screen
A popular screen design is the Bayer screen. This screen has a greater
number of green squares, because of the human eye’s relative high
sensitivity to green areas of the colour spectrum.
Figure 67 The Bayer screen
The Bayer screen is very popular in single CCD cameras.
The pseudo random Bayer screen
The problem with the Bayer screen is that there is a strong pattern. It
may be possible, at certain low resolutions, for the human eye to pick
out this pattern.
121 Sony Broadcast & Professional Europe

130. Part 14 – CCD sensors
Figure 68 The psuedo random Bayer screen
Thus by jumbling the Bayer pattern in a particular way it is possible to
retain the same ratio of the three primary colours as the basic Bayer
pattern, but with no easily definable pattern. This design also makes
sure that there are no large ares of one colour by ensuring that there are
no squares of the same colour next to each other.
Reading out from a single CCD camera
Filter screens need to be very accurately fitted to the sensor. If the filter
position is defined and fitted exactly, it is possible to define which pixel is
responsible for which primary colour.
Reading single CCD’s then becomes reasonable straightforward, if a
little more complex, than with triple CCD designs.
Figure 69 Pixel interpolation
The camera will read each pixel out in sequence in the same way.
However the read-out circuitry knows which pixel is responsible for each
Sony Training Services 122

131. Broadcast Fundamentals
primary colour, and sequentially demultiplexes all the pixels for each
primary to a different part of the electronics for further processing.
Pixel interpolation
With a Bayer filter screen half the pixels are responsible for the green
primary, a quarter for red and a quarter for blue. This looks like Fig.
,where the each pixel shows the brightness of each primary colour at
that point in the picture.
Some single CCD cameras rebuild a full image of pixels for each primary
by interpolating the pixels it has to make up the missing pixels. Putting
these three separate images back together gives a much more pleasing
result.
A newer approach is to follow the interpolation process by a small
amount of sharpening to improve the perceived quality of the image. (If
you squint at the three images the sharpened one looks better!)
Noise reduction
Noise is a problem in any image sensor. The first area where noise in
introduced is in the pixel itself where thermal noise is an enduring
problem (see page 110). The only way of eliminating this kind of noise is
to freeze the sensor to absolute zero to eliminate thermal electron
movement. This is not practical and the sensor would fail to work at all
anyway! It is really down to good image sensor design to accept thermal
noise will exist and reduce its effect.
O u tp u t
b u ffe r
S ense
F ro m s w itc h
C C D O u tp u t
a rra y s ig n a l
S e n s in g A u to -z e ro
c a p a c ito r s w itc h
0V 0V
Figure 70 Auto zeroing
Another area where noise can be introduced is during the charge
transfer period, where the charge collected under each pixel is
transferred to the ouput.
123 Sony Broadcast & Professional Europe

132. Part 14 – CCD sensors
The last area where noise can be a problem is in the output gate itself,
where the charge is placed into a capacitor and measured as a voltage.
This capacitor needs to be carefully and quickly drained of any charge
from a previous pixel, or from anywhere else, before the pixel charge is
put into it.
Auto zeroing
Auto zeroing is the traditional way of cancelling noise in sensing
comparators, analogue to digital convertors, and image sensors. It aims
to pull the capacitor charge down to zero just before the pixel charge is
put in. This is done by building a switch into the circuit just before the
capacitor. The clocking circuit closes this switch just before the pixel
charge is input.
Auto zeroing switches need to be very low impedance if they are to work
effectively
P ix e l
A c tiv e s a m p le
p ix e l & h o ld
F ro m c lo c k
C C D +
a rra y O u tp u t
s ig n a l
-
R e fe re n c e
R e fe re n c e s a m p le
p ix e l & h o ld
c lo c k
Figure 71 Correlated double sampling
Double sampling
Double sampling (DS) is a method of eliminating the effects of noise in
the sense capacitor. Firstly the sense capacitor zeroed, and set to a
predetermined charge that is sensed as a reference voltage. Then the
capacitor is zeroed again and the pixel charge placed across the
capacitor.
Noise will be common to both samples. Therefore any difference
between the sensed reference voltage and the voltage this reference
should be is noise and is subtracted from the pixel voltage.
Sony Training Services 124

133. Broadcast Fundamentals
Correlated double sampling
Correlated double sampling (CDS) eliminates some of the drawbacks of
auto zeroing, and provides for a simpler sampling method that for DS.
CDS places the operation range of the sensing capacitor away from
zero, where it becomes noisy and difficult to sense accurately.
CDS operates by placing a predetermined charge on the sensing
capacitor, and sensing this as a voltage, just as with DS. The pixel
charge is then input to the capacitor, without zeroing, and the voltage
sensed again. The difference between the two values is the pixel itself.
D u m m y p ix e ls
A c tiv e p ix e ls
D u m m y p ix e ls
D u m m y p ix e ls
D u m m y p ix e ls
A c tiv e p ix e ls
P ix e l
A c tiv e s a m p le
p ix e l & h o ld
F ro m c lo c k
C C D +
a rra y O u tp u t
s ig n a l
-
R e fe re n c e
D um m y s a m p le
p ix e l & h o ld
c lo c k
Figure 72 CDS with dummy pixels
125 Sony Broadcast & Professional Europe

134. Part 14 – CCD sensors
CDS and dummy pixels
One method of CDS involves the sampling of pixels outside the active
region on the CCD array. Many CCD sensors are designed such that
there are a few pixels on the edge of the array that are not focussed on.
These dummy pixels are masked and are effectively giving out a signal
corresponding to black.
The whole array is given a small reference precharge voltage. This pulls
the whole array from zero and gives the masked pixels a specific value.
The output gate has a sample and hold circuit that samples the masked
pixels and holds their average charge as an output voltage.
The active pixels are routed to a separate sample and hold and
compared to the reference.
P ix e l
s a m p le
A c tiv e & h o ld
F ro m p ix e l
CC D c lo c k
a rra y +
O u tp u t
s ig n a l
-
R e fe re n c e R e fe re n c e
s a m p le s a m p le
Dum m y & h o ld 1 & h o ld 2
p ix e l
c lo c k
Figure 73 CDS with dummy pixels and triple sample & hold
Triple sample and hold
A further improvement on the double sample and hold design is to add
another sample and hold circuit directly after the reference sample and
hold circuit. This seemingly silly idea removes transition errors within the
sample and hold switches themselves. This is shown in Fig. . Sample
and hold circuits 2 and 3 are switched for every active pixel. Any
switching transition errors in the active pixel circuit will be eliminated by
circuit 3.
Correlated triple sampling
A further method of sampling the CCD array is called correlated triple
sampling (CTS). This method is not used very much. It improves the
noise cancelling effect of CDS by taking a third sample part way through
the pixel reset period. This third sample allows for more information to
get gained about the noise pattern within the array.
Fowler sampling
A natural conclusion of the progression from DS, CDS and CTS is to
take multiple samples. This is particularly useful for long exposures.
Multiple array samples are taken at the beginning and end of the
integration time. These are averaged out to eliminated noise.
Sony Training Services 126

135. Broadcast Fundamentals
Fowler sampling is not appropriate for broadcast camera design as the
exposure time is relatively short. This kind of sampling is more
appropriate for imaging in low light situations, like space exploration.
The future of CCD sensors
CCD sensors, in one form or another, have gained universal dominance
within the video camera world. They have completely replaced the older
tube sensor designs of old (See Early Image Sensors.)
However a newer technology has begun to emerge that promised to
replace conventional CCD sensor technology is CMOS. Although CMOS
technology has been around longer than CCD technology, it was difficult
to produce. CMOS is now a very important technology for
microprocessor design, digital signal processing chip design and for
many other types of chip design.
Production processes are now reliable and versatile enough for some
manufactuers to use CMOS for image sensors.
However the move from conventional CCD to CMOS technology will not
be the revolution that the move from tube to CCD was. CMOS
represents more of an evolution rather than a revolution.
127 Sony Broadcast & Professional Europe

136. Part 15 – The video tape recorder
Part 13 The video tape recorder
A short history
There are many different video tape formats. Some have been more
successful than others. Some have been technically more superior than
others. There is often very little correlation between commercial success
and quality or technical excellence.
Beginnings
AEG Magnetophons, Bing Crosby and the beginnings of Ampex
The video tape recorder has been in existence for about 50 years. The
only way to record video before video tape recorders was to use film.
Film did not lend itself well to television. It required a telecine to convert
the film to a television signal.
During the second world war John (Jack) Mullin was stationed in
England as part of the Signal Corps. He became intrigued by the fact
that the Germans were able to transmit propaganda and music in the
middle of the night. The music was particularly interesting, as the quality
was very good, but the music was orchestral. It seemed unlikely that this
quality came from 78rpm record disks, and it seems even more unlikely
that entire orchestras were being employed to play every night.
As the Allied forces pushed forward into Germany Jack found the
machine that was able to play out with such quality. It was an AEG
Magnetophon. He grabbed two and proceeded to ship the mechanical
parts back to the States using the war souvenir parcel service. He
reassembled a machine rebuilding the electronics and making a few
improvements on the original AEG design. He demonstrated the
improved Magnetophon at a number of venues.
Bing Crosby saw one of the demonstrations and became interested. He
desperately wanted to avoid doing live radio shows every day, and knew
that quality of the shellac based recording methods available at the time
were so poor that audiences at home could tell the difference between a
live show and a recorded one.
Bing Crosby invested $50000 in Ampex to have John Mullin’s machines
produced on a more commercial basis. Ampex was a very small
company based in Redwood City, California. It had been making electric
motors for aircraft during the was and was looking for new projects to get
involved in. Bing Crosby agreed the investments on the basis that his
company Crosby Enterprises would become the sole marketing channel
for Ampex machines.
Ampex audio recorders were very successful and allowed radio
programmes to be pre-recorded and edited before going live, but still
maintaining such quality that radio listeners could not tell the difference.
In the mean time 3M had made significant advances in the formulation of
magnetic tape, and Ampex had set up a division specifically to make
tape to supply the rapid increase in demand.
Sony Training Services 128

137. Broadcast Fundamentals
The first commercial video recorders
At the end of the 1940’s television was enjoying a jump in popularity.
Television producers had to use modified film technology to record
television shows. The quality was poor. A new method of recording
video was desperately required, so that producers could edit shows and
delay transmission for different time zones.
In 1949 John Mullin approached Bing Crosby and proposed that the
same plan that was used design a commercial audio tape machines be
used for video. A team within Ampex was put together and the first
working machine was demonstrated in 1950.
Other early machines
In England Axton began research work in 1952 on a video recorder they
called Vera (Vision Electronic Recording Apparatus).
By the mid 1950’s other companies like RCA had started projects to
develop a video tape recorder.
The 1970’s
During the early 1970’s several companies, including Sony, Teac and
JVC introduced a semi-professional format based on a ¾” tape, called
U-Matic
The 2” machine was superseded by machines using a 1” tape during the
mid 1970’s, with companies like Sony and Bosch entering the fray. This
format was truly helical, with the tape now wrapped around the drum
which spun almost in line with the tape, rather than at right angles to it.
The Bosch machines were ratified as the B format while the SMPTE
ratified the 1” tape standard as the more successful C format.
The 1980’s
By the early 1980’s ½” tape based broadcast machine began to appear.
The most successful of these was the Sony Betacam and Betamax
formats. Betacam, the broadcast format, was later improved with the
introduction of Betacam SP (Superior Performance).
The domestic ½” format, Betamax, eventually lost the commercial battle
with the VHS (Video Home System) format, although it was technically
superior.
In the late 1980’s digital video tape recorders had begun to appear with
the D1 and D2 tape formats. A digital version of Betacam SP was
introduced. Called Digital Betacam this format became a widely
accepted standard for high quality digital television recording.
The 1990’s
½” tape formats based on the original Betacam format were introduced
during the second half of the 1990’s. They introduced MPEG
compression to mainstream broadcast tape recording and the idea of
high quality low bit rate recordings, metadata, and offering a bridge
between streaming technology (tape) and file based technology
(computers).
129 Sony Broadcast & Professional Europe

138. Part 15 – The video tape recorder
The DV format was introduced during the mid 1990’s. Originally
designed as a digital replacement for VHS and Hi8, it has been more
popular as a domestic camcorder format rather than for recording
television programmes at home.
In the professional and broadcast arenas manufactuers have squeezed
extra quality and performance out of the DV format to produce the
DVCam (Sony) and DVCPro (Panasonic) formats suitable for more
professional and broadcast use.
DV, and its high quality derivatives, are helical scan systems, indeed
there is actually little fundamental mechanical different from the very first
true helical scan video recorders. They are just a lot smaller.
The present day
The domestic arena
In the domestic environment VHS is still king, although its days are
almost certainly numbered. The general quality of television output (from
an image point of view) has been steadily increasing over the last few
years and people are starting to realise just how bad VHS is. Even from
a convenience point of view VHS is starting to look cumbersome and
fragile.
DV was designed as a possible replacement to VHS. However
manufacturers have never produced DV equivalents of the ubiquitous
domestic VHS recorder, with remote control and timer functions. It looks
likely that VHS will be superseded by recordable optical disk rather than
tape.
The professional and broadcast arenas
The transition to hard disk
Broadcast television has seen an increased use of hard disk technology.
Indeed people have predicted that tape will be replaced by disk for many
years, and yet new tape formats keep appearing, and broadcasters are
still buying tape based technology.
It is true that hard disk technology is being used a lot more than it used
to be, and it is slowly replacing areas of the broadcast chain previously
occupied by tape technology. Hard disk is now used heavily in post
production and editing.
However tape is still cheaper, and more robust than hard disk. It is still a
popular choise in acquisition, and archive storage. All popular
camcorders in use today use tape. It is removable, can be treated with a
fir amount of disrespect, and is readily available.
Archive and long term storage systems use tape, although video and
audio material is now generally stored as digital data, and is often
compressed to further save space on the tape. Tape robotics machines
allow large amounts of tapes to be stored safely, with the advantage of
automatic scheduling and database support, so that video and audio
archives can be searched
Sony Training Services 130

139. Broadcast Fundamentals
It is unlikely that hard disk will entirely replace tape. It is more likely that
optical disk, using blue laser technology, will replace tape.
The cost/quality balance
Cost is now more of an issue than it ever was. The newer tape formats
are a careful balance between cost and quality. ‘Cost’ means total cost,
not just the price of the equipment, but also maintenance costs and
running costs, commonly referred to as ‘total cost of ownership’. ‘Quality’
means the image quality, as well as manufacturing quality and quality of
after sales service and support.
Digital tape technology satisfies this careful balance much better than
analogue tape technology. All major television companies now use
digital tape technology, and almost all of these companies record new
material in digital format. However, with large analogue tape archives,
analogue tape players are still in popular use.
For very high quality DI is still used. It is expensive but offers the kind of
quality not attainable by any other broadcast format. Many companies
use Digital Betacam, and a few the M2 format. While these formats are
slightly compressed they still offer superb image quality at a much more
realistic price.
Betacam SX, DVCam and DVCPro are popular for news gathering
where convenience and price are more important. Indeed domestic DV
is being used in many professional areas.
The stream/file bridge
The major thrust in digital tape technology is in bridging the very difficult
gap between streams and files.
Video and audio are basically streams. They have no beginning and no
end, and do not contain any kind of header, label, or other information.
Video and audio are also strongly related to time. They are continuous
and must be played at the correct speed without breaks.
Files, on the other hand, are contained chunks of data. They have a
header, information about the contents of the file, etc.. Files are also not
related to time. When copying files from one location to another it really
does not matter how long it takes, how the data is actually transferred, or
if the beginning of the file gets to its destination before the end.
With increasing use of computer technology in broadcast, television
companies require a way of bridging this gap between these two basic
methods of storing media.
Manufactures are starting to produce tape recorders that can place extra
information into the video or audio stream, much like a computer file can.
This so called ‘metadata’ is the focus of a lot of research work.
Manufacturers are also starting to introduce tape formats that can output
video and audio in packets, or in file structures, so that they can be
saved as files on hard disk, and treated as files within the television
station.
131 Sony Broadcast & Professional Europe

140. Part 15 – The video tape recorder
The MPEG IMX format is a good example of this. Although this format is
essentially a stream recording system just like the original Ampex
VR-1000 of the mid 1950’s, it is able to output a stream as a series of
chunks, or packets of data.
The E-VTR takes this one step further and allows bits of video or audio
to be marked, and played out as a file. Although the material has been
recorded from a video or audio connection, as a stream, locked to time,
it is now being played out to a computer network, as a file with
associated metadata, and at a speed governed by the network, in
bursts, faster or slower than real time.
Sony Training Services 132

141. Broadcast Fundamentals
Magnetic recording principles
Although magnetic recording heads may have got a lot smaller, the
materials used are better, and manufacturing tolerances a lot higher, all
video tape recorders depend on the same basic principles of recoding a
signal to magnetic tape.
Principle of a magnetic field
Man has known of the existence of magnets for several thousands of
years. The Chinese used them to invent the compass about one
thousand years ago.
However the fact that an electric current develops a magnetic field was
not discovered until much later. In 1820 Hans Christian Oersted
discovered that a straight wire, carrying an electric current, developed a
magnetic field, which circulates around the wire. Andre-Marie Ampere
discovered that the magnetic field could be concentrated and magnified
by winding the wire into a coil. William Sturgeon later discovered that
placing an iron core inside the coil greatly increased the strength of the
magnetic field, and bending the coil and iron into a ‘U’ shape further
concentrated the field at the two ends of the ‘U’ shape. A little later
Joseph Henry insulated the wire, thus enabling a larger and tighter coils
to be wound.
C o il w in d in g s
C u rre n t
F lu x
Figure 74 Magnetic field around a wire and coil
It is perhaps a pity that Oersted, Ampere, and Henry have all has their
names immortalised as units of magnetic flux density, current and
inductance, but Sturgeon’s name has sunk into relative obscurity.
The magnetic field is known as flux, and its strength, the magnetic flux
density. Flux finds less resistance through some materials than through
others. Many materials are magnetic. This means that they become
magnetised if subjected to a magnetic field. The ability to become
magnetised is called the remenance.
133 Sony Broadcast & Professional Europe

142. Part 15 – The video tape recorder
W in d in g s
F e rrite c o re (to ro id )
F lu x
Figure 75 The toroid
Principle of electromagnetic induction
The opposite of the principle of a magnetic field is that of
electromagnetic induction. When a magnetic field is applied to a wire it
induces a current to flow in the wire. To be more exact, when a magnetic
field changes, it induces a current. Indeed it does not matter how strong
the magnetic field is, no current will be induced if this field remains
constant. Conversely, a small magnetic field could induce a high current
if it changes rapidly.
Figure 76 The basic magnetic record head
Sony Training Services 134

143. Broadcast Fundamentals
Using the qualities of magnetic materials in video tape recorders
The purpose of a video tape recorder is to use a record head to
magnetise the magnetic material on the tape, and for the playback head
to detect this magnetisation.
Record heads use the same basic principles discovered by Sturgeon. By
bending an iron core into a ring a high flux density could be made to flow
around the ring. A small gap is left in the ring. Flux jumps this gap,
bulging slightly outwards, and could be used to magnetise the surface of
the tape.
However, although a large gap results in a greater flux bulge which has
a greater influence on the tape, it also offers greater resistance to flux
and therefore reduces flux density. Thus the design of the record head
meets its first compromise.
Playback heads use the same design. When the magnetised tape
passes across the head gap, it induces a small magnetic flux in the head
core. As the flux changes, so a small current is induced in the coils.
Thus record and playback head cores need to be made from magnetic
material with low flux resistance and low remenance. Conversely the
magnetic material used in tape needed high remenance so that the
maximum amount of signal could be recorded.
The essentials of helical scan
The bandwidth problem
Humans can hear audio from about 20Hz to about 20kHz. The
bandwidth of audio is therefore about 20Khz. If we consider modulation
or sampling, the Nyquist criteria doubles this bandwidth to about 40kHz.
No matter how we look at it, the frequencies involved are well within the
capability of magnetic tape and recording head technology, using
stationary record and playback heads.
Video is very different. Broadcast channels have a total bandwidth of
about 6MHz. In component form we would expect to retain as much of
the quality as possible, and give the luminance signal as near to 6MHz
as we can. Each colour difference signal may have a bandwidth of about
3MHz.
Add these bandwidths, and take into account the Nyquist criteria and
any recording system will need at least 24MHz of bandwidth!
If nothing else, these somewhat crude calculations show us that we
cannot record video on magnetic tape in the same way we do with
audio. Either the recording system has to be radically different, or the
bandwidth must be reduced.
Head to tape speed
Key to the problem of bandwidth is the relative head to tape speed. In
analogue audio recorders this is achieved by pulling the tape across a
static head.
135 Sony Broadcast & Professional Europe

144. Part 15 – The video tape recorder
If the bandwidth of the video signal is higher one could simply increase
the tape speed. This idea was tried in the first video recorders. Ampex
were at the forefront of video tape recorder development and
demonstrated video recorders with high speed tape at the beginning of
the 1950’s
Other groups were working on the design of a video recorder, like the
BBC Vera (Vision Electronic Recording Apparatus) which ran through
tape at 21 metres per second. With reels 21” in diameter just 15 minutes
could be recorded. In the States RCA built a prototype that ran through
tape at 9 metres per second. An improvement, maybe, but it only gave 9
minutes of recording time.
It became clear to all those groups working on video recorder designs
that the kind of tape speed required for the kind of bandwidth normally
found in video made the machine difficult to control and used up vast
amounts of tape.
Designers eventually decided that, to achieve the relative head to tape
speeds required, the recording head itself could not stay still.
Early scanning techniques
Many of the earliest video recorders used a moving head to increase the
relative head to tape speed. From the very first prototypes this was
achieved by mounting the heads on a rotating drum.
Ampex Mark 1 arcuate recorder
An early notable attempt to increase the head to tape speed was the
Ampex Mark 1 arcuate recorder. Built in 1952, this machine wrote the
video information onto tape as arcs using three heads on the drum. It
proved unreliable and difficult to regulate. The geometry also wasted
more tape than was necessary. Arctuate tape machines were not
successful.
Tap e
G u id e
D ru m
R e c o rd e d
tra c k s
Head
Figure 77 The arcuate recorder
Sony Training Services 136

145. Broadcast Fundamentals
The Ampex Mark 1 did give rise to the transversal scanning technique
used by their quadruplex machines.
The Ampex VR-1000 quad recorder
The original Ampex VR-1000 machine used 4 heads fitted 90 degrees
apart on a spinning drum. (Hence the name “quadruplex” or simply
“quad”.) The drum spins in line with the tape. The tape itself is 2 inches
wide and is curved by a vacuum chamber, to fit around the drum.
Each head records a stripe of video across the tape, called a track. As
soon as one head breaks contact with the tape the next one is ready to
carry on. The tape moves a little more than the width of one of these
tracks before the next head comes along. This keeps tape speed slow.
The drum spins quickly. This makes the head to tape speed high and
allows a high bandwidth signal to be recorded.
A u d io /c u e R /P
h e a d s ta c k
A u d io tr a c k
A u d io /c u e e r a s e
Id le r h e a d s ta c k
Id le r
C a p s ta n
D ru m m o to r C o n tro l
h e a d s ta c k P in c h w h e e l
Vacuum
cham ber
V id e o h e a d s C u e tra c k
H e lic a l (v id e o ) t r a c k s
C o n tro l tra c k
D r u m d e ta il
(V a c u u m c h a m b e r re m o v e d )
Figure 78 The Quad tape path
The drum spins at 14400rpm (USA machines), recording 960 tracks per
second. 16 tracks make up each video field. Each head recorded or
played back just 16 lines of video. Quad recorders manage to use up 15
137 Sony Broadcast & Professional Europe

146. Part 15 – The video tape recorder
inches of tape per second. With a 4800 foot reel of tape it was possible
to record just over 1 hour of video
The quad machine is generally not considered a true helical scan
machine, because the tracks’ angle is almost perpendicular to the tape’s
direction. The scanning method is generally referred to as transversal.
A longitudinal control track is also recorded along the edge of the tape
so that the machine can lock to it, with the playback heads exactly
following the tracks as they were recorded.
Audio can also be recorded with a static head and a longitudinal track,
and there is also provision for a lesser quality audio track called cue.
A u d io tr a c k H e lic a l v id e o tr a c k s G u a rd b a n d s
C o n tro l tra c k C u e tra c k
Figure 79 The Quad tape footprint
Quad remains the most long lasting video tape format of all time. This is
partly because there was no alternative video recording format for many
years, and partly because no video recording format since has been
able to last the length of time Quad was used for, before it has been
superseded by something else.
However Quad has its quirks. Tape operators would have to regularly
ensure that the machine was aligned correctly. The drum and vacuum
chamber were particularly sensitive. The forces applied literally stretched
the tape. Applying exactly the same level of brutal treatment to the tape
every time it was played was not easy.
True helical scanning
Helical scanning uses the same basic geometry as transversal scanning.
However, instead of the drum spinning vertically in transversal scanning,
producing tracks on tape that are almost vertical, helical scanning uses a
drum that is nearly horizontal.
Another difference between transversal and helical scanning is the tape
wrap. In transversal scan the vacuum chamber achieves a slight wrap
around the drum, stretching and distorting the tape as it does so.
However in helical scan designs the wrap is huge, by comparison. In
some formats wraps of almost 360 degrees have been used, although a
little over 180 degrees is more common. Even though the wrap is large
Sony Training Services 138

147. Broadcast Fundamentals
none of the stresses common in Quad machines is placed on the tape in
helical scan machines.
E x it g u id e T r a c k e n d p o in t a n d H e a d (w r itin g tr a c k )
E x it e n d o f a c tiv e w r a p
p o in t
U p p e r d ru m
(r o ta tin g )
H e a d (n o t w r itin g )
Tap e
E n tra n c e
p o in t T r a c k s ta r t p o in t a n d
b e g in n in g o f a c tiv e w r a p
E n tr a n c e g u id e
Figure 80 Helical scan (from above)
During recording and playback the tape moves slowly through the
machine and round the drum. The point where the tape meets the drum
is called the entrance side. The point where the tape leaves the drum is
called exit side.
H e a d (w r itin g tr a c k ) E n tra n c e
U p p e r d ru m Tape P r e v io u s ly g u id e
(r o ta tin g ) w r itte n
E n tra n c e Tap e tra c k s
g u id e
Tape
E n tra n c e
g u id e E x it
H ead P r e v io u s ly g u id e
(n o t w r itin g ) E x it
w r itte n g u id e
L o w e r d ru m R a b b e t in tra c k s
(s ta tic ) lo w e r d r u m L o w e r d r u m (s ta tic ) R a b b e t in
lo w e r d r u m
Figure 81 Helical scan (from the side)
The drum assembly itself sits in the machine at a slight angle, and in
most cases consists of two halves. The top half spins and the bottom
139 Sony Broadcast & Professional Europe

148. Part 15 – The video tape recorder
half is static. The record and playback heads are fitted to the bottom
edge of the top half.
The bottom half has a rebate cut into it, called a rabbet. The rabbet is cut
at an angle, and in most machines is very close to the top of the lower
drum near to the entrance side, and much lower at the exit side. Taking
into account the angle of the drum assembly as a whole, the rabbet is
effectively horizontal. The bottom edge of the tape rests on the rabbet as
it passes around the drum assembly.
The angled drum assembly and the way that the rabbet is cut, mean that
the heads prescribe a helical scan across the tape as the drum spins. In
some formats the track angle goes upwards, in others it goes
downwards, depending on the angle of the rabbet and the rotational
direction of the drum. In most modern tape formats the drum spins anti-
clockwise. The tracks recorded on tape are about 5 degrees from the
line of the tape, and very long, compared to those in transversal
scanning machines.
Modern video recorder mechadeck design
The mechadeck is the mechanical part of a video recorder, consisting of
the tape reels, any servo mechanism, including the pinch wheel and
capstan mechanism, tension regulation, the drum and all the record and
playback heads, tape cleaners, head cleaners and cassette handling
mechanism.
Most modern video tape recorders now have similar mechadeck
designs. Tape, normally enclosed in a cassette, travels from the left
(supply) reel to the right (take-up) reel during normal recording or
playback. The route taken by the tape from the supply real to the take-up
reel is called the tape path.
The tape path from the supply reel to the capstan/pinch wheel is called
the supply side. From the capstan/pinch wheel to the take-up reel is
called the take-up side. The supply side of the tape path is by far the
most important part. It contains all the record and playback heads. Good
supply side tension regulation is important. There is often no take-up
side tension regulation at all.
Many machines have a tape cleaner placed in the tape path as the tape
leaves the supply reel. There will also be a tension regulator in the tape
path between the supply reel and the drum, to ensure that the tension
around the drum is correct.
Many video tape machines have static heads before the drum. A full
erase head will be fitted to all recorders to erase everything on the tape
before any new recording is made. Some machines also include a
control head. This head records special pulses on a longitudinal track
either along the top or bottom edge of the tape, and plays these pulses
back to help the servo system lock during playback.
There is a guide just before the tape wraps around the drum. Called the
entrance guide this guide has a flange that touches the top of the tape
stopping it from riding up as it is wrapped around the drum. The tape is
prevented from dropping by the drum rabbet.
Sony Training Services 140

149. Broadcast Fundamentals
There is another top touching guide on the exit side of the drum, called
the exit guide. There may also be one or more static heads between the
exit side of the drum and the capstan/pinch wheel. These are commonly
used for timecode and audio, but may also be used for control.
The capstan is a precision servo controlled motor responsible for pulling
the tape through the tape path at the correct speed and position. A pinch
solenoid will force a soft rubber pinch wheel against the capstan
squeezing the tape between the two. This force is strong enough to stop
the tape from slipping but not so strong as to damage it.
Although the capstan rotates at essentially a constant speed the servo
system constantly speeds up and slows down by minute amounts to
keep the tape in the correct position relative to the drum, and to ensure
that the video heads are moving exactly up the centre of the helical
tracks. The control head and track are used in some machines to
accomplish this. Others use the RF signal from the helical tracks.
The capstan and pinch wheel isolates the supply side of the tape path
from the take-up side.
Various guides guide the tape back into the cassette and onto the take-
up reel. Some machines include a take-up tension regulator to ensure
that there is a small amount of slack on the take-up side to allow for
sudden changes in direction during normal operation, but not so much
as to start throwing tape loops. The drum assembly is at an angle and
some machines include an angled guide to ensure that the tape is
straight as it re-enters the cassette.
Guard bands
Older helical scan machines record analogue composite video. Each
track contained both the luminance and the colour video information. As
with every video recorder before and since it was important to ensure
that the playback heads followed the recorded tracks exactly.
All early machines used a longitudinal control track running along the top
or bottom edge of the tape. The pulses recorded on this track helped the
machine find the beginning of each helical track.
The control track is not an exact method of finding the beginning of each
helical track. The helical tracks themselves are very thin, and it is
possible for the control head to be in the wrong place. Any error in the
position of the control head, and the video heads will not move up the
centre of the helical tracks.
Tape H e lic a l tr a c k s G u a rd b a n d s
Figure 82 Guard bands
141 Sony Broadcast & Professional Europe

150. Part 15 – The video tape recorder
If the helical tracks are packed together tightly there will be a risk of
playing back a portion of video from another helical track. Editing also
becomes problematic as recorders run the risk of over-recording
material on tape that it should not.
A guard band is a space between helical tracks with no recorded signal.
Early machines used the concept of a guard band to prevent the video
heads from picking up the recording from adjacent tracks during play
back, if the control head is slightly in the wrong position, and to prevent
the machine overwriting the wrong helical track during edits.
However guard bands use up tape. Later machines abandon guard
bands in favour of track azimuth, and thus save tape.
Helical tracks with track azimuth
Later video machines recorded component video, with separate circuitry,
record heads, playback heads, and tracks for luminance and colour.
R e c o rd / C o ils
P la y b a c k
H ead gap
head
No P o s itiv e N e g a tiv e
A z im u th A z im u th A z im u th
H e lic a l tr a c k s
Tape
Figure 83 Track azimuth
Sony Training Services 142

151. Broadcast Fundamentals
Making a tape machine that is able to record and playback entirely in
component increases the quality of recording over older composite
machines, and removes the problems associated with editing composite
material. However component video recorders are more expensive than
composite ones as they are effectively two video recorders in one.
Designers needed a way for these machines to differentiate the helical
tracks responsible for luminance from those for colour. The method
adopted was track azimuth.
Track azimuth involves tilting the head gaps over at an angle. The
luminance head gaps are tilted over positively and the colour head gaps
negatively.
During recording the luminance tracks are recorded with a positive
azimuth and the colour tracks with a negative azimuth.
If the machine is badly aligned and a luminance playback head is trying
to play back a colour track, the angle of the recording will be incorrect. In
fact it will be incorrect by twice the azimuth angle. This will severely
reduce the signal.
Azimuth angles of about 15 degrees are popular. This gives a total error,
if the each head is on the wrong track, of about 30 degrees.
Azimuth replaces the need for a guard band. The colour tracks are
effectively guard bands for the luminance heads and visa versa. Helical
track can be packed next to each other, saving a lot of tape, and
increasing the tape’s recording capacity.
Video head design
The principles used by video tape recorders to record a signal onto
magnetic tape have not changed since the very first tape recorders.
They still rely on a donut shaped head made from ferrite, or some similar
material, with a slot cut in its front face and a coil wrapped around its
back. The dimensions used in modern video recorders may be a lot
smaller, but you can still find a donut idea somewhere in every video
record and playback head.
The video heads, and the tracks they record, are thin. Older formats
used heads close to 100um thick. Modern formats are using heads less
than 10um thick.
Video heads are no longer the classical round donut shape. They are
square. The surface in contact with the tape is along rectangle. This
reduces tape bounce at the head gap and reduces wear.
The coils are wound on the sides of the head. Only a few turns are
required on each side for the head to be effective.
Channeling flux
One of the challenges facing head designers is to channel as much flux
to the front of the record head gap as possible, where the head is in
contact with the tape, and where recording and playback will take place.
Likewise the front of the playback head gap needs to be as sensitive as
possible to achieve maximum signal off the pre-recorded tape.
143 Sony Broadcast & Professional Europe

152. Part 15 – The video tape recorder
H ead gap
C o re Ty p ic a l v id e o h e a d
E tc h in g
T r a c k w id th
G a p d e p th
C o ils
C h i p w id th
S end u st or
S o f tm a x r e g i o n s
B a c k p r o fi l i n g
T h e fe rrite h e a d
S p u tte r e d S e n d u s t r e g i o n
T h e M IG h e a d
The TS S he ad
Figure 84 Video head designs
The first modification is to cut away at the back of the head, where the
head gap is. This forces flux lines forwards to the front of the head.
The second modification is to introduce a different material with low
reluctance to the front face of the head, just where the head gap is. Flux
prefers to jump the gap at this material, rather than the ferrite behind.
Several material are used, often with exotic names to hide their true
composition. Materials like Softmax and Sendust are used. However
these materials tend to be softer than ferrite and therefore tend to wear
away quicker. Only a thin slither is placed on the head, and only close to
the head gap, rather than across the whole front face.
Sony Training Services 144

153. Broadcast Fundamentals
Automatic tracking
Another important technology that has been a vital part of modern
professional and broadcast video tape machines is the automatic
tracking playback head.
While servo systems using a control track were able to bring the video
heads, in particular the playback heads, close to the centre of the helical
tracks, there was a certain degree of error due to badly adjusted servo
electronics, or an imperfectly adjusted control head.
Another problem is even more annoying. The geometry of all helical
scan video recorders will play back correctly at play speed, because that
was how the tape was recorded, at play speed. If the machine is
speeded up slightly, slowed down, or paused altogether, the geometry
changes. Now the playback heads will not travel exactly up the centre of
the helical tracks, and will wander off track and possible cut across
adjacent helical tracks.
This is annoying for editors who regularly want to play back at other that
play speed, or pause the video machine altogether and look at one
frame or field on its own.
Automatic tracking video playback heads eliminate these problems.
Introduced by Ampex in 1977 as the Automatic Scan Tracking (AST)
system and by Sony in 1984 as the Dynamic Tracking (DT) system, both
systems relied on moving the playback heads to keep them in the centre
of the helical tracks.
Automatic tracking playback heads
Automatic tracking video playback heads use piezo-electric crystal
bimorphs The bimorph consisted of two piezo-electric crystals, bonded
together. When a voltage is applied across the bimorph one crystal
expands while the other contracts. This causes the bimorph to bend.
Reversing the voltage reverses the bend. One end of the bimorph is
fixed to the drum. The playback head is placed on the other end. In early
designs, including the Ampex AST designs, one bimorph was used. Two
bimorph are used in later designs, because this keeps the head itself
perpendicular with the tape surface.
One disadvantage with this kind of tracking system is that the bimorphs
require a high voltage to bend sufficiently. Any machine with automatic
tracking heads needs brushes and slip rings to transmit these high
voltages to the drum. Furthermore the brushes and slip rings must
maintain good contact and the drum contain smoothing circuitry. Any
intermittence in the supply to the bimorphs could generate
electromagnetic radiation that could be disastrous to the delicate record
and playback process.
Automatic tracking in operation
A small alternating voltage of about 450kHz is applied to the bimorphs
causing the heads to continually wobble. The wobble continually takes
the head slightly off track, causing a slight drop in the RF signal. The
servo system continually checks the level of the RF signal from the
145 Sony Broadcast & Professional Europe

154. Part 15 – The video tape recorder
heads, keeping the drops in RF as small as possible, by adding a DC
voltage to the wobble voltage.
Automatic tracking playback heads allow operators to change the
playback speed of a helical scan tape machine and still maintain a
steady picture. They have become an essential part of professional and
broadcast tape machines.
Tension regulation
It is vital that the tape tension around the drum is correct and maintained
within a small range. If the tape is too tight the video heads and tape will
wear out rapidly. If the tape is too loose the video heads will not be able
to maintain good contact with the tape surface and there is a risk that the
machine will throw tape loops or stick around the drum.
There are two types of tension regulator, the purely mechanical type and
the electromechanical type. Most domestic machines, cheaper and
smaller professional machines, use purely mechanical tension
regulators. They are simple, light and cheap.
All high-end professional and broadcast tape machines, especially those
intended for studio use, use electromechanical tension regulators.
Although they are generally heavier, more complex and more expensive,
they offer much finer tension regulation, and a change for the servo
system to monitor the tension regulation process. This in turn allows
machine to have different tension regulation response times for different
modes of operation and fault detection in case the tape sticks or breaks.
The principles behind good tension regulation
All tension regulators operate in the same basic way. During recording
or playback the capstan and pinch wheel pull the tape out of the supply
reel and around the drum. The take-up reel motor will apply a constant
pull on the tape. This pull is very light but it ensures that any tape that
has come through the capstan and pinch wheel is drawn into the take-up
reel in a tidy fashion.
The supply reel motor will be trying to resist tape from being drawn out
of the supply reel. This is what produces the tension. The higher the
resistance, the higher the tension.
Mechanical tension regulators
Mechanical tension regulators have a sensing arm with a roller on the
end of it, around which tape moves. The arm is connected to a spring
and to a friction belt which is wrapped around the supply reel table. If the
tape tension drops the spring will pull the arm further out, tightening the
friction belt around the supply reel table, increasing its resistance. The
capstan will continue to pull more tape out and the tension will increase.
Likewise if the tape tension increases, the arm will be pulled in against
the spring loosening the friction belt around the supply reel table, and
decreasing its resistance to allow tape out.
Sony Training Services 146

155. Broadcast Fundamentals
Mechanical tension regulators cannot handle loose tape by pulling it
back into the supply reel. This is because the tension regulator can only
stop the supply side reel motor, it cannot make it turn backwards.
Electromechanical tension regulators
Electromechanical tension regulators use a sensing arm with a roller on
the end of it, just as mechanical tension regulators. Likewise the arm is
connected to a spring, however the spring tends to be better quality, and
in some cases may be more that one spring to give a more accurate
response.
The arm will also have a strong magnet attached to it. One or more hall
effect detector are fixed to a circuit board either on the mechadeck or on
the tension regulator assembly. The hall effect detector will output a
signal corresponding to the position of the tension regulator arm. With a
properly aligned spring the position of the arm will also provide a
measure of the tension in the tape.
The signal from the hall effect detector is send to the machine’s servo
system which controls the supply side reel motor. Reel motors in this
kind of machine are more complex than those in machines with
mechanical tension regulators. The servo system is able to control the
direction, speed and amount or torque very precisely. Rather than using
friction the supply reel is effectively trying to turn backwards.
This ability to control the backward rotation of the supply reel motor also
allows electromechanical tension regulators to draw loose tape quickly
back into the supply reel.
Variation in tape path designs
Before the universal acceptance of cassettes for video tape recorders,
manufacturers designed several exotic tape paths that often required the
tape operator spend a while lacing up before the machine could be
used. Indeed we often take it for granted as we slam another cassette
into the machine that it was not always that easy.
Tape path designs have now settled to a just a few variations since the
introduction of cassettes, because the machine must be able to
automatically draw the tape out of the cassette before recording and
playing can take place, and put it neatly back into the cassette before it
is ejected. Any complicated lacing cannot be performed.
Terms of confusion
Various terms have been given to various tape wrap patterns, and there
appears to be a fair degree of confusion as to which one is which. The
term ‘omega wrap’ has been associated with many wrap pattern that are
not that similar.
Alpha wrap
The alpha wrap takes its name from the Greek letter ς . The tape passes
around the drum for a full 360 degrees. The wrap is sideways, with the
entrance and exit guides on the left or right. Alpha wrap would be very
147 Sony Broadcast & Professional Europe

156. Part 15 – The video tape recorder
difficult to achieve with cassettes because the tape passes over itself.
The tape must be manually laced. It is only used in machines with
spools. As an example alpha wrap is performed by the old Philips
EL3402 1” machine.
Omega wrap
Omega wrap takes its name from the Greek letter Σ . The tape passes
around the drum for almost 360 degrees. The active wrap is about 270
degrees. Although the term omega is used with many cassette machines
it is not actually possible to perform and true omega wrap with a
cassette. The tape must be laced. As an example, omega wrap is
employed by 1” C format machines.
C wrap
The wrap pattern is actually in the shape of an backward ‘C’ a little like
the Cyrillic letter ‘t ’. C wrap is possible and popular with cassette
machines. The tape is drawn from the cassette at one point and taken
between 200 and 300 degrees round the drum in an anti clockwise
direction, giving an active wrap of anything between 180 and 270
degrees. As an example, C wrap is popular with broadcast studio
machines using the Betacam SP and Digital Betacam formats and other
similar tape formats.
M Wrap
This is the most popular wrap pattern, and is used in cassette based
machines. Tape is drawn from the cassette at two points. It is drawn
round the left side of the drum, and round the right side of the drum, to
give a total wrap of between 250 and 300 degrees, and an active wrap
of anything between 180 and 270 degrees. As an example M wrap is
popular with domestic VHS machines and some broadcast machines
like the Sony D1 and D2 machines and the PVW range of Betacam SP
machines.
Definition of a good tape path
A perfect tape path with contain a perfectly circular supply and take-up
reel. The tape would move from the supply reel to the take-up reel in a
straight line without touching anything. Video and audio would be
recorded and played back without any heads touching the tape.
Clearly this is an impossibility! Compromises have to be made. The
record and playback heads must touch the tape. Furthermore helical
scanning requires that the tape be wrapped around the drum. Thus the
tape must change direction dramatically. Helical scanning also requires
accurate tape tension control.
The speed of the tape must be governed and regulated. Reel motors are
simply not good enough to accomplish this. A capstan is required.
Any item like a guide, drum or static head, cleaner of capstan changes
the tapes direction and adds friction. Spinning guides, and the drum
itself, are never absolutely central and always add a slight wobble to the
Sony Training Services 148

157. Broadcast Fundamentals
tape’s motion. There are therefore opportunities for the tape to stick, be
forced into the wrong position or the timing to be altered.
The important part of any video tape recorder tape path is the distance
between the supply side reel and the capstan. This is where all the
heads are and this is where the tape must be at the correct tension and
in the correct position. This length of tape should be as short as
possible, and should pass across as few items as possible.
Static heads, the cleaner, the drum, entrance and exit guides, the supply
side tension regulator and the capstan/pinch wheel are vital and
therefore always present. Designers will ensure that any other guides
will only be added to the design if they are absolutely necessary.
A perfect tape path does not need top touching or bottom touching
guides, or a rabbet on the lower drum. The tape would pass around the
various items on the mechadeck in exactly the correct position.
Designers calculate the angle of guides, drum, static heads, etc. so that
the tape runs smoothly through the tape path. Although it is impractical
to expect a perfect level of mechanical accuracy, the rabbet and any
guides touching the top or bottom of the tape should do so very lightly.
It is not very critical what happens to the tape after the pinch wheel and
capstan. The pinch wheel and capstan act as a wall, isolating the drum
and static heads from any minor wobbles in the tape afterwards.
Therefore the amount of tape, number of guides and other hardware is
not important.
The servo system
Modern video machines consist of a number of servo loops. Normally
one item is the master, and servo loops slave off the master.
When a video machine is playing back the master is the drum. It obeys
the incoming reference taking no regard of anything else, and spins at a
constant rate, somehow related to the reference itself. The drums in
early analogue machines spin at frame rate, 25Hz for PAL (625 line)
based machines and 29.97Hz for NTSC (525 line) based machines.
Later digital machine drums spin at a multiple of frame rate.
The rest of the servo system slaves off the drum. The first servo loop to
consider uses signals from the drum and the control head and uses the
capstan as a control. Although the machine’s servo system will control
the capstan to pull the tape through at almost a constant rate, signals
from the drum and the control head inform the servo system of the
relative position of the tape and the spinning drum. By slightly altering
the speed of the capstan the servo system will ensure that the timing
between the pulses from the drum and the control head is correct, thus
ensuring that each playback head finds the beginning of each helical
track.
Another servo loop slaves off the capstan servo loop. This loop uses a
signal from the tension regulator to control the supply reel, trying to
maintain the tension around the drum at a constant predefined level. In
simpler machines this is done mechanically. In more complex machines
this is done electronically.
149 Sony Broadcast & Professional Europe

158. Part 15 – The video tape recorder
In some machines there is also a servo loop between a take-up tension
regulator and the take-up reel, to maintain the take-up tension.
Analogue video tape recorder signal processing
Video tape recorder signal processing can be divided into a number of
distinct areas. The first division is between record and playback
processing.
The problems of recording to tape
The earliest pioneers of tape recoding technology discovered that it was
impossible to record an audio or video signal directly to tape and expect
a reasonable playback signal. Two characteristics ensure that the record
process was not going to be that straightforward, the basic behaviour
characteristics of inductors and magnetic hysteresis.
When a record head records a signal the current applied to the head
generates a flux which magnetises the tape. The strength and direction
of magnetisation is directly related to the current. When the playback
head plays the tape back, current is generated at the output of the head
that is proportional to the rate of change of magnetisation in the tape.
This term ‘rate of change’ is crucial. If a high DC signal is recorded to
tape, it will magnetise the tape a lot, but nothing will be played back,
because the rate of change is zero. Conversely if a smaller high
frequency signal is recorded to tape a large high frequency signal will be
played back because the rate of change will be high.
This characteristic is evident when looking at the control track of most
professional VTR’s. The control head records a 25Hz or 29.95Hz square
wave signal on tape. The resulting control track consists of positively
and negatively magnetised region. The resulting playback signal is a
series of large negative and positive spikes for each negative and
positive transition.
This ‘rate of change’ characteristic makes the record/playback process
non-linear, it integrates the record signal, and introduces a phase shift
between the record and playback sine waves.
The second characteristic is hysteresis. This defines the ‘memory’’
magnetic materials have. Apply a magnetic flux to a magnetic material
and it will remember this by becoming magnetised.
The answer to these problems is modulation. Modulation is the process
of combining a low and high frequency signal together into one signal.
There are two types of modulation, amplitude modulation (AM) and
frequency modulation (FM). AM involves changing the amplitude of the
high frequency signal with the low frequency signal. FM involves
changing the frequency of the high frequency signal with the low
frequency signal.
AM is the easier modulation system to design, and was used in the first
attempts at modulating the video signal before recording it to tape.
However FM is more resilient and was chosen as the modulation of
choice for video tape recorders.
Sony Training Services 150

159. Broadcast Fundamentals
Input processing
Reference input selection
Another important part of the input circuitry is the reference input. The
machine should be able to play a tape back on its own, maintaining good
and consistent timing. It should also be able to lock to an incoming
reference when playing back, locking the entire playback process to the
incoming reference. The machine should also be able to lock to an
incoming reference while recording, or lock to the incoming video signal
it is recording.
Therefore every tape machine will include a precision oscillator and sync
pulse generator (SPG). This module can either free run to provide a
good reference for the machine, or it can be genlocked to either an
incoming reference signal or video input. Part of the input processing will
include a switch to select which input will be directed into the oscillator
and SPG.
Video input processing
All video tape recorders have input circuitry. This is required to convert
the input video into a common form appropriate for recording on tape.
For instance, component video recorders require any video input be in
component form before any final encoding or modulation can occur prior
to recording to tape. Therefore input circuits will include a composite
decoder, or S-video decoder, and a selection switch to allow the
operator to select which type of input to record.
Input audio processing
As with video inputs, all video tape recorders will include switches,
equalisers noise reduction encoders (Dolby for instance) and even
provide microphone power, to convert and process the incoming audio.
Tape encoding
The video signal needs processing prior to recording to tape. This will
include FM. It may also include pre-emphasis to improve the recorder’s
ability to capture sharp transitions and detail in the image. It may also
include a small amount of AM after the FM to reduce the possibility of
over modulation problems that sometimes manifest themselves as
bearding on the playback image.
The audio signal needs little further processing other than standard bias
modulation, before being recorded to longitudinal tracks to tape.
Signal transfer to the drum
Once a decision had been made to use a rotating drum to increase the
relative head to tape speed, a way was need to transfer the video
signals onto the spinning drum and to the record heads. Wire connection
could hardly be used. They would very quickly tie themselves in knots
and wrench themselves free. Slip rings and brushes also presented
problems. It was impossible to maintain good enough connection.
151 Sony Broadcast & Professional Europe

160. Part 15 – The video tape recorder
The answer lies in the rotary transformer. This operates in a similar way
to a standard transformer with two windings (coils) sitting close to one
another. Current in the input coil produces a magnetic flux. As the
current changes the flux changes. The rate of change of this flux excites
a current in the output coil. If an AC signal is input to the input coil and
AC output will appear and the output coil, albeit phase shifted.
Rotary transformers have one coil build into the static lower drum and
the other built into the spinning upper drum. An RF signal can be
transferred from the lower to upper drum during recording, and from the
upper to lower drum during playback.
Modern video tape recorders have many rotary transformers for
transferring more than one signal onto and off the upper drum. This is
essential in component video machines where there is a separate path
for the luminance and colour signals. A separate transformer is often
also used to transfer switching information to the upper drum, so that the
drum itself can switch record or playback signals between different
heads either as the drum rotates, or for multi-format machines where
different playback heads are used to play back tapes from different
formats.
Output processing
One of the challenges facing designers of early tape recorders was how
to play the tape back with smooth consistent timing. The timing
requirements of a standard broadcast video signal are very accurate.
Video tape recorders are essentially mechanical. No matter how well the
tape machine is built, and no matter how good the servo system is, there
will still be a slight amount of mechanical wobble that will introduce huge
timing fluctuations compared to the timing accuracy required of
broadcast video signals. The answer lies in a clever piece of circuitry
called a timebase corrector which is explained in a separate section
below.
Signal transfer from the drum
RF signals from the playback heads are transferred off the drum through
the rotary heads described above. Head switching is required for those
machines with a drum wrap of less than 360 degrees. Many machines
use an active wrap of 180 degrees. Two sets of playback heads are
used, 180 degrees apart. The machine will switch between heads at
exactly the correct point, and thus maintain a continuous video signal
from the tape.
This switching can be performed on or off the drum. Switching on the
drum means that fewer rotary transformers are required to transfer the
video signals from the upper to lower drums. However an extra rotary
transformer is required to transfer switching information to the upper
drum.
Switching off the drum removes the need to transfer switching
information to the upper drum, but increases the number of rotary
transformers required to transfer the video signals off the upper drum.
Sony Training Services 152

161. Broadcast Fundamentals
Once the signal is off the drum it is buffered and equalised. There may
also be an automatic gain control (AGC) to automatically correct small
irregularities in the amplitude of the playback RF signal.
Output video and audio processing
The final piece of circuitry before the outputs processes the video and
audio signals to provide whatever outputs the machine is designed to
provide. Component analogue machines may include a composite
encoder for a composite output, or analogue to digital converters for
either digital audio or digital video outputs.
The timebase corrector (TBC)
Sitting between the playback equalisers and the final output processing
is the TBC. A TBC evens out the irregularities in timing of the signal
coming from the helical tracks of a video tape recorder. All TBCs do this
by storing a certain amount of the signal and then playing it out at a
constant rate.
Clock generation
An important part of any TBC is the ability to generate accurate clocks.
Clocks are used to write video into the store and out of it. The read clock
needs to accurately follow the timing irregularities in the signal coming
from tape. The write clock needs to be locked to the machine’s SPG
keeping constant smooth timing.
Of the two clocks the read clock present the greatest challenge. A
horizontal syncs detector sends the horizontal sync pulses off tape to a
timed monostables which output a voltage depending on the rate of the
sync pulses. The detector may also include and window discriminator
which will ignore any false horizontal syncs and half line pulses during
the vertical interval.
The signal from the timed monostables is fed into a voltage controlled
oscillator (VCO). The VCO is designed to run at the same rate as the
read clock when there is no signal from the timed monostables.
Timing irregularities in the signal off tape will increase or decrease the
horizontal sync rate. This will cause the control voltage from the timed
monostables to increase or decrease, shifting the VCO frequency up or
down.
Charge coupled device (CCD) delay line TBC
The first TBCs used CCD delay lines. These half analogue, half digital,
devices consist of a long line of cells. Each cell could contain an
analogue charge. A clock input transfers all the charges one cell towards
the end of the line. The input is connected to the first cell and the output
to the last cell.
CCD delay lines cannot input and output at the same time. Therefore
two delay lines are used, one for writing and the other for reading. They
are designed to store enough for one line of video. One line later and the
delay lines are switched. The one that was writing is now reading, and
153 Sony Broadcast & Professional Europe

162. Part 15 – The video tape recorder
visa versa. The clocks are also switched. The delay line that is writing
uses the write clock and the one that is reading uses the read clock.
Semiconductor TBC
All new TBC designs use semiconductor memory devices instead of
CCD delay lines. Semiconductor memory devices are totally digital.
They therefore require a digital input and give out a digital output. All
semiconductor memory TBC used in analogue video recorders use
analogue to digital converters at the TBC input and digital to analogue
converters at the output.
Dual TBC designs
All modern broadcast analogue video tape recorders record component
video, keeping the luminance and colour parts of the video signal
separate throughout the whole record playback process, even on tape.
Thus the luminance and colour playback signals experience their own
timing inconsistencies. Each signal must be timebase corrected
independently if quality is to be maintained.
A dual TBC has a separate horizontal sync detector, timed monostable
and VCO for luminance and colour. It also has two stores. The read
clock is the same for both luminance and colour.
Popular analogue video recording formats
This is by no means an exhaustive list. There are many analogue video
tape formats not mentioned here, that were only moderately successful
and others that were more of a failure.
Quadruplex (1956)
Ampex introduced the Quadruplex tape format, commonly known as
Quad. Quad is a professional 4 head transversal composite format. It
uses spools of 2” tape. Helical tracks were originally 10 mils wide, 33
minutes from vertical. Drum is just over 2” dia. spinning at 1,400 rpm for
the original NTSC machines.
The most long lasting of video tape formats. The Ampex VR-1000
machine was the first commercial video tape machine. There are still
archives of Quad tape, and it is still in use in a few places, although it
has been superseded for new recordings by other formats.
U Matic (1970)
Developed by JVC, Matsushita and Sony in 1971, sometimes called
Type E. U Matic is a professional 2 head helical scan composite format.
It uses spools of ¾” (19mm) tape. Helical tracks are 84um wide and 4.95
degrees from horizontal. Drum is 110mm dia. spinning at 1800 rpm
(1500 rpm for NTSC). U Matic will record 2 longitudinal audio tracks.
LTC was not designed in as a separate track to start with but was given
a dedicated track under the helical tracks. This meant that LTC had to
be recorded first and could not be re-recorded without over writing part
of the helical tracks. Later provision for VITC.
Sony Training Services 154

163. Broadcast Fundamentals
U Matic was a very successful format because of its wide user base,
from high end broadcast to professional and industrial use. Although not
as long lasting as Quad it was probably more popular. Eventually
available in higher quality SP form, in lo-band and hi-band modes.
Betamax (1975)
Developed by Sony. Betamax is a domestic 2 head composite helical
scan format using the colour under system. It uses cassettes containing
½” tape. Helical tracks are just over 30 um wide and 5.85 degrees from
horizontal. Drum is 74.487mm dia. spinning at 1800 rpm (1500 rpm for
NTSC). Betamax will record 1 longitudinal audio track and no timecode.
Betamax was head to head with VHS in the late 1970’s, but eventually
lost. Many reasons have been given for this. The reluctance of Sony to
licence the format. The reluctance of video rental firms to accept pre-
recorded Betamax tapes, the lesser record times and features of
Betamax machines.
1” type C (1976)
Developed by Sony and Ampex. C is a professional helical scan
composite format. It uses spools of 1” tape. Helical tracks are 5.1 mils
wide, and almost flat at 2.5 degrees from horizontal. The format uses a
large drum of 132mm dia spinning at 3600 rpm (3000 rpm for NTSC). C
will record 3 longitudinal audio tracks (4 in Europe) and LTC normally
recorded on the last audio track.
VHS (Video Home System) (1976)
Developed by JVC, adopted by many other manufacturer. VHS is a
domestic 2 head composite helical scan format using the colour under
system. It uses cassettes containing ½” tape. Helical tracks are 2.3 mils
wide in standard play mode, 1.15 mils wide in long play mode and 5.96
degrees from horizontal. Drum is 60.5mm dia. spinning at 1800 rpm
(1500 rpm for NTSC). VHS will record 2 longitudinal audio tracks and no
timecode.
VHS was head to head with Betamax in the late 1970’s, but eventually
won. VHS went on to become the most popular domestic and industrial
format.
Video 2000 (1979)
Developed by Philips and Grundig. Video 2000 is a domestic 2 head
composite helical scan format using the colour under system. It uses
cassettes containing ½” tape. Helical tracks are 22.6 um wide and 15
degrees from horizontal. Drum is 65mm dia. spinning at 1800 rpm (1500
rpm for NTSC). Video 2000 will record 2 longitudinal audio tracks and no
timecode.
In Europe Video 2000 was the ‘other domestic format’ while VHS and
Betamax battled for supremacy. It boasted automatic tracking, using
bimorphs similar to those used in professional machines, and dual sided
cassettes. However Video 2000 was never going to win against either
VHS or Betamax. While video rental firms were pushed to provide two
155 Sony Broadcast & Professional Europe

164. Part 15 – The video tape recorder
versions of each movie in their shops, VHS and Betamax, it was
inconceivable that they would supply three.
Betacam (1982)
Developed by Sony, sometimes called Type L. Betacam is a
professional 4 head component helical scan format. It uses cassettes
containing ½” tape. Helical tracks are 86um wide, and +15.25 degree
azimuth, for luminance and 72um, and -15.25 degree azimuth, for colour
tracks, and 4.679 degrees from horizontal. Drum is 74mm dia. spinning
at 1800 rpm (1500 rpm for NTSC). Betacam will record 2 longitudinal
audio tracks, LTC and VITC.
Betacam became a popular professional format using oxide tape similar
to that used by domestic Betamax. However Betacam was really just a
‘rehearsal’ for the improved version, Betacam SP, which became a
workhorse professional and broadcast video tape format.
8mm (1983) and Hi8 (1989)
Developed by a Japanese consortium. 8mm is a domestic 2 head
composite helical scan format using the colour under system. It uses
cassettes containing 8mm tape. Helical tracks are 20.6um wide 4.88
degrees from horizontal. Drum is 1.6” dia. spinning at 1800 rpm (1500
rpm for NTSC). 8mm will record 2 PCM audio channels and 2 AFM
audio channels and no timecode.
Hi8 is an enhancement of 8mm, developed by Sony, using metal oxide
tape
Both 8mm and Hi8 have gained reasonable success as a domestic
camcorder tape format.
Betacam SP (1986)
Developed by Sony, sometimes called Type L. An improvement over the
Betacam format Betacam SP, with the same format dimensions, uses
higher FM frequencies and metal instead of oxide tape. Betacam SP
introduced 2 AFM audio tracks inserted into the colour helical track
signal, providing the format with 4 audio tracks altogether.
The BVW-75 and BVW-75P machines became workhorse machines
within the broadcast industry selling thousands of machines and millions
of tapes worldwide.
M2 (1986)
Developed by Panasonic. M2 is a professional 4 head component helical
scan format. It uses cassettes containing ½” tape. Helical tracks are
44um wide for luminance and 36um for colour tracks, with a 15 degree
azimuth, and 4.29 degrees from horizontal. Drum is 76um dia. spinning
at 1800 rpm (1500 rpm for NTSC). M2 will record 2 longitudinal audio
tracks, 2 AFM audio tracks, LTC and VITC.
M2 was introduced as a competitor to Betacam SP and some
broadcasters adopted it as a standard. Although technically very similar
the machines gained a reputation for unreliability, probably due more to
Sony Training Services 156

165. Broadcast Fundamentals
spare parts availability and service rather that the machine’s reliability,
M2 did not gain the universal acceptance that Betacam SP gained.
S-VHS (1987)
Developed by JVC, adopted by every other manufacturer. VHS is an
enhancement of the VHS format with improved luminance bandwidth. It
gained popularity because of its compatibility with VHS.
Digital video tape recorders
Practical broadcast digital video recorders began to appear at the
beginning of the 1980’s with the publication of CCIR 601 in 1982 and
CCIR 656 in 1986. These two documents proposed a method if digitising
component video signals and conveying them in digital form over a
multicore cable. Sony designed the D1 video recorder specifically to
record CCIR-601 signals without any loss.
The original CCIR 601 document specified 8 bit samples. However the
CCIR 656 document also specified two spare data bits which were for
‘future development’. The industry grabbed these two spare bits using
then as ½ and ¼ resolution, increasing the samples sizes to 10 bits.
About the same time as the transition from 8 bits to 10 bits, there was
transition from the original multicore cable, parallel method of conveying
CCIR 601 data to a serial version using standard 75 ohm coaxial cable
and BNC connectors.
Sony and other manufacturers, notably Panasonic and Ampex, followed
over the years by producing broadcast quality digital video recorders to
record either 8 or 10 bit CCIR 601 samples, either entirely transparently,
or with compression.
Although digital video recorders have gained almost universal
acceptance in broadcast, the domestic and industrial markets continue
to use analogue tape formats, due to the overwhelming use of VHS, and
the introduction of analogue formats like Hi-8 which have sufficient
quality for most peoples’ needs.
DV has gained wide acceptance as a camcorder standard for domestic
use. The lack of any domestic DV television recorders has helped to
keep VHS as the only practical home television recording format.
It is unlikely that there will ever be a de-facto standard digital home tape
recording format. The imminent release of Blu-Ray optical disk recorders
will certainly now kill any change for a manufacturer introduce one.
The advantages of digital video tape recorders
Digital video recorders have a number of distinct advantages over
analogue machines. The first is the record transparency. A digital video
signal can retain all its quality through the record playback process. In
theory exactly the same digital data that is recorded to tape can play
back. Although this is not exactly true, it is certainly true that professional
digital video recorders allow video to be recorded, played back and re-
recorded many more times than is possible with analogue recorders.
This is important for editing and post production.
157 Sony Broadcast & Professional Europe

166. Part 15 – The video tape recorder
The second is robustness. Digital data can be protected with error
correction data far more easily than an analogue signal. Furthermore
digital data can be shuffled and scrambled before recording to tape. If
there is a large error on tape, either during recording or during playback,
due to, for instance, dust, the highly concentrated group of errors can be
diluted over a large amount of data, as a widely spread group of small
errors that can easily be corrected one by one.
Other advantages have become apparent over the years, as digital
video tape formats have developed, and with the introduction of
computers into the broadcast production chain.
Later digital video recording formats now offer long recording time, and
small tape sizes. They also offer the possibility of transferring digital data
directly to the IT world, without loss, where computer based non-linear
editors and effects processors can perform a wide range of previously
unavailable creative possibilities.
Digital video recorder or digital data recorder
It is important to remember that all digital video recorders are not
recording digital video, or audio. The data is always processed,
scrambled, shuffled, sometimes compressed, and has extra error
correction data added. What is actually recorded to tape is just digital
data and bears very little resemblance to the original video and audio it
came from.
Manufacturers are now using stripping the video and audio input and
output processing out of their digital video recorders to produce very
competent data recorders for the IT backup and archive markets.
Digital video recorder mechadeck design
There is no difference in the requirements of a digital video recorder
mechadeck from that of an analogue one. Digital video recorder
mechadecks differ more as a result of general developments in
mechadeck design rather than any special requirements. Most digital
video recorder mechadecks use a spinning upper drum and static lower
drum. They employ either M wrap or C wrap, and they all incorporate
sophisticated supply side tension regulation, and capstans on the exit
side of the drum.
A notable difference employed by the Sony Digital Betacam, D1 and D2
machines was the rotating mid drum assembly. The lower drum is static,
as normal, but these machines also have an upper drum is fixed to the
lower drum, leaving a narrow slot between the two. A mid drum
assemble spins inside and between the upper and lower drums, with the
record, playback and flying erase heads fixed to its circumference and
protruding through the slot to touch the tape. This technique is more
expensive but produces equal strain on the tape at every point round the
drum resulting in very straight helical tracks on tape.
From the start, broadcast digital video recorders have recorded audio as
digital data somewhere on the helical tracks. Although some digital
formats still retain a low quality longitudinal cue track, this development
has resulted in very high quality audio recording and the removal of
Sony Training Services 158

167. Broadcast Fundamentals
much of the mechadeck hardware required for analogue longitudinal
audio recording.
Some digital formats have even removed the need for a conventional
longitudinal control track transferring all the servo lockup to the helical
tracks. These formats only have one longitudinal heads on the
mechadeck, the full erase head.
Digital video recorder channel coding
Analogue video recorders use FM as a method of coding the video
signal before recording it to tape, and decoding it on playback, to
overcome the problems of recoding to magnetic tape. In general terms
this technique of coding and decoding is called channel coding. The tape
is the channel.
Digital video recorders cannot use FM. It is both inappropriate and
impossible considering the available bandwidth on tape and the required
recording bandwidth. Digital video recorders use a combination of Partial
Response type 4 (PR4 or PRIV) and Viterbi as a channel coding
scheme.
Popular digital video tape formats
D1 (1987)
Developed by Sony. D1 is a professional digital 4 head component
helical scan format. It records 8 bit CCIR 601 video data with no
compression. It uses cassettes containing 19mm tape. Helical tracks are
40um wide and 5.4 degrees from horizontal. Drum is mm dia. spinning at
rpm ( rpm for NTSC). D1 will record 4 audio channels on the helical
tracks, and one on a longitudinal track. It will also record LTC and VITC.
D1 is expensive, both for machines and tape cassettes, but is used in
post production where quality is the prime concern.
D2 (1989)
Developed by Sony. D2 is a professional digital 4 head composite helical
scan format. It records a digitised PAL or NTSC (depending on the
machine version) composite video signal with no compression. It uses
cassettes containing 19mm tape. Helical tracks are um wide and
degrees from horizontal. Drum is mm dia. spinning at rpm ( rpm for
NTSC). D1 will record 4 audio tracks on the helical tracks, and one on a
longitudinal track. It will also record LTC and VITC.
D1 is expensive, both for machines and tape cassettes, but is used in
post production where quality is the prime concern.
Digital Betacam (1993)
Developed by Sony. Digital Betacam is a professional digital 4 head
component helical scan format. It records 10 bit CCIR 601 video data
with DCT based compression at just over 2:1. It uses cassettes
containing ½” tape. Helical tracks are 24um wide and 5 degrees from
horizontal. Drum is 80mm dia. spinning at 4500rpm ( 5400rpm for
159 Sony Broadcast & Professional Europe

168. Part 15 – The video tape recorder
NTSC). Digital Betacam will record 4 audio tracks on the helical tracks,
and a low quality cue channel on a longitudinal track. It will also record
LTC and VITC.
Certain machines capable of playing back Betacam and Betacam SP
tapes.
Digital Betacam is cheaper than D1 but offers indistinguishable image
quality and 10 bit sample recording. It is widely used in post production
where quality is the prime concern. However the compression scheme is
closed and proprietary. Present broadcasters are now looking to output
the digital stream directly from the tape machine.
DV & Mini DV (1995)
A consortium of 10 companies agreed and created DV, sometimes
called MiniDV. DV is a domestic digital 4 head component helical scan
format. It records intraframe 4:1:1 or 4:2:0 compressed video data with
5:1 compression ratio. It uses cassettes containing ¼” tape. Helical
tracks are 18um wide and 9.18 degrees from horizontal. Drum is
21.7mm dia. spinning at 9000rpm. DV will record 2 audio tracks on the
helical tracks. Timecode is recorded on helical track as data (not VITC).
DV has become the most popular domestic digital video tape format,
and is available from a wide range of manufacturers. Camcorders and
decks offer direct compressed data outputs via the IEEE1394 interface,
otherwise known as FireWire (Apple) and iLink (Sony). Software
companies also offer good support for DV, with drivers and plug-ins for
DV data input to graphics, editing, and rendering software.
DVCPRO (1995)
Developed by Panasonic and based on the DV format. DVCPRO is a
professional digital 4 head component helical scan format. It records DV
data but with a wider track on metal evaporated tape to increase
robustness and quality. It uses cassettes containing ¼” tape. Helical
tracks are 18um wide and 9.18 degrees from horizontal with a +20.03
-19.97 degree azimuth. Drum is 21.7mm dia. spinning at 9000rpm.
DVCPRO will record 2 audio tracks on the helical tracks and 1
longitudinal cue track. It will also record LTC and VITC.
DVCPRO is the Panasonic professional DV format, and initially gained
wide acceptance due to its low price and compact design.
DVCAM (1996)
Developed by Sony and base on the DV format. DVCAM is a
professional digital 4 head component helical scan format. It records DV
data. It uses cassettes containing ¼” tape. Helical tracks are 15um wide
and 9.18 degrees from horizontal with a +20.03 -19.97 degree azimuth.
Drum is 21.7mm dia. spinning at 9000rpm. DVCPRO will record 2 audio
tracks on the helical tracks and 1 longitudinal cue track. It will also
record LTC and VITC.
DVCPRO is the Sony professional DV format. Introduced after DVCPRO
is lagged behind in popularity but is now beginning to gain widespread
Sony Training Services 160

169. Broadcast Fundamentals
support as an industrial format and for low budget television work.
Machines like the PD-150 have almost gained ‘classical’ status.
Betacam SX (1996)
Developed by Sony using the Digital Betacam mechadeck. Betacam SX
is a professional digital 4 head component helical scan format. It records
8 bit CCIR 601 video data with MPEG 4:2:2P@ML based compression.
Betacam SX uses IB frame compression to maintain broadcast quality at
18Mbps and 10:1 compression ratio. It uses cassettes containing ½”
tape. Helical tracks are 22um wide and 5 degrees from horizontal with a
15.25 degree azimuth. Drum is 80mm dia. spinning at 2250rpm
( 2700rpm for NTSC). Betacam SX will record 4 audio tracks on the
helical tracks. It will also record LTC and VITC.
Certain machines capable of playing back Betacam and Betacam SP
tapes.
Sony introduced a hybrid Betacam SX machine combining a
conventional tape mechadeck with hard disks. Compressed video and
audio material could be transferred to and from the tape and disks. This
allowed for linear and non-linear editing in one unit. However the hybrid
machine proved too complex for many and was not widely adopted.
Betacam SX was introduced as a replacement to Betacam SP and has
comparable digital quality. Widely used as a news gathering format.
However although the compressed digital stream is available at the
output for direct high speed transfer the compression scheme was not
ratified, with the standard authorise preferring 50Mbps data instead.
Sony responded with IMX.
Digital S (1996)
Developed by JVC and otherwise known as D9. Digital S is a
professional digital 4 head component helical scan format. It uses 4:2:2
sampling, like MPEG, making it technically better than DV and the same
as MicroMV. It uses cassettes containing ½” tape. Helical tracks are
20um wide.Digital S will record 4 audio tracks on the helical tracks and 2
longitudinal tracks. It will also record LTC and VITC.
HDCAM (1997)
Developed by Sony and based on the Digital Betacam mechadeck.
HDCAM is a professional digital 4 head component helical scan format.
It records high definition video data with mild 3:2 compression. It uses
cassettes containing ½” tape. Helical tracks are 22um wide and 5
degrees from horizontal with a 15.25 degree azimuth. Drum is 80mm
dia. HDCAM will record 4 audio tracks on the helical tracks and one
longitudinal cue track. It will also record LTC and VITC.
HDCAM was introduced as an alternative to film, and thus records
progressive 24 fps (24P), but can be switched to a number of television
based recording methods. HDCAM is expensive and exclusive, but
offers very high quality recording.
161 Sony Broadcast & Professional Europe

170. Part 15 – The video tape recorder
Machines due for release about the time this book is published will
include uncompressed high definition and machines based on the IMX
mechadeck.
DVCPRO 50 (1998)
Developed by Panasonic as an enhancement of the original DVCPRO
format, with 50 Mbps recorded data to comply with the requirements of
standards authorities. Machines can now be equipped with a IEEE1394
interface, allowing high speed transfer of 50Mbps DV data.
IMX (2000)
Developed by Sony using a new design of mechadeck loosely based on
the Digital Betacam mechadeck. IMX is a professional digital 4 head
component helical scan format. It records 8 bit CCIR 601 video data with
MPEG 4:2:2P@ML I frame only based compression at 50 Mbps. It uses
cassettes containing ½” tape. Helical tracks are 22um wide and 5
degrees from horizontal with a 15.25 degree azimuth. Drum is 80mm
dia. spinning at 4500rpm ( 5400rpm for NTSC). IMX will record 8 audio
tracks on the helical tracks. It will also record LTC and VITC.
Certain machines capable of playing back Betacam, Betacam SP and
Digital Betacam tapes.
IMX was introduced to comply with the standards authorities
requirement for a 50Mbps I frame only MPEG video recorder. Although
recording 8 bit samples ( a requirement of MPEG) IMX quality is
indistinguishable from Digital Betacam. However, unlike Digital Betacam,
the compressed stream is available at the output for direct transfer to
other machines, computer hard disk, or video servers.
A later modification, the E-VTR, allows video and audio material from
tape to be packaged and send directly out on a computer network cable.
Micro MV (2001)
Developed by Sony. Micro MV is a new format intended for the domestic
market. However it records true MPEG data on a tiny cassette, giving it
comparable, if not better, quality than DV. Although technically superior
than DV, MicroMV has a lot of work to do to gain any ground on DV and
DV based formats like DVCPRO and DVCAM. Software manufacturers
have still to offer the kind of support for MicroMV that DV enjoys.
)
Sony Training Services 162

171. Broadcast Fundamentals
Part 14 Betacam and varieties
Variations of the original Betacam format have dominated the broadcast
industry for the last 20 years. The same basic scheme is now used in
analogue and digital video recorders, high definition recorders and data
recorders.
The first of these broadcast machines recorded to ½” oxide tape
encased in a cassette. Two cassette sizes were made available, the
smaller being exactly the same size as the domestic Betamax tape, and
was suitable for portable devices and short programme content. The
larger was about twice the size. It had longer record time and was more
suitable for studio use.
Mechanics
All Beta formats are true helical scan. The drum assembly is about
81mm diameter and consists of two halves. The lower half is static and
acts simply as a support. The upper half is about 15mm thick and spins
horizontally. The whole assembly leans at about 5 degrees writing tracks
that are only about 5 degrees from the tape’s direction.
The tape wrap is a little over 180 degrees with the record/playback
heads fitted in pairs, on opposite sides of the drum. During recording
each head writes one track for 180 degrees. At the end of the track, just
before the head leaves contact with the tape, the record signal is
switched to the opposite head, which has just begun its 180 degrees
contact with the tape. This head then writes the next track.
The tape moves slowly through the machine, so that each track sits next
to the last.
With the original Betacam format each track writes one field of video.
Therefore PAL based machines have a drum that spins at 25Hz. Each
complete revolution of the drum records or plays back one complete
frame.
Several other tracks are recorded on the tape. There are four of these
and all are longitudinal. These tracks are recorded and played back by
two sets of static heads placed in the tape path just before and just after
the drum itself.
Two tracks run along the top edge and two along the bottom. The top
two tracks are responsible for audio channels 1 and 2. Channel 1 is on
the inside (bottom track). This is intentional. If a single channel is
recorded it is likely to be channel 1 and is therefore less likely to be
corrupted if the edge of the tape is damaged.
The bottom two longitudinal tracks are responsible for control and
timecode. The top track is responsible for control, the most important of
the two. If the bottom edge of the tape is damaged, timecode will be lost
and the machine will switch to the control track to keep timecode
counting, until a god timecode signal can be found again.
163 Sony Broadcast & Professional Europe

172. Part 16 – Betacam and varieties
The standard Betacam tape path
Even though the internal structure of Betacam tape machines may differ
from one machine to another, the basic tape path is identical in all
machines. It has to be, if one is able to take a tape recorded in one
machine and play it back in another.
Supply reel and tape cleaner
Tape exits from the left hand cassette reel, commonly called the supply
reel. Most machine have a tape cleaner. This is a blade, made either
from steel of artificial sapphire and cleans any debris off the tape before
the machine attempts to record or play back the tape.
Supply side tension regulator
The tape then passes around a tension regulator, often called the supply
side tension regulator. This important device measures the tape tension
on the whole supply side of the machine, including the drum, and sends
a signal back to the supply side reel to either let more tape out, if the
tension is too high, or hold the tape back, if the tension is too low.
It is very important that the tension around the drum is correct. Too tight
and both the tape and drum will wear out quickly. Too loose and head to
tape contact will be broken with resulting loss in recording and playing
back.
Full erase and control heads
Now the tape passes across a static head responsible for erasing the
whole tape when a crash record is being performed. This head blasts
the tape with a strong alternating magnetic field deleting anywhing
previously recorded on it. The tape then passes across another static
head responsible for recording and playing back the control track, the
control head.
The exact position of the erase head is not important. As long at it is
before the control head, it really does not matter. The position of the
control head is exact, and the same on every Betacam machine.
The drum, entrance and exit guides
Now the tape runs across an entrance guide, round the drum for at least
180 degrees, and leaves the drum to run across an exit guide. The lower
drum has a small step or ledge milled in its surface, called the rabbet.
This rabbet is at the top of the lower drum at the entrance side and
slopes down towards the exit side. The tape rests on the rabbet. This,
and the slope of the whole drum causes the helical motion of the drum
heads.
The entrance and exit guides have flanges that touch the top of the tape,
holding the tape down so that is enters and exits the drum at exactly the
correct point.
The upper drum spinning anti-clockwise at 25Hz (for PAL machines)
draws air between the tape and the drum itself. If the tape tension is
Sony Training Services 164

173. Broadcast Fundamentals
correct this makes a cushion of air between the two, and the heads
protrude from the surface of the drum, penetrating this cushion to touch
the tape.
Figure 85 The basic Betacam tape path
165 Sony Broadcast & Professional Europe

174. Part 16 – Betacam and varieties
Audio/timecode head stack
The tape now passes across another pair of static heads. The first is
responsible for erasing the two longitudinal audio tracks and the
timecode track. The second is responsible for recording and playing
back the audio and timecode tracks.
Some Betamax have a third static head that is responsible for playing
back the audio and timecode signals while in record mode. This so-
called confidence mode gives the operator confidence that the audio and
timecode has been recorded correctly.
The exact position of the audio/timecode stack is critical to ensure
proper syncronisation between the video, audio and timecode.
Capstan and pinch wheel
The head now passes between the capstan and pinch wheel. The pinch
wheel is a small soft rubber cylinder. The capstan is a precision motor
with a shaft about 5mm diameter sticking out of the top of it.
Normally there is a gap between the pinch wheel and the capstan shaft.
However during recording or playing back a solenoid pushes the pinch
wheel against the capstan shaft squeezing the tape between the two. As
the capstan motor turns it pulls the tape through at a steady speed.
The control track signal is passed into the machines computer where it is
processed and converted into control signals to adjust the speed of the
capstan motor so that the tape is passing round the drum as the correct
speed and position.
Take-up side tension regulator
The tape then makes its journey back into the cassette and onto the
take-up reel. Some machines include a tape-up tension regulator, to
measure the take-up tension and pass a signal back to the take-up reel
motor, to ensure that the tape is reasonably loose but not sloppy.
Other guides
The tape path will include a series of other guides along the tape path.
Some of these touch the top of the tape, some the bottom and some
neither.
Definition of a good tape path
The important part of a ½” tape path is the distance between the supply
side reel and the capstan. This is where all the heads are and this is
where the tape must be at the correct tension and in the correct position.
This length of tape should be as short as possible, and should pass
across as few items as possible.
Any item like a guide, drum or static head changes the tapes direction
and adds friction. Spinning guides, and the drum itself, are never
absolutely central and always add a slight wobble to the tape’s motion.
There are therefore opportunities for the tape to stick, be forced into the
wrong position or the timing to be altered.
Sony Training Services 166

175. Broadcast Fundamentals
Static heads, the cleaner, the drum, entrance and exit guides, and the
supply side tension regulator are vital and cannot be removed. However
any other guides should only be added to the design if they are
absolutely necessary.
A good tape path design will therefore have a short length of tape
between the supply side reel and the capstan, and as few extra guides
as possible.
What happens to the tape after the capstan is not at all critical. The
amount of tape, and number of guides between is not important.
Electronics
The basic electronics elements of a Betacam VTR consists of two
halves, the audio/video circuitry and the control circuitry. The
audio/video circuitry can be further divided into two halves, the record
circuits and the playback circuits, finally these two halves can be
subdivided into audio and video circuitry.
Control circuitry
The control circuitry is responsible for taking control signal from the VTR
keyboard and from any remote control ports at the back of the machine,
and converting them into control signals for the mechanics.
Control circuitry is also responsible for recording the control and
timecode tracks. The control track has a 25Hz signal recorded to it. This
signal is used during playback to ensure that the tape is sitting in the
correct place relative to the spinning drum, so that the playback heads
are moving directly up the centre of the helical tracks on tape.
Audio/video circuitry
If the composite input is used, the video recording circuitry decodes this
input into three component signals, Y, (R-Y) and (B-Y). It then combines
the two colour difference signals (R-Y) and (B-Y) into one signal using a
compressed time division multiplexing (CTDM) technique.
The Y and CTDM signal are then emphasised and modulated onto FM
carriers. Special horizontal sync signals are added before the signals are
sent to the record heads for recording to tape.
The video playback circuitry takes the modulated Y and CTDM signals
from the tape, checks for tape drop-out and extracts a clock from the
horizontal syncs signals.
The signals are then demodulated and de-emphasised. The clock is
used to perform timebase correction on the signals before the CTDM
signal is broken down into individual (R-Y) and (B-Y) signals.
If needed the resulting component signals are combined to make a
composite output. If not the Y, (R-Y) and (B-Y) signals are output as
component signals.
Record audio circuitry takes the two incoming analogue channels and
passes them through a Dolby noise reduction system before recording
them directly to the two longitudinal tracks at the top edge of the tape.
167 Sony Broadcast & Professional Europe

176. Part 16 – Betacam and varieties
In playback the two audio signals off tape are passed through the same
Dolby noise reduction system and directly out to analogue connectors.
Betacam video record techniques
The normal horizontal sync is replaced by a large tri-level pulse. The
CTDM colour signal has no horizontal sync pulse. So a large negative
going pulse is added so that timbase correction can be performed on the
CTDM signal independently.
Prior to modulation the signals are emphasised. During playback the
signals will be de-emphasised again. This improves the signals to noise
ratio of the whole record/playback path.
Betacam uses frequency modulation to record the Y and CTDM signals
to tape. This is a form of channel coding. The signals themselves would
not playback correctly if they were recorded to tape without any
modulation.
The FM modulation signal has a straight line sloped frequency response
that drops to zero and about 15MHz. Again this improves the signal to
noise ratio.
The Y and CTDM signals are modulated onto their respective FM
carriers so that the most positive and negative excursions on these two
signals fits between two specific FM carrier frequencies.
Betacam video playback techniques
The video playback circuitry is more complex than the record circuitry.
Even with all the precise mechanical engineering of a Betacam
mechadeck it impossible to playback a perfectly timed signal.
Tape speed fluctuations, drum speed fluctuations, rotary head impact
and tape tension fluctuations all serve to alter the exact timing of the
video signal as it comes off the tape.
Dirt and debris can get between the tape and the rotary heads, as the
drum is spinning. The tape may also be damaged, old or just bad quality.
All these factors can prevent a signal from being recorded to tape, or
prevent a good signal on tape from being played back. This is called
drop-out
The playback circuitry must correct any playback signal timing
fluctuations and somehow hide drop-out.
The timebase corrector
An important part of the playback circuitry of any Betacam VTR is the
timebase corrector, This piece of electronic circuitry smoothes out timing
fluctuations in the video playback signals.
The timebase corrector works by storing a small amount of the signal,
holding it for a short while, and then releasing it smoothly. This is a little
like using a bucket to provide a smooth water flow from a fluctuating
water source. As with the bucket analogy a certain amount of the signal
must be stored to allow for flucturations. Hopefully the fluctuations are
not so great as to either completely fill or empty the store.
Sony Training Services 168

177. Broadcast Fundamentals
Timebase correctors normally use a semiconductor memory as its store.
Thus the analogue playback signal must be passed through analogue to
digital converter before the store, and through a digital to analogue
converter afterwards.
169 Sony Broadcast & Professional Europe

178. Part 17 – The special effects machine
Part 15 The video disk recorder
History
The idea of recording video to a disk has existed about as long as those
of tape. However, whereas tape technology was technically possible,
with good recoding times and playback quality, technology was too
crude to allow comparable machines to be built for many years after
tape recorders became popular.
Ampex designed and built a prototype disk recorder in 1965. It became a
commercial product in 1967. Called the HS-100 it used an open hard
disk and allowed just 30 seconds of analogue video to be recorded. It
was used mainly for instant and slow motion replay.
However it was not until the 1990’s that disk recorders started to
appears that had a real practical use in broadcast. Abekas produced the
A-64 component video recorder which recorded parallel CCIR-601 digital
video on two large hard disks. The total storage time was a little under 1
minute, but it was full uncompressed broadcast quality.
At the time the Abekas disk recorders were one of only a few methods of
manipulating full uncompressed digital video in a non linear fashion, and
they became popular in post production and any high quality complex
short form editing.
The secret to Abekas’s success was their ability in modifying the hard
disks available at the time to make them do things that they would not
otherwise be able to do. Hard disks generally had integrated controllers
mounted on them that acted as an interface between the outside world
and the disks themselves. Abekas bypassed this to allow the video data
to be recorded directly to the disk platters.
Abekas built a business based to a great extent on their ability to modify
standard hard disk technology, and successfully sold a variety of hard
disk digital video recorders to post production facilities, advertising
companies, etc..
As hard disk technology improved, it became less necessary to bypass
the disk controller. There was less need for the kind of specialist
techniques employed by Abekas. Companies could produce video disk
recorders by using standard hard disks.
Hard disks also became cheaper. It became possible for companies to
offset the low bandwidth and speed of standard hard disks by simply
designing in more that one hard disk. Systems became available that
used an array of disks to both spread the bandwidth and increase the
capacity.
Present day
Video disk recorders can now be grouped into two areas, although some
products are able to cross the grey area between.
Sony Training Services 170

179. Broadcast Fundamentals
Transmission servers
The first group are the video disk recorders intended specifically for
transmission. These machines trade quality for storage capacity. The
whole system may require enough storage capacity for a day, or several
days of transmission. It may be required to supply many channels of
transmission. However the quality need not be supreme. These servers
can employ high compression ratios, with low data rates.
Reliability is very important in these servers. They need to work all the
time without fail. Therefore this kind of server often employs a high
degree of redundancy technology and hot swappable elements, like hard
disks, controllers and power supplies.
Transmission servers also do not need to be able to perform many
‘tricks’. They are intended to play, and perform some simple real-time
switching. Nothing else. So while remote control will need to be accurate
and fast, it will not need to be very versatile.
Production servers
This kind of video disk recorder has opposite requirements to those of
transmission servers.
Production servers need only have capacity for the programme that is
being worked on – far lower than the requirements in transmission
servers. Generally the material held on them is not the master, but the
work-in-progress material. Regular backups are normally performed.
Absolute reliability is les important than in transmission servers.
However production servers must maintain quality. As video is edited,
copied from one location to another, and generally fiddled with, there
must be no loss in quality. Any loss in quality would accumulate through
each edit generation until it becomes noticeable.
Production servers also need to be ‘clever’. Operators will need to be
able to be able to perform complex edits, and move, copy, and cut
material on the hard disk as though they were using a word processor.
Remote control ports need to be both fast, accurate and versatile
This kind of server may also need to be accessed by several users at
the same time. Therefore there may need to be more than one remote
control port fitted.
171 Sony Broadcast & Professional Europe

180. Part 17 – The special effects machine
RAID technology
An important technology for video disk recorders is RAID (redundant
array of inexpensive, or independent, drives). RAID consists of an array
of disks, and a RAID controller. The device (normally a computer) that is
accessing the array will see one logical drive. The RAID controller acts
as an interface, organising and sending data backwards and forwards
between the device and the array.
History
In 1988 Professor Randy Katz, while at Berkeley University in California,
was chiefly responsible for writing a paper entitled “A case for
Redundant Array of Inexpensive Disks”. This paper became the model
for disk array design. It specified five RAID levels, 1 to 5. These levels
define how the disks are logically arranged and how the data, and any
error correction codes, are spread across them.
Since the paper’s publication further levels have been designed. The
most important of these is levels 0 and 6, both of which have gained
general adoption in the industry. Level 7 was later added by Storage
Computer Corporation. Although proprietary is has gained general
acceptance because the company is a major producer of RAID solutions
and level 7 offers some real benefits.
Other levels have been added more recently. These are all combinations
of the existing levels, and have all been added for marketing reasons.
Reasons for RAID
Disk arrays have the benefit of increasing capacity. The “inexpensive”
part of RAID becomes important. The raid controller makes it appear
that there is one big expensive disk drive where there are actually many
smaller cheaper drives instead.
Another important reason for RAID is to increase the performance of the
array. The “redundant” part of RAID is not important in this case. Indeed
in some forms of RAID there is actually no redundancy at all. Using an
array simply makes it look as though one very fast disk has been
installed.
Redundancy
The basic idea of RAID was originally designed to ensure that extra data
is written to the disk array. This so-called redundant data could be used
if any of the data had errors in it during read operations.
The simplest redundant data is to make a complete copy of the data to
another set of disks. However it became obvious that a complete copy
did not have to be made. Instead some kind of error correction data
could be written. These codes generally took up less space that the
original data. There are two kinds of error correction codes used, Parity
codes, dual parity codes and Hamming codes.
Sony Training Services 172

181. Broadcast Fundamentals
Parity
Parity is a simple one bit code applied to a byte, word, or block of data.
There are two kinds of parity, even and odd. With both kinds the number
of “1”s on the data is added together. With even parity the parity code is
a “1” is there is an even number of “1”s in the data and a “0” if there is an
odd number of “1”s. With odd parity the logic is reversed.
Parity codes are not very powerful. They can only detect 1 bit errors. It is
perfectly possible for more than 1 bit to be wrong, and the parity code to
still be correct.
Hamming codes
Hamming codes (names after the inventor) are multi-bit codes derived
from the data, that can be used to reconstruct the data if it is read back
with errors. Specific data pattern produce specific Hamming codes.
Hamming codes are more powerful than parity codes. They can detect
and correct 1 bit errors, and detect 2 bit errors.
Dual parity codes
Dual parity codes are an enhanced version of simple 1 bit parity codes.
With these codes many bytes, words or blocks of data are grouped to
give a two dimensional array. Parity codes are generated in two
dimensions, for the array columns and rows.
Dual parity codes are more powerful that simple parity codes because
parity checks can be applied in two dimensions, detecting and correcting
a greater number of possible errors. They take up more space that
simple parity codes but a lot less space than Hamming codes.
Dual parity codes have been incorrectly called Reed Solomon codes.
However Reed Solomon codes are not single bit codes but multi-bit
codes. They are also generated from a complex polynomial algorithm
giving very powerful error correction capability. Dual parity codes cannot
achieve the error correction capability of Reed Solomon codes, but take
up far less space.
RAID levels
Level 0 (Disk striping)
Data is written in blocks, in sequence, to each disk in turn. Not really
RAID. Specifically designed for high performance, with increased
bandwidth and performance, and no redundancy.
Advantages : High bandwidth and performance. Simple design.
Disadvantages : Not a true RAID. No error correction (other than the
individual drives’ own internal error correction).
Level 1 (Mirroring)
Exact copy of each disk written to another disk. Sometimes achieved
within the computer software for simplicity, but this loads the computer
resources. Best achieved within the RAID controller instead.
173 Sony Broadcast & Professional Europe

182. Part 17 – The special effects machine
Advantages : Best error protection. Increased read performance. Good
for multi-user environments.
Disadvantages : Expensive (twice the hard disks).
Level 2 (Bit level disk striping & Hamming code disks)
With this level data is striped at bit level across multiple disks. Hamming
codes are generated and written to a separate disk or disks.
Hamming codes are multi-bit error correction codes. Although more
powerful they take up more space than simple 1 bit parity codes.
Therefore Hamming codes need more disk space making level 2 disk
requirements closer to level 1.
Level 2 is a dead RAID level. None of the RAID suppliers supports this
level. It is said that level 2 is not used because it requires special disks.
This argument comes from the fact that standard hard disks have their
own internal error correction, and that if you are using Hamming codes
the disks themselves need to be non-standard, with no internal error
correction.
However any RAID error correction is applied before the data is written
to disk. The disk’s internal error correction just adds another level of
security underneath anything applied by the RAID system.
In truth, level 2 is probably not used because the internal error correction
provided by present day disks, with their overall level of reliability, is
good enough that Hamming code protection supplied by the RAID
system would be more protection than is necessary considering the
extra capacity required. Simple parity codes are generally sufficient.
Advantages : Very good error protection.
Disadvantages : Dead. High ratio of ECC disks to data disks.
Level 3 (Byte level disk striping & parity disks)
RAID level 3 stripes across disks at the byte level rather than at the bit
level. 1 bit parity codes are written to a separate disk or disks.
Similar to level 2 but parity codes are smaller than Hamming codes and
take up less disk space.
Advantages : Good error protection. Low ratio of ECC disks to data
disks. Good for small and scattered file read/writes.
Disadvantages : Error protection not as powerful as levels 1 and 2.
Inefficient handling of large sequential files. Single parity drive is a
performance bottleneck.
Sony Training Services 174

183. 175
D a ta b its
1 ,5 ,9 ,1 3 ,1 7 ,2 1 ...
D a ta b lo c k s D a ta w o r d s D a ta b lo c k s
1 ,5 ,9 ,1 3 ,1 7 ,2 1 ...
Figure 86
1 ,5 ,9 ,1 3 ,1 7 ,2 1 ... D a ta b lo c k s D a ta b lo c k s 1 ,5 ,9 ,1 3 ,1 7 ,2 1 ...
1 to 1 0 2 4 . 1 to 1 0 2 4 .
D a ta b its
2 ,6 ,1 0 ,1 4 ,1 8 ,2 2 ...
D a ta b lo c k s D a ta w o r d s D a ta b lo c k s
2 ,6 ,1 0 ,1 4 ,1 8 ,2 2 ... D a ta b lo c k s D a ta b lo c k s 2 ,6 ,1 0 ,1 4 ,1 8 ,2 2 ... 2 ,6 ,1 0 ,1 4 ,1 8 ,2 2 ...
1 0 2 5 to 2 0 4 8 . 1 0 2 5 to 2 0 4 8 .
D a ta b its
R A ID C o n tr o lle r
3 ,7 ,1 1 ,1 5 ,1 9 , 2 3 ...
D a ta b lo c k s D a ta w o r d s D a ta b lo c k s
R A ID C o n tr o lle r
3 , 7 ,1 1 ,1 5 ,1 9 ,2 3 .. .
R A ID C o n tr o lle r
D a ta b lo c k s D a ta b lo c k s 3 ,7 ,1 1 ,1 5 ,1 9 ,2 3 ...
R A ID C o n tr o lle r
3 ,7 ,1 1 ,1 5 ,1 9 ,2 3 ...
R A ID C o n tr o lle r
2 0 4 9 to 3 0 7 2 2 0 4 9 to 3 0 7 2
D a ta b its
4 ,8 ,1 2 ,1 6 ,2 0 ,2 4 ...
Broadcast Fundamentals
code
D a ta b lo c k s D a ta w o r d s D a ta b lo c k s
H a m m in g
4 ,8 ,1 2 ,1 6 ,2 0 ,2 4 ...
g e n e ra to r
4 ,8 ,1 2 ,1 6 ,2 0 ,2 4 ... D a ta b lo c k s D a ta b lo c k s 4 ,8 ,1 2 ,1 6 ,2 0 ,2 4 ...
3 0 7 3 to 4 0 9 6 3 0 7 3 to 4 0 9 6
P a r ity
P a r ity
g e n e ra to r
g e n e ra to r
H a m m in g c o d e s
1 ,3 ,5 ,7 ,9 ,1 1 ,1 3 ...
P a r ity c o d e s P a r it y c o d e s
1 ,2 ,3 ,4 ,5 ,6 ,7 ,8 ... 1 ,2 ,3 ,4 ,5 ,6 ,7 ,8 ...
Level 0 Level 1 Level 2 Level 3 Level 4
H a m m in g c o d e s
2 ,4 ,6 ,8 ,1 0 ,1 2 ,1 4 ...
D a ta b lo c k s D a ta b lo c k s
1 ,6 ,1 1 ,1 6 ,2 1 . .. 1 ,6 ,1 1 ,1 6 ,2 1 ...
P a r ity c o d e s D u a l p a rity c o d e s D a ta b lo c k s
1 .6 .1 1 .1 6 .2 1 . .. 1 .6 .1 1 .1 6 .2 1 ... 1 ,5 ,9 ,1 3 ,1 7 ...
D a ta b lo c k s
C ache
1 , 5 ,9 ,1 3 ,1 7 ,2 1 ... D a ta b lo c k s
D a ta b lo c k s
1 to 1 0 2 4
D a ta b lo c k s D a ta b lo c k s 1 ,5 ,9 ,1 3 ,1 7 ...
2 ,7 ,1 2 ,1 7 ,2 2 ... 2 ,7 ,1 2 ,1 7 ,2 2 ...
P a r ity c o d e s D u a l p a rity c o d e s
2 ,7 ,1 2 ,1 7 ,2 2 ... D a ta b lo c k s
2 ,7 ,1 2 ,1 7 ,2 2 ... 2 ,6 ,1 0 ,1 4 ,1 8 ...
D a ta b lo c k s
2 , 6 ,1 0 ,1 4 ,1 8 ,2 2 .. .
D a ta b lo c k s
D a ta b lo c k s
1 0 2 5 to 2 0 4 8
D a ta b lo c k s 2 ,6 ,1 0 ,1 4 ,1 8 ...
D a ta b lo c k s
3 ,8 ,1 3 ,1 8 ,2 3 ... 3 ,8 ,1 3 ,1 8 ,2 3 ...
P a r ity c o d e s D u a l p a rity c o d e s D a ta b lo c k s
3 ,8 ,1 3 ,1 8 ,2 8 ... 3 ,8 ,1 3 ,1 8 ,2 8 ... 3 ,7 ,1 1 ,1 5 ,1 9 ...
D a ta b lo c k s
3 , 7 ,1 1 ,1 5 ,1 9 ,2 3 .. .
A s y n c h ro n o u s
D a ta b lo c k s
R A ID C o n tr o lle r
2 0 4 9 to 3 0 7 2 D a ta b lo c k s
R A ID C o n t r o lle r
R A ID C o n tr o lle r
R A ID C o n tr o lle r
3 ,7 ,1 1 ,1 5 ,1 9 ...
R A ID C o n tr o lle r
D a ta b lo c k s D a ta b lo c k s
4 ,9 ,1 4 ,1 9 ,2 4 ... 4 ,9 ,1 4 ,1 9 ,2 4 ...
P a r it y c o d e s . .. D u a l p a rity c o d e s ...
D a ta b lo c k s
4 ,9 ,1 4 ,1 9 ,1 4 .... 4 ,9 ,1 4 ,1 9 ,1 4 ....
D a ta b lo c k s 4 ,8 ,1 2 ,1 6 ,2 0 ...
4 , 8 ,1 2 ,1 6 ,2 0 ,2 4 .. . D a ta b lo c k s
3 0 7 3 to 4 0 9 6 D a ta b lo c k s
P a rity
D a ta b lo c k s 4 ,8 ,1 2 ,1 6 ,2 0 ...
g e n e ra to r
D a ta b lo c k s
P a rity
5 ,1 0 ,1 5 ,2 0 ,2 5 ... 5 ,1 0 ,1 5 ,2 0 ,2 5 ...
g e n e ra to r
g e n e ra to r
D u a l p a rity c o d e s
D u a l p a rity
P a r ity c o d e s
5 ,1 0 ,1 5 ,2 0 ,2 5 ... 5 ,1 0 ,1 5 ,2 0 ,2 5 ...
P a r ity c o d e s S t r ip e d s e t M irro re d s e t
1 , 2 ,3 ,4 ,5 ,6 ,7 ,8 ...
Level 5 Level 6 Level 7 Level 10 Level 10
C o m p u te r to C o m p u te r to
D a ta c o n n e c tio n R A ID c o n tr o lle r D a ta c o n n e c tio n
R A ID c o n tr o lle r c o n n e c tio n
D a ta P a r it y D ual H a m m in g c o n n e c tio n
codes p a r ity codes
codes
RAID levels
Sony Broadcast & Professional Europe

184. Part 17 – The special effects machine
Level 4 (Block level disk striping & parity disks)
RAID level 4 is similar to level 3 except that the data is striped across
the disks in blocks rather than in bits. Block reads and writes tend to
increase the overall performance over level 3 for large and sequential
file read and write operations.
Advantages : Good error protection. Low ratio of ECC disks to data
disks. Good for large sequential file read/writes.
Disadvantages : Error protection not as powerful as levels 1 and 2.
Inefficient handling of small files. Single parity drive is a performance
bottleneck. Seldom used.
Level 5 (Block level & parity disk striping)
This is the most popular RAID level. It is very similar to level 4, except
that the parity codes are not written to a separate disk. All the parity
codes are striped across the same disks as the data. This improves the
performance over RAID levels 3 and 4 by removing the bottleneck
associated with the separate ECC disk or disks.
However, because the data and parity codes are scattered over all the
disks, it is difficult to rebuild a new drive if one of the drives in the array
fails.
Advantages : Higher performance than level 4. Good error protection.
Low ratio of ECC codes to data. Good for large sequential file
read/writes.
Disadvantages : Error protection not as powerful as levels 1 and 2.
Inefficient handling of small files. Complex controller design. Complex
rebuilds.
Level 6 (Block level & dual parity disk striping)
This is a similar scheme to level 5. It uses block level read and write
operations, and spreads the parity across the disks rather than writing all
the parity codes to a separate disk or disks. However level 6 processes
blocks of data and produces 2 parity codes, one set for the columns and
another for the rows.
This increase in the amount of parity code generation greatly increases
the complexity of the RAID controller, and decreases the overall
performance of the array.
Advantages : Very good error protection. Low ratio of ECC codes to
data (but not as low as levels 2-5). Good for large sequential file
read/writes.
Disadvantages : Poor performance due to dual parity calculations.
Inefficient handling of small files. Very complex controller design.
Level 7 (Asynchronous cached data & parity striping)
RAID Level 7 is a proprietary technology from Storage Computer
Corporation. It borrows ideas from levels 3 and 4, but incorporates a
large memory cache between the disk array and the controller. The
Sony Training Services 176

185. Broadcast Fundamentals
controller uses the cache to read and write data to the disks
asynchronously. This means that each disks in the array can operate
independently, greatly improving overall performance.
This extra workload mean that RAID level 7 controllers are complex.
Power failures stand a greater chance of data loss, because data
spends time in the cache. If power is lost the cache is emptied.
Advantages : Very good performance for any file type. Very good error
protection. Low ratio of ECC codes to data.
Disadvantages : Proprietary design. Very complex controller design.
Expensive. Possible loss of data during power failure.
Level 10 (Striped array with mirroring)
Level 10 is a combination of level 0 and level 1. This is not a true
standard and there are different definitions of exactly what level 10 is,
some which are no different to level 0+1. The most consistent definition
is that part of the array is a striped set and part a mirror set. This
combines the advantages of both sets.
Advantages : Simple design. Same error correction capability as level
1. Good performance. Low ratio of ECC codes to data.
Disadvantages : Not efficient. Not a rigid standard. No ECC data.
Level 0+1 (Mirrored array with striping)
Level 0+1 is another non-standard array definition. However level 0+1
definitions appears to be more consistent than level 10.
This level has two complete striped sets. It is a level 1 implementation of
a level 0 array. It therefore has some of the advantages and some of the
disadvantages of each level.
Level 0+1 is not efficient but have good error protection, like level 1. Is
also has good bandwidth, like level 0.
Advantages : Simple design. Same error correction capability as level
1. Good performance. Low ratio of ECC codes to data.
Disadvantages : Not efficient. Not a rigid standard. No ECC data.
Other levels
There are a variety of other levels specifies within the storage industry
that attempt to combine advantages of the established levels.
Level 30 is a level 0 array where each stripe is a level 3 array. Level 50
is a level 0 array where each stripe is a level 5 array. Level 53 is a level
5 combined with a level 0. Strictly this level should be called level 50, but
this description has already been used.
JBOD
JBOB (just a bunch of disks) is a name given to a group of hard disks
that have no particular array pattern. It is a name often given to disk
systems in an attempt to give them the same importance as RAID levels.
177 Sony Broadcast & Professional Europe

186. Part 17 – The special effects machine
Stripe size considerations
Most RAID solutions require the files be split into small pieces called
stripes. Each stripe is written to a different disk in the array. Stripe size is
important in defining the performance of the array.
Striping a file means that parts of the file can be written to and read from
multiple disks at the same time. This greatly increases the performance
of the array as a whole. However the task of splitting files into stripes
takes time and effort. Therefore it is important that the stripe size is
correct.
Ideally every file would be divided into exactly the number of disks in the
stripe set. This gives the highest performance of all, with the lowest
splitting workload. However stripe size if fixed.
Working with small stripes
A RAID system that uses small stripes works well for files systems with
many small files and few large files. It is easy to split the files into stripes
as they are already small, but most files can be divided so taking
advantage of the increased performance of striping. Any large files take
longer to split and give lots of stripes that have to be handled. But there
are few of these files in this kind of file system.
Working with large stripes
A system that uses large stripes works well for file systems with
predominantly large files, and with few small files. The large files splits
easily into a few stripes that can be stored on the disk array quickly. Any
small files may well be smaller that the stripe size and cannot therefore
be split. There is no performance advantage of striping for these files,
but there are few of then in this file system, so overall performance is still
high.
Software v hardware RAID
RAID systems are designed to perform all the RAID processing either as
a software solution or in dedicated hardware. In fact both these solutions
perform the same function, just in a different place and in a slightly
different way.
RAID controller processing
The RAID controller performs three basic operations during write
operations. Firstly is splits the files into stripes. Secondly, it may
calculate an error correction code. This may be a simple parity code, a
dual parity code or a Hamming code. Thirdly it arbitrates and controls
data write operations to each disk’s interface.
Software RAID
Software RAID performs all the file splitting and error correction
calculation in the computer’s processor using a small piece of software
resident in the computer’s memory. This processing robs processor
Sony Training Services 178

187. Broadcast Fundamentals
resource and is somewhat inefficient, but is simple to achieve, and
relatively simple to modify.
The calculation of error correction codes is generally very resource
hungry. Software RAID solutions are more popular for RAID level 0 and
1 where no error correction codes are used.
Hardware RAID
Hardware RAID performs all the controller processing in dedicated
hardware. This removed all the workload from the central processor
allowing it to perform other tasks with greater efficiency. The dedicated
hardware is often actually a fast processor coupled to some dedicated
integrated firmware.
Hardware RAID controllers are faster than software RAID controllers, but
are more expensive and dedicated to specific RAID levels and stripe
sizes.
Realising RAID systems
At the time RAID was proposed by Randy Katz, and others, in 1988 the
idea was to allow large storage elements to be built from lots of small
cheap drives, with redundancy built in to allow for errors and disk failure.
Direct disk connection
The fastest method of writing to and reading from hard disks is to
communicate directly with the disk platters. Early video disk recorders
used direct access as the only way of achieving the bandwidth to and
from the disks for full bandwidth broadcast video.
However this imposes extra loading on the computer that is using the
hard disks. All the sector, and cylinder allocation had to be done by the
computer’s central processor.
Direct access also imposes a risk of drive retirement. Disk drive
manufactures often improve their products and alter the layout of platters
and their density. This may not change the overall disk capacity, and
therefore will make little difference to normal computer systems.
However it does effect video disk recorders that rely on direct access to
the hard disk platters.
Bus RAID connections
All hard disks are now built with some kind of integrated controller. This
handles all the sector and cylinder addressing. The computer is
presented with logical addressing that has nothing to do with the actual
sector and cylinder addressing the drive itself will use.
This removes the loading from the computer and also removes the risk
of disk retirement. However hard disk integrated controllers add a layer
between the computer and the disk platters themselves and slow data
transfer to and from the disks. Hard disk technology had to evolve before
hard disk with integrated controllers could be used in video disk
recorders.
179 Sony Broadcast & Professional Europe

188. Part 17 – The special effects machine
The popularity of IDE/ATA, SCSI, and later, serial SCSI, fibre and
IEEE1394, coupled with the advances in hard disk technology made it
possible and easy to build RAID systems for broadcast quality video.
SCSI
SCSI (small computer systems interface) was originally designed as a
method of connecting peripherals to a computer with a very fast data
link. Although SCSI has been popular as a method of connecting
scanners, and a number of other peripherals, the most popular
peripheral that uses SCSI is the hard disk.
The original version of SCSI allows for 8 nodes on the whole bus. One of
these must be the controller. The other 7 nodes can be hard disk drives.
Later versions of SCSI allow for 16 devices (1 controller and 15 drives).
The length of the SCSI bus is also important. Different versions of SCSI
have different maximum bus lengths. Originally the SCSI bus could not
be any longer than about 6 metres. Now it is possible to achieve a SCSI
bus of about 25 metres, although this imposes other restrictions.
There are about 10 different flavours of SCSI in current use, and a few
new ones starting to appear. The table on the next page shown these
different types.
Name is the common name given to this type of SCSI. In some cases
two names are in-fact identical SCSI types.
Standard is the official SCSI standard to which this type refers.
Bus is the size of the SCSI bus, in bits. This is not the number of pins in
the connector, or conductors in the cable, although there is a general
relationship.
Rate is the data rate for this SCSI type. This is not the bus speed. Some
SCSI types use special tricks to multiply the bus speed to increase the
actual data throughput on the bus.
Connectors is a guide to the kind of connectors used in the various
SCSI types. There is no hard and fast rule to the connector type, and
each SCSI equipment manufacturer has their own preference. However
certain connectors are specifically for external, and others for internal,
use.
Cables is a guide to the kind of cables used in the various SCSI types.
Signal is the kind of electrical signal this type of SCSI uses. All SCSI
types use a pair of wires on the bus for each bit. The original electrical
signal was single ended (S), in which case one of the wires in each pair
was the data itself while the other was connected to ground. Later types
used differential connections to increase the possible bus length. Each
pair had a positive and negative version of the signal. The original used
relatively high voltages for the signals and was so called high voltage
differential (H). The latest differential type is low voltage differential (L),
which allows for longer bus lengths and high data rates.
Max devices is the maximum number of devices allowed on the SCSI
bus, including the controller. In some cases the maximum number of
devices allowed will depend on the signal type, and will also have an
Sony Training Services 180

189. Broadcast Fundamentals
effect on the maximum length of the bus. For example Ultra 2 SCSI will
allow high voltage or low voltage differential connection, with low voltage
differential connections allowing either 2 of 8 devices depending on the
bus length.
Max cable is the maximum length of the whole SCSI bus, including the
cable itself, any internal connections and any circuit board tracks.
Different SCSI types allow for different bus lengths. For instance Ultra
SCSI allows for single ended or high voltage differential bus
connections. If 3 metres of single ended connection are used, only 4
devices can be connected. If the bus length is halved to 1.5 metres, the
number of devices doubles to 8.
181 Sony Broadcast & Professional Europe

190. Bus R a te
N am e S ta n d a r d (b its ) (M B /s ) C o n n e c to rs C a b le s S ig n a l M a x d e v ic e s M a x c a b le
S C S I-1 S C S I-1 8 5 1 ,2 ,3 ,4 ,5 1 ,2 S ,H 8 S :6 H :2 5
N a rro w S C S I S C S I-1 8 5 1 ,2 ,3 ,4 ,5 1 ,2 S ,H 8 S :6 H :2 5
Fast S C S I S C S I-2 8 10 2 ,3 ,7 1 ,2 ,3 S ,H 8 S :3 H :2 5
Sony Training Services
F a s t N a rro w S C S I S C S I-2 8 10 2 ,3 ,7 1 ,2 ,3 S ,H 8 S :3 H :2 5
W id e S C S I S C S I-2 16 10 6 ,1 1 4 S ,H 16 S :6 H :2 5
F a s t & W id e S C S I S C S I-2 16 20 6 ,1 1 4 S ,H 16 S :3 H :2 5
U ltr a S C S I S C S I-3 8 20 2 ,3 ,7 1 ,2 ,3 S ,H S :4 o r8 H :8 S 4 :3 S 8 :1 .5 H :2 5
N a r r o w U ltr a S C S I S C S I-3 8 20 2 ,3 ,7 1 ,2 ,3 S ,H S :4 o r8 H :8 S 4 :3 S 8 :1 .5 H :2 5
W id e U ltr a S C S I S C S I-3 16 40 6 ,1 1 4 S ,H S :4 o r8 H :8 S 4 :3 S 8 :1 .5 H :2 5
U ltr a 2 S C S I S C S I-3 8 40 2 ,3 ,7 1 ,2 ,3 H ,L H :8 L :2 o r8 H :2 5 L 2 :2 5 L 8 :1 2
N a r r o w U ltr a 2 S C S I S C S I-3 8 40 2 ,3 ,7 1 ,2 ,3 H ,L H :8 L :2 o r8 H :2 5 L 2 :2 5 L 8 :1 2
W id e U ltr a 2 S C S I S C S I-3 16 80 6 ,1 1 4 H ,L H :1 6 L :2 o r1 6 H :2 5 L 2 :2 5 L 1 6 :1 2
U ltr a 3 S C S I S C S I-3 16 160 6 ,1 1 4 L 2 o r1 6 L 2 :2 5 L 1 6 :1 2
U ltr a 1 6 0 S C S I S C S I-3 16 160 6 ,1 1 4 L 2 o r1 6 L 2 :2 5 L 1 6 :1 2
U ltr a 1 6 0 + S C S I S C S I-3 16 160 6 ,1 1 4 L 2 o r1 6 L 2 :2 5 L 1 6 :1 2
U ltr a 3 2 0 S C S I S C S I-3 16 320 6 ,1 1 4 L 2 o r1 6 L 2 :2 5 L 1 6 :1 2
U ltr a 6 4 0 S C S I S C S I-3 16 640 ? ? L ? ?
182
Part 17 – The special effects machine

191. Broadcast Fundamentals
SCSI connectors
1 : 25 pin, D25 connector. 1
2 : 50 pin Centronics connector. 12
3 : 50 pin IDC connector. 1 2 ultra
4 : 50 pin D50 connector. 1
5 : 37 pin D37 connector. 1
6 : 68 pin HD68 connector. Ultra2 lvd & ultra wide scsi3
7 : 50 pin HD50 connector. 23
8 : 30 pin HDI30 connector. Apple
9 : 50 pin HPCN50 connector.
10 : 60 pin HDCN60 connector.
11 : 68 pin VHDCI connector. Ultra scsi 2 & 3
SCSI cables
1 : 50 conductor Centronics C50.
2 : 50 conductor ribbon cable.
3 : 50 conductor high density D50M cable.
4 : 68 conductor high density D68 cable.
IDE/ATA
Early PC designs placed the hard disk on a card, integrating it with the
controller and providing a simple connection through one of the ISA
connectors into the motherboard. However this was awkward because it
made the card large, heavy and cumbersome.
Western Digital produced a card that provided an interface between the
16 bit ISA bus connector on the motherboard and the drive. Controller
electronics were placed on the drive, providing a simple interface without
having to communicate directly with the disk platters, just as SCSI does.
This was called integrated drive electronics (IDE)
Because the PC design this was first used in was called the PC/AT the
adaptor was called the AT adaptor, or ATA.
Several other manufacturers saw the simplicity of the IDE/ATA design
for PC’s. These computers did not need any of the complexity or
performance of SCSI, and IDE/ATA became the de-facto standard for
fitting hard disks into PC’s.
Every PC required a hard disk, some more than one. It became obvious
that the ATA controller should be fitted to the PC motherboard, rather
than wasting one of the ISA slots.
In the early 1990’s ATA packet interface (ATAPI) was introduced. This
enhancement allowed CDROM’s and tape drives to be integrated into
183 Sony Broadcast & Professional Europe

192. Part 17 – The special effects machine
the same bus connection as the hard disks rather than connecting them
to some other proprietary interface.
Later versions of the IDE/ATA interface allowed direct memory access
(DMA) modes, and later, faster DMA modes, called Ultra DMA. (UDMA).
UDMA mode 2 allowed for a data transfer rate of 33MB/s and was often
called Ultra DMA-33 or Ultra ATA-33, or simple UDMA-33 or ATA-33.
Later improvements to the bus appears as UDMA-66 (ATA-66) and
UDMA-100 (ATA-100).
The performance of the whole drive/controller configuration will drop to
the item with the slowest speed. Therefore it is important to ensure that
both the ATA controller and the drives have to be designed to operate at
the correct bus speed.
Most IDE drives can be connected via a standard 40 way ribbon cable.
However any ATA controller and drive faster than UDMA-33 must use
special 80 way cable. This cable is exactly the same overall size as the
40 way cable, and has the same number of signal connections as 40
way connections. However every other wire in this 80 way ribbon cable
is connected to ground and separates the signal wires, improving
performance.
Modern PC’s integrate the ATA into the motherboard’s chipset. Intel’s
PCI chipset now integrates the entire ATA into the PCI chipset. All
motherboards now include two 40 pin connectors on the motherboard.
Each connector provides one IDE/ATA bus. Each bus allows for one
master and one slave drive. Thus four IDE drives can be fitted. It should
be remembered that ATAPI allows all drives, including CDROM and
DVD drives to be connected to these IDE connectors. Most PC’s have
this connection, with very few drives connections available to build a
RAID from.
IDE/ATA RAID
Manufacturers have now produced plug-in cards that have multiple ATA
connections, and an interface controller. These cards allow small RAID
systems to be built into the PC.
Some of these cards rely on software to perform the RAID control.
These are little more than the standard ATA controllers found integrated
to motherboards. They generally only have two 40 way connectors
allowing four drives to be fitted. These RAID solutions are somewhat
restricted, and slow.
Other cards offer hardware RAID. Free from the constraints of the
normal two connector scheme these cards often include four or more 40
way connectors, allowing for more drives to be connected. They include
coprocessors and memory to perform proper interface and control.
Some motherboards include more ATA connectors other than the
normal two. These are specifically designed to allow small RAID
systems to be added to the PC without the need for any plug-in card.
Just as with the plug-in solutions, motherboard integrates solutions can
offer either software or hardware based RAID.
Sony Training Services 184

193. Broadcast Fundamentals
Hardware based plug-in IDE RAID cards and the motherboard
integrated IDE RAID controllers tend to use hardware based RAID level
0, 1 and 5. These are by far the most popular RAID levels for PC RAID
designs.
Serial SCSI and Fibre
The argument for serial SCSI
To an engineer it may appear that a parallel interface should be faster
than a serial one. After all if you can sent data down a serial bus 8, 10,
16, 20, 32, or even 64 bits as one big chunk, each clock cycle, this must
surely be faster than sending it one bit at a time. Surely if, for instance
you have a 16 bit parallel bus connection the data rate would have to be
16 times faster to achieve the same data rate with a serial bus.
However various transmission effects conspire to ensure that serial bus
SCSI connections in fact offer greater performance than most parallel
connections.
As the data rate increases there is an increase in cross-talk between
one conductor in a parallel bus to another. Each data bit becomes more
corrupted and the data rate is stepped up.
As the cable length is increased there is an increase in bit slippage. This
is where the data in one conductor gets to its destination before another
conductor simply because of the data pattern in each wire and its
corresponding delay. All the bits from one word of data arrive at the far
end of the bus at slightly different times and become more difficult to
read.
With many pins in a serial connector there is a greater chance that any
one pin will not make proper contact. This may invalidate transmitted
data.
Some parallel SCSI connections are still the fastest connection method.
Ultra 320 and the up and coming Ultra 640 versions of SCSI are still
parallel connections. However most SCSI installations are based around
10, 20 or 40MB/s buses where serial connection are better.
Which flavour?
With the introduction of SCSI-3 the whole structure of the format was
altered giving it more of a layered and modular structure, with each
module communicating with others in the structure. Any one
implementation need not use all the modules, just enough to ensure
messages and data are properly transferred.
Although appearing complex the new approach allowed elements of the
interface to be altered while still maintaining the overall format.
An important concept for SCSI-3 is that the physical elements are now
separated from the command definitions. A range of different physical
modules exist, for traditional parallel connections as well as some serial
connections and network connections.
185 Sony Broadcast & Professional Europe

194. Part 17 – The special effects machine
The important serial connections are Serial Storage Architecture (SSA),
Fibre Channel (FC) and IEEE1394.
IEEE1394 is more of a multimedia interface. Although popular as a
media interconnection, it is not popular in RAID design.
There is plenty of discussion on the relative merits of SSA and FC with
both supporters backing their own preference and decrying the other. I
appears that FC have more universal backing irrespective of any
technical merits (although it seems FC is also technically better).
Fibre Channel
The main advantages for using FC in a RAID environment are :-
1 : FC is a network protocol, SCSI and ATA/IDE are not. This allows
drives to be addressed just as anything else on a computer network.
2 : Different connection topologies possible. Point to point (the nearest to
parallel SCSI and ATA/IDE), arbitrated loop and fabric.
3 : Huge connection distances compared to parallel SCSI or ATA/IDE.
4 : Each computer can access a huge number of disks, not the 2 per bus
on ATA/IDE or 16 on SCSI.
5 : Each disk can be accessed by a huge number of computers. This is
not possible with either ATA/IDE or SCSI and allows easy file and
directory sharing.
6 : Easy connection.
7 : Hot swappable connection.
8 : Fast. As already discussed there are some versions of SCSI that are
faster but these are exotic and not universally supported at present. FC
is as fast or faster than most common SCSI types and every form of
ATA/IDE.
Sony Training Services 186

195. Broadcast Fundamentals
Part 16 Television receivers & monitors
The basic principle
A television receiver’s task is to turn the electrical video signal that is
connected to it back into a moving image. In most cases television
receivers include a tuner to accept signals from an aerial, cable or
satellite feed. These signals also include audio information, which the
television receiver will turn back into a recognisable audio signal.
In a domestic arena the television receiver is often referred to simply as
a “television”.
Monitors are designed for use in professional and broadcast situations.
They have the same technology as a television, although the input signal
possibilities are normally restricted to those used in professional and
broadcast situations. They normally have no tuner and cannot be
connected to an aerial, cable or satellite feed. Monitors sometimes have
provision for audio but its quality is not normally very good.
Monitor video quality has a far greater range of quality than for television
receivers. At the lowest end of the quality scale are mini-monitors
intended for CCTV and surveillance. Compact design, robustness and
price tend to be the defining factors in these monitors. Picture quality is
not so important and is often poorer than for domestic televisions.
Broadcast monitors have very good quality because they are used as a
reference in the broadcast station. They are expensive and often need
periodic alignment checks to retain their quality. Studio monitors are
graded according to their quality. A grade 1 monitor is the best quality.
The tube is selected for its definition and colourimetry, and the circuitry
is designed with no compromise to picture quality.
Input signals
Analogue inputs
Terrestrial
By far the most popular input signal to a television is the UHF signal.
Called terrestrial because the signal is sent over land as a radio signal. It
uses a transmitter mast at the broadcast station and an aerial at the
receiver. The radio frequency carrier holds a composite video signal and
its associated audio, in a bandwidth of about 6MHz.
Televisions include radio frequency tuners that can tune into one of
these terrestrial signals and demodulate the video and audio signals,
turning them back into baseband analogue signals.
Composite
Composite video is a baseband signal, and does not include audio.
Audio must be input separately. This presents a difficulty in a domestic
environment where simple installation is very important. The most
187 Sony Broadcast & Professional Europe

196. Part 18 – Television receivers & monitors
popular method of connecting composite video to a domestic television
in Europe is through a Scart connector. The Scart connector is a multi-
pin connector providing component, composite and audio connections in
one connector.
Many European domestic television peripheral equipment, like video
tape players, DVD players, and video games machines have Scart
connectors fitted and provide a simple way of connecting to domestic
television receivers.
Composite is common in broadcast stations and post-production.
Generally regarded as a low quality connection for monitors, compared
to analogue component and digital connections, composite is easy to
connect and still provides a reasonable monitoring image.
Component
Component video connections are normally more difficult to connect
because they involve three connectors. Add to this the fact that, like
composite, component is a baseband video signal with no audio, and
component is not an easy option for domestic use where simplicity is
paramount. However analogue component provides a very high quality
connection for domestic purposes. An increasing amount of peripheral
equipment, like video tape players, DVD players, and video games
machines are being designed to output component signals using the
Scart connector, providing the home viewer with a relatively high quality
image.
Component is the preferred analogue video connection within the
broadcast station, and most studio monitors provide for component
analogue input, using three separate, usually BNC, connectors.
Digital inputs
Satellite/cable
Although there are an increasing amount of people subscribing to
satellite and cable channels, it is still rare to find a television receiver
with a built-in satellite decoder. Most receivers use an external decoder
and the television takes a decoded signal from the decoder. This is often
a UHF signal, similar to a terrestrial signal, but could be either a
baseband composite or component analogue signal.
Digital terrestrial
Like satellite and cable, digital terrestrial is still a rare option as an input
for television receivers. Most people subscribing to domestic digital
terrestrial broadcast services use an external decoder box with the
television taking a UHF, composite or component analogue signal.
Sony Training Services 188

197. Broadcast Fundamentals
Part 17 Timecode
A short history
Splicing tape
Ever since video was first recorded there has been a need to edit video
material. At first this process consisted of little more than removing
errors and any material not required in the final recording. This was done
by simply cutting the video tape.
As technology progressed efforts were made to perform the same
editing tasks that were already in common use in film, i.e. making up a
complete program from bits and pieces of video joined together. As with
film this was done by simply cutting the required sections of video and
splicing them together. Edits were often badly made, causing picture
breakup and rolls at the edit points, and once the edit was made there
was no turning back.
Electronic editing
A little later electronic editing was introduced. Rather than cutting the
video tape up into pieces, a copy would be made from the original tape
onto a new tape. By electronically organising how the various bits of
video material were copied from the original tape to the new one it
became possible to edit a complete program together without affecting
the original video tape.
It soon became necessary to index the tape in some way so that
particular edit points could easily be found. In the early 60’s Ampex
introduced a system called Editek. This system allowed the editor to
insert an audio tone into the audio channel of the video tape at the
chosen edit point. The recorder and player VTR would then use the tone
to switch at the edit point and perform the edit electronically.
Although providing editors with a technical advantage over anything that
had gone before, Editek was still slow and not as easy to use as could
be. Further more, Editek was not frame accurate.
Frame accuracy
Film uses sprocket holes to mechanically move it through the projector.
By linking the mechanics of the projector to a counter it was therefore
easy to get an accurate frame count as the film progressed through the
projector.
Early efforts were made to do the same thing with video tape, by
counting capstan rotations. This was however very inaccurate due to
slippage.
A little later the control track was used. As video uses a control track to
lock the player’s mechanics to the helical video tracks recorded on tape
a simple counter could be attached to the control track servo system to
count frames in much the same way as one would counting sprocket
holes in film.
189 Sony Broadcast & Professional Europe

198. Part 19 – Timecode
However, if the film was damaged sprocket holes could be missed and
the overall count would slip. In much the same way, if video tape was
stopped and started, and wound backwards and forwards repeatedly,
control track pulses could be missed and the count would slip. Also if the
film or video tape was loaded somewhere in the middle, one would have
no idea how far from the beginning one was.
What was needed, not only for video tape editing, but also for film
editing, was a method of individually “marking” each frame with a unique
number.
Timecode
In the late 60’s timecode was introduced. Simply called ‘timecode’ this
coding method would later be called ‘longitudinal timecode’ when the
alternative ‘vertical interval timecode’ was introduced some ten years
later.
Timecode provided editors with the system they had been waiting for. A
coding system that was designed to be read both in the forward and
reverse directions, at a wide range of tape speeds with a numbering
system that was related to real time, with a unique code for each and
every video frame.
Computer based editing systems soon became popular, allowing edits to
be programmed as a number of timecode related points. The idea of an
edit list came about, and it became common to carry an edit list, either in
paper form or on disk, with video tapes, when moving from one edit suite
to another.
Future uses of timecode include very sophisticated computer controlled
equipment using timecode and related video clips or snapshots for
versatile off-line editing suites.
Timecode’s basic structure
Timecode is represented as 8 digits split into 4 pairs of 2 digits each,
separated by colons, as shown in figure . Each digit pair conveys hours,
minutes, seconds, and frames, reading from the left.
Figure 87 Timecode’s basic structure
Sony Training Services 190

199. Broadcast Fundamentals
Timecode gives a 24 hour count. This amount is considered longer than
any single piece of video would last, and can also be set to be the time
of day, thus allowing the time a video recording was made to be
recorded on tape as well.
Both LTC and VITC are conveyed and recorded on tape as a serial data
stream. This data stream is a collection of binary bits, 80 bits for LTC
and 90 bits for VITC. Groups of these bits define various elements of the
timecode.
Timecode address bits (BCD)
Figure 88 Binary coded decimal
There are 26 address bits separated into 8 BCD (binary coded decimal)
groups, of either 2, 3 or 4 bits each. Each BCD group defines either the
tens or units digit of the hours, minutes, seconds or frames count.
User bits (binary groups)
There are 32 user bits separated into 8 groups of 4 bits each. These
groups can either be used in any way the user sees fit, or can be
specified to comply with 7 and 8 bit standard ISO character sets, or can
define another timecode value which can be the same, or different, from
the one defined by the timecode address bits.
Thus by using the user bits as timecode, LTC or VITC can hold 2
unrelated timecode counts.
Sync bits
In LTC there are 16 sync bits placed as one group (word) at the end of
each code. They define the end of each code, so that an LTC timecode
reader can find the beginning of each code. They also define the tape
direction because the sync word is different in the forward direction as it
is in the reverse direction.
In VITC there are 18 sync bits separated into 8 pairs occurring
throughout the code after each user group.
191 Sony Broadcast & Professional Europe

200. Part 19 – Timecode
Flags
There are 6 special single bit flags. They are placed in the bits ‘saved’ by
2 and 3 bit BCD timecode address groups, as explained as part of the
timecode address bits on page.
The flags are defined as follows :-
Drop frame flag
Used to identify if the timecode increments according to the NTSC drop
frame counting method. See page 131.
Colour frame flag
Used to identify if the timecode is related to composite video material, or
component video material that has been decoded from composite video
material. The exact field in the colour frame sequence is found by
mathematically calculating on the basis that timecode 00:00:00:01 is the
first field of the colour frame sequence.
Phase correction flag (LTC only)
Used to ‘switch’ the LTC phase if there are an uneven number of 1’s in
the complete code. This makes sure that each code starts low at the
beginning of each video frame.
(In VITC this flag is used as a field mark flag.)
Binary group flags
Used to define how the user binary group bits are to be used as shown
in the table below.
Binary group flags Function
2 1 0
0 0 0 Character set not specified
0 0 1 Eight bit character set
0 1 0 Unassigned
0 1 1 Unassigned
1 0 0 Page/Line
1 0 1 Unassigned
1 1 0 Unassigned
1 1 1 Unassigned
The first state, with all bits at ‘0’ specify that all the user bit groups are
undefined and can be used in any way the user sees fit.
The second state specifies that the user bit binary groups are taken in
pairs, giving either 7 or 8 bit groups that are used to specify an ISO
character.
Sony Training Services 192

201. Broadcast Fundamentals
Field mark flag (VITC only)
VITC is field sensitive, unlike LTC which is only frame sensitive. This flag
is used to define the field. For field 1 this flag is ‘0’, for field 2 it is ‘1’. For
NTSC and PAL based video material this flag is ‘0’ for odd field and ‘1’
for even fields
(In LTC this flag is used as a phase correction flag.)
Unassigned timecode flags
There are two unassigned flags in timecode. These have been left for
future expansion if a new technology is devised that may require more of
timecode than in presently provided.
193 Sony Broadcast & Professional Europe

202. Part 19 – Timecode
Longitudinal timecode
Figure 89 The longitudinal timecode head
LTC (longitudinal timecode) was the first timecode to be proposed and
extensively used, during the latter part of the 1960’s. It uses a linear, or
longitudinal, track running along the edge of the video tape. The actual
position on tape varies from one standard to another. C format, for
instance ‘borrows’ audio track 3 along the bottom edge of the tape for
timecode. U-Matic machines have an extra track placed at the bottom
end of the helical track for timecode.
½” tape formats like Betacam, Betacam SX Digital Betacam, IMX and
HDCAM place LTC along the bottom edge of the tape below the control
track.
LTC has the advantage that it can be read at high tape speed when
VITC cannot be read. However it has the disadvantage that it cannot be
read at zero tape speed (stop mode) because the tape is no longer
moving passed the longitudinal timecode head.
LTC signal structure
LTC consists of 80 bits of data recorded serially on tape beginning at the
same time as line 5 of either the 525 or 625 line sequence is being
written to the helical tracks on tape.
Sony Training Services 194

203. Broadcast Fundamentals
(There is an obvious physical displacement between these two points on
tape, but this displacement is the same within each tape format, and
therefore is not a problem.)
195 Sony Broadcast & Professional Europe

204. Figure 90
0 4 8 12 16 20 24 28 32 36 40 44 48 52 56 60 64 68 72 76 79
Sony Training Services
F ra m e s U s e r b in a r y U s e r b in a ry S econds U s e r b in a ry S econds U s e r b in a ry M in u te s U s e r b in a ry M i n u te s U s e r b in a ry H o u rs U s e r b in a ry H o u rs U s e r b in a ry S y n c r o n is a ti o n w o r d
u n it s c o u n t g ro u p 1 g ro u p 2 u n its c o u n t g ro u p 3 te n s g ro u p 4 u n its c o u n t g ro u p 5 te n s g ro u p 6 u n its c o u n t g ro u p 7 te n s g ro u p 8
count count co u n t
The longitudinal timecode signal
Part 19 – Timecode
196

205. Broadcast Fundamentals
B it U s e (5 2 5 ) U s e (6 2 5 ) B it U s e (5 2 5 ) U s e (6 2 5 )
0 F r a m e u n its b it 0 40 M in u t e s t e n s b it 0
1 F r a m e u n its b it 1 41 M in u t e s t e n s b it 1
2 F r a m e u n its b it 2 42 M in u t e s t e n s b it 2
3 F r a m e u n its b it 3 43 B in a r y g r o u p fla g 0 B in a r y g r o u p f la g 1
4 U s e r g ro u p 1 44 U s e r g ro u p 6
5 U s e r g ro u p 1 45 U s e r g ro u p 6
6 U s e r g ro u p 1 46 U s e r g ro u p 6
7 U s e r g ro u p 1 47 u s e r g ro u p 6
8 F r a m e te n s b it 0 48 H o u r s u n it s b it 0
9 F r a m e te n s b it 1 49 H o u r s u n it s b it 1
10 D r o p fr a m e fla g U n a s s ig n e d ( s e t t o 0 ) 50 H o u r s u n it s b it 2
11 C o lo u r f r a m e f la g C o lo u r f r a m e f la g 51 H o u r s u n it s b it 3
12 U s e r g ro u p 2 52 U s e r g ro u p 7
13 U s e r g ro u p 2 53 U s e r g ro u p 7
14 U s e r g ro u p 2 54 U s e r g ro u p 7
15 U s e r g ro u p 2 55 U s e r g ro u p 7
16 S e c o n d s u n its b it 0 56 H o u r s t e n s b it 0
17 S e c o n d s u n its b it 1 57 H o u r s t e n s b it 1
18 S e c o n d s u n its b it 2 58 B in a r y g r o u p f la g 1 B in a r y g r o u p f la g 2
19 S e c o n d s u n its b it 3 59 B i n a r y g r o u p f l a g 2 P h a s e c o r r e c t io n f la g
20 U s e r g ro u p 3 60 U s e r g ro u p 8
21 U s e r g ro u p 3 61 U s e r g ro u p 8
22 U s e r g ro u p 3 62 U s e r g ro u p 8
23 U s e r g ro u p 3 63 U s e r g ro u p 8
24 S e c o n d s te n s b it 0 64 S y n c w o rd (s e t to 0 )
25 S e c o n d s te n s b it 1 65 S y n c w o rd (s e t to 0 )
26 S e c o n d s te n s b it 3 66 S y n c w o rd (s e t to 1 )
27 P h a s e c o r r e c t io n f la g B i n a r y g r o u p f l a g 0 67 S y n c w o rd (s e t to 1 )
28 U s e r g ro u p 4 68 S y n c w o rd (s e t to 1 )
29 U s e r g ro u p 4 69 S y n c w o rd (s e t to 1 )
30 U s e r g ro u p 4 70 S y n c w o rd (s e t to 1 )
31 U s e r g ro u p 4 71 S y n c w o rd (s e t to 1 )
32 M in u t e s u n it s b it 0 72 S y n c w o rd (s e t to 1 )
33 M in u t e s u n it s b it 1 73 S y n c w o rd (s e t to 1 )
34 M in u t e s u n it s b it 2 74 S y n c w o rd (s e t to 1 )
35 M in u t e s u n it s b it 3 75 S y n c w o rd (s e t to 1 )
36 U s e r g ro u p 5 76 S y n c w o rd (s e t to 1 )
37 U s e r g ro u p 5 77 S y n c w o rd (s e t to 1 )
38 U s e r g ro u p 5 78 S y n c w o r d ( s e t to 0 ) S y n c w o r d ( s e t to 1 )
39 U s e r g ro u p 5 79 S y n c w o rd (s e t to 1 )
Figure 91 LTC bits
197 Sony Broadcast & Professional Europe

206. Part 19 – Timecode
LTC is recorded as simple polarised regions on tape according to the bi-
phase mark channel coding method. See page for an explanation on the
bi-phase mark method of channel coding.
The bi-phase mark signal must obey certain criteria which are outlined in
Fig. on page.
Figure 92 Longitudinal timecode signal detail
The 80 bits are evenly spaced over the whole frame. They are separated
into groups of bits responsible for the timecode itself, user binary groups,
flags and syncs. The usage of these groups is described on page .
The LTC signal structure is shown in Fig. and on page 122.
LTC sync bits
LTC contains 16 sync bits at the end of each code. These bits have a
particular sequence, ‘0011111111111101’. The 12 bits in the middle of
the sync bits are all ‘1’s. Because the timecode groups are all BCD (see
page this particular pattern of 12 bits cannot occur anywhere else in the
LTC code. Thus an LTC reader can determine where the end of the LTC
code is and therefore know to begin looking for the beginning of the next
code.
The bi-phase mark signal structure is direction independent (see page ).
In the reverse direction the code reads ‘1011111111111100’. Because
the reader finds ‘10’ before the 12 ‘1’s in the middle of the sync bits, and
‘00’ at the end it knows the tape is running backwards. It therefore
knows the code will occur after the sync, not before it, and that it will be
backwards, and that the appropriate adjustments therefore have to be
made to read the code correctly.
Sony Training Services 198

207. Broadcast Fundamentals
Bi-phase mark coding
LTC uses bi-phase mark as a channel coding method, otherwise known
as the Manchester code 1.
Bi-phase mark places a transition at every bit boundary, and a transition
in the middle of each bit period for each ‘1’ bit. This makes it polarity
independent i.e. anything reading bi-phase mark is only concerned with
the transitions, not whether the transitions are going high or low.
Figure 93 Bi-phase mark coding
It is also direction independent, and the original data can always be
decoded from a bi-phase mark signal even though it may be read
backwards. It is also self clocking, i.e. no matter what speed the tape is
going, all an LTC reader has to do is to look for regular transitions
corresponding to the bit boundaries, and once locked to them, search for
any bit periods with a transition in the middle. Any with no transitions are
0’s and those with, are 1’s.
Adjusting the LTC head
The LTC head is a static head as shown in Fig 66.
199 Sony Broadcast & Professional Europe

208. Part 19 – Timecode
Head to tape contact
The head itself hangs off a bracket. The head to tape contact can be
adjusted by loosening the fixing screws between the bracket and the
head and rotating the head about the vertical axis.
The head gap must be in direct contact with the tape if it is to record and
playback timecode properly.
Head height
The bracket is held on a plate by a large spring underneath the whole
assembly. The spring is trying to pull the whole head downwards. Thus
by adjusting a small screw between the bracket and the plate you can
adjust the head height.
Figure 94 LTC head adjustments
The head assembly must be at the correct height if each head is to
cover its track properly.
Sony Training Services 200

209. Broadcast Fundamentals
Head zenith
The plate is fixed to the base plate by a screw and spring arrangement.
There is a small pivot at the back of the head which keeps the plate and
base plate apart. Thus by adjusting a small screw at the front whish is
also separating the plate and base plate, you can adjust the head’s
zenith, i.e. the amount of lean forwards or backwards.
If the head zenith is incorrect either the timecode or audio portions of the
head will not be in good contact with the tape and both recording and
playback will be bad. Furthermore, as the tape moves across the head it
will forced either upwards or downward by the head. This may make
video tracking difficult and may force the tape against the tape guides
and damage the edge of the tape.
Head azimuth
The plate is also held from the base plate by another screw at the side of
the assembly. This screw can be used to adjust the head azimuth, i.e.
the sideways lean.
Incorrect head azimuth will result in incorrect audio phase and incorrect
relative position between the audio heads and the timecode head.
Head position
The base plate is fixed to the mechadeck with a number of screws.
The fixing holes in the base plate are actually slots. Thus by loosening
the screws the head position on the tape path can be adjusted.
If the head position is incorrect the relative timing between the timecode
and audio signals compared to the control and video signals will be
wrong. Lip sync will be incorrect and timecode may be incorrectly read,
resulting in bad edits.
201 Sony Broadcast & Professional Europe

210. Part 19 – Timecode
Vertical Interval Timecode
The basis for VITC
A form of VITC was proposed at the same time as LTC. However
machines at the time generally found it difficult to maintain a good video
signal at anything other than normal play speed, thus VITC offered no
advantage over LTC.
As VTR technology progressed, video playback heads were designed
that could move to follow the helical tracks on tape at other than normal
play speed.
As designs improved it soon became possible for the video heads to
follow the helical tracks at very slow speeds and even in still mode, while
maintaining a steady picture.
Figure 95 LTC and VITC speed comparison
At slow and still speeds LTC becomes unreadable, and editing becomes
difficult. Eventually a workable VITC was proposed about ten years after
LTC, to get over this problem. It uses two lines during the field blanking
Sony Training Services 202

211. Broadcast Fundamentals
period, (otherwise known as vertical blanking) to store serial data with
much the same format and content as LTC.
Because VITC is written into the video signal itself as part of the helical
tracks it has the advantage over LTC in that it can be read at zero tape
speed (still mode), because even at this speed the flying heads on the
scanner are still moving over the tape.
Another advantage of VITC over LTC is that it is field accurate, because
a complete code is placed during the vertical interval of each field,
whereas LTC requires a whole frame to convey one code.
Because VITC is included in the video signal itself, it also has the
advantage that cabling can be made simpler, with no extra cable
required specifically for timecode.
However at high tape speeds, no matter how good the heads are at
following the helical tracks, they eventually lose their position on the
helical tracks, and consequently lose the VITC signal.
Another disadvantage of VITC is that there is no agreed set standard for
the vertical blanking interval lines that should be used for the VITC
signal. The proposal was simply put to the industry too late, and other
uses had already been found for the vertical interval, teletext and vertical
interval test signals being 2 examples.
Therefore VITC can occur on just one line on field 1 anywhere between
lines 9 and 22 and on field 2 anywhere between lines 322 and 335.
VITC signal structure
Because VITC is stored in the vertical interval it therefore conforms to
the same basic rules as the video signal itself does. In fact VITC
conforms to the same criteria as a monochrome video signal, i.e. the
same bandwidth limitations, maximum slew rate, maximum and
minimum voltage levels, and so on. Peak white level represents a ‘1’ and
a black level represents ‘0’.
VITC consists of 90 bits of data recorded serially during the chosen
vertical interval line. The first bit must occur between 10us and 11us
after the leading edge of the line sync pulse. It usually occurs at about
10.5us.
203 Sony Broadcast & Professional Europe

212. Part 19 – Timecode
0 1
200 +_ 50 ns
_
1 = 80 + 1 0 IR E
L e s s th a n 5 %
90%
Peak
to 50%
peak
(1 0 0 % )
_
0
1 0% 0 = 0 + 10
IR E
Figure 96 VITC signal details
The whole VITC code normally take up most of the vertical line, and the
last bit, bit 79, cannot occur less than 2.1us before the leading edge of
the next line sync pulse.
The VITC signal must also obey certain criteria which are outlined in Fig.
on page .
VITC sync bits
VITC has a ‘10’ sequence on bits 0 and 1 and every 10 bits after. These
allow the VITC reader to overcome timing jitter when reading the signal.
Sony Training Services 204

213. Broadcast Fundamentals
89
C R C C b it s
84
S y n c r o n is a ti o n b it s
80
U s e r b in a r y g r o u p 8
76
H o u r s te n s c o u n t
72
S y n c r o n is a ti o n b it s
68
U s e r b in a r y g r o u p 7
64
H o u r s u n i ts c o u n t
S y n c r o n is a ti o n b it s
60
U s e r b in a r y g r o u p 6
56
M i n u t e s te n s c o u n t
52
S y n c r o n is a ti o n b it s
U s e r b in a r y g r o u p 5
48
44
M i n u te s u n i ts c o u n t
S y n c r o n is a ti o n b it s
40
U s e r b in a r y g r o u p 4
36
S e c o n d s te n s c o u n t
32
S y n c r o n is a ti o n b it s
U s e r b in a r y g r o u p 3
28
24
S e c o n d s u n i ts c o u n t
S y n c r o n is a ti o n b it s
20
U s e r b in a r y g r o u p 2
16
12
S y n c r o n is a ti o n b it s
U s e r b in a r y g r o u p 1
8
F ra m e s u n its c o u n t
4
S y n c r o n is a ti o n b it s
0
Figure 97 The vertical interval timecode signal
205 Sony Broadcast & Professional Europe

214. Part 19 – Timecode
B it U s e (6 2 5 ) U s e (5 2 5 ) B it U s e (6 2 5 ) U s e (5 2 5 )
0 S ync 1 45 M in u te s u n its b it 3
1 S ync 0 46 U s e r g ro u p 5
2 F r a m e u n its b it 0 47 U s e r g ro u p 5
3 F r a m e u n its b it 1 48 U s e r g ro u p 5
4 F r a m e u n its b it 2 49 U s e r g ro u p 5
5 F r a m e u n its b it 3 50 S ync 1
6 U s e r g ro u p 1 51 S ync 0
7 U s e r g ro u p 1 52 M in u te s te n s b it 0
8 U s e r g ro u p 1 53 M in u te s te n s b it 1
9 U s e r g ro u p 1 54 M in u te s te n s b it 2
10 S ync 1 55 B in a r y g r o u p fla g 0 B in a r y g r o u p fla g 1
11 S ync 0 56 U s e r g ro u p 6
12 F r a m e te n s b it 0 57 U s e r g ro u p 6
13 F r a m e te n s b it 1 58 U s e r g ro u p 6
14 U n a s s ig n e d D r o p fr a m e fla g 59 U s e r g ro u p 6
15 C o lo u r fr a m e fla g C o lo u r fr a m e fla g 60 S ync 1
16 U s e r g ro u p 2 61 S ync 0
17 U s e r g ro u p 2 62 H o u r s u n its b it 0
18 U s e r g ro u p 2 63 H o u r s u n its b it 1
19 U s e r g ro u p 2 64 H o u r s u n its b it 2
20 S ync 1 65 H o u r s u n its b it 3
21 S ync 0 66 U s e r g ro u p 7
22 S e c o n d s u n its b it 0 67 U s e r g ro u p 7
23 S e c o n d s u n its b it 1 68 U s e r g ro u p 7
24 S e c o n d s u n its b it 2 69 U s e r g ro u p 7
25 S e c o n d s u n its b it 3 70 S ync 1
26 U s e r g ro u p 3 71 S ync 0
27 U s e r g ro u p 3 72 H o u r s te n s b it 0
28 U s e r g ro u p 3 73 H o u r s te n s b it 1
29 U s e r g ro u p 3 74 B in a r y g r o u p fla g 1 B in a r y g r o u p fla g 2
30 S ync 1 75 B in a r y g r o u p fla g 2 F ie ld m a r k fla g
31 S ync 0 76 U s e r g ro u p 8
32 S e c o n d s te n s b it 0 77 U s e r g ro u p 8
33 S e c o n d s te n s b it 1 78 U s e r g ro u p 8
34 S e c o n d s te n s b it 2 79 U s e r g ro u p 8
35 F ie ld m a r k fla g B in a r y g r o u p fla g 0 80 S ync 1
36 U s e r g ro u p 4 81 S ync 0
37 U s e r g ro u p 4 82 C R C C
38 U s e r g ro u p 4 83 C R C C
39 U s e r g ro u p 4 84 C R C C
40 S ync 1 85 C R C C
41 S ync 0 86 C R C C
42 M in u te s u n its b it 0 87 C R C C
43 M in u te s u n its b it 1 88 C R C C
44 M in u te s u n its b it 2 89 C R C C
Figure 98 VITC bits
Sony Training Services 206

215. Broadcast Fundamentals
Drop frame timecode
Drop frame timecode is only applicable to 525 line, NTSC based
systems. It arises from a basic problem associated with the fact that
NTSC systems do not count an exact number of frames per second.
Instead NTSC systems run at a rate of 29.97 frames per second.
This means that if timecode were to count at a rate of 30 frames per
second continuously it would eventually count an extra 108 frames per
hour, which amounts to about 3.6 seconds.
Over a 24 hour period, the maximum possible with timecode this extra
amounts to almost 1 minute!
Drop frame timecode was devised to allow for this problem by jumping
the timecode generator’s counter at certain specific times. Frames are
therefore dropped from the count.
Losing 108 frames from the timecode count is performed by first jumping
the timecode generator two frames at the beginning of every minute.
Therefore when the timecode generator reaches 09:42:59:29, for
instance, it will increment to 09:43:00:02 instead of 09:43:00:00, missing
out frames 09:43:00:00 and 09:43:00:01.
This effectively losses 120 frames per hour. This is too much so the
second counting scheme is used whereby the timecode generator is not
jumped at the beginning of every 10 minutes, i.e. at 00, 10, 20, 30, 40,
50 & 60 minutes.
Thus when the timecode generator reaches 09:49:59:29 it will increment
to 09:50:00:00 as normal instead of jumping to 09:50:00:02. This chops
12 frames off the 120 frames that would originally be lost using the first
counting scheme on its own to leave 108 frames lost in total per hour,
just the number required!
Which timecode am I using ?
The generally accepted standard timecode is LTC. There are a number
of reasons why. Firstly it was the first timecode to be proposed and used
extensively, and therefore had a head start over VITC in general
acceptance.
Secondly, as explained in the description on VITC, there is an inability to
standardise the VITC line on either field 1 or 2 of a video signal.
This is because VITC was proposed later and other uses had already
been found for the vertical interval lines before VITC could ‘grab’ any
particular vertical interval line for its own exclusive use.
Thus the VITC standard allows VITC to be put on any one of a number
of vertical interval lines, and VITC timecode reader/generators often
have to be altered to operate with the particular lines chosen, having first
made sure that it is free from use by anything else.
Thirdly, tapes used in an edit suite are often striped with continuous LTC
timecode. This then becomes the timing reference for the tape.
207 Sony Broadcast & Professional Europe

216. Part 19 – Timecode
As editing takes place, using insert edits, the LTC track will not be
recorded. Insert edits are made to the video and audio tracks only.
Thus LTC became accepted as the timecode that would be guaranteed
not to change in editing.
This third reason is a little difficult to justify in modern VTR’s where there
is the capability to replace VITC during video insert edits and guarantee
the same code is replaced after the edit.
However editors habits soon tended to regard LTC as the timecode to
depend on. Habits account for a lot, to the extent that the Sony DVW-
A500P series Digital Betacam machines, which could only playback
analogue Betacam SP tapes, where modified so that the LTC track
could be re-recorded to analogue tapes even though nothing else could.
Timecode use in video recorders
Most modern professional tape recorders have the capacity to read,
generate, record and playback both LTC and VITC.
When recording, timecode is used in two distinctive ways. The first is to
record the time of day the recording was made. Camcorders and
portable machines are often set to record the time of day as they are
often used to record news or sport, and this means the time is also
recorded.
The second is to provide continuous timecode throughout a tape. Studio
machines are often set to record continuous timecode. This means that
although an edit may take days or weeks to complete, and is made up
from many bits and pieces of video put together, the final master tape
will have a continuous seamless timecode, starting from zero at the
beginning of the tape.
Typical VTR timecode controls
There are a number of controls that can be found in a typical modern
professional video tape recorder. It is not possible to mention all of them
here, or the different names that might be given to each control within a
particular machine. However a few are considered here to give a rough
idea of the kind of things to look for.
Rec Run / Free Run switch
This switch allows a VTR operator to either select continuous timecode
recording (Rec Run), or time of day recording (Free Run). With the
switch set to Rec Run the machine’s internal timecode generator
increments only when the machine is recording.
With the switch set to Free Run the timecode generator continues to
increment all the time. If the time of day is to be recorded the timecode
generator then needs to be set to the time of day.
VITC On Off switch
As LTC is the industry accepted timecode, rather than VITC, it is often
possible to switch the internal VITC reader/generator off, if it isn’t in use.
Sony Training Services 208

217. Broadcast Fundamentals
VITC/AUTO/LTC switch
This switch is often included in a machine. It allows the user to either
force the machine to operate with VITC or with LTC, or to automatically
select the timecode signal it is able to find. If both VITC and LTC have
been recorded to tape, and played back at a variety of speeds, ranging
from stop to fast-forward at 50 times play speed, the machine’s
capacity to pick up both LTC and VITC is similar to Fig. below.
With the machine stopped LTC cannot be read off the tape. At play
speed both LTC and VITC can be read. As the machine’s speed is
increased eventually VITC becomes unreadable. (If the machine’s speed
is increased further,eventually it becomes difficult to pick up even LTC,
but that situation is beyond this discussion.)
Most machines will default to LTC, the preferred industry standard, at
any speed where both LTC and VITC are detectable off tape.
Drop Frame On Off switch
This switch is only found on 525 line (NTSC) machines to allow the
operator to work with continuous timecode, which ironically would not be
related to time at all, or drop frame timecode which keeps to time by
dropping frames at certain specific points.
625 line (PAL) machines do not include this switch, and a blank space
will often be found where one would be fitted.
Real Time to LTC/VITC User Bits switch
As explained on page , timecode includes groups of bits in amongst the
time address bit groups that can be used to store anything the user
might wish. This switch allows the user to store the real time (time of
day) into the user bits of either LTC of VITC on tape.
Timecode Reset, Advance or Set buttons
Machines generally have one or more buttons which allow the user to
either reset the machine’s internal timecode generator to 00:00:00:00, or
to set it at any predetermined count.
If running in Rec Run mode the user might reset to 00:00:00:00 before
beginning an edit session. Alternatively in Free Run mode the user might
preset the timecode generator to some time count, say just one minute
in the future, using something like the Advance button, and press Set the
moment that time comes up, to ‘synchronise’ the machine to the time of
day.
External User Bits switch
Some machines might include a switch to allow the user to either record
the user bit code input to the machine or to use the user bit code set
within the internal timecode generator, irrespective of what is happening
to the time address bits of timecode.
209 Sony Broadcast & Professional Europe

218. Part 19 – Timecode
VITC line selector
This selector may be a switch or a menu item. It allows the user to select
which of the vertical lines will be used by VITC.
LTC phase correction switch
A bi-phase mark signal can be read no matter which way round it is (see
page . However the LTC specification includes a polarity bit (bit 59 in
625 line, and bit 27 in 525 line systems). This bit ensures that each LTC
code begins low. Normally this makes no difference to the code, and
many VITC readers do not care if the code has been inverted. However
if two signals are to be edited together, and one of the LTC signals has
somehow been inverted there will be an error at the edit point.
This switch inverts the incoming LTC timecode signal if there appears to
be a problem with the edit.
The future
Most modern machines that can store video can also store timecode.
LTC was designed specifically for tape recorders, where the timecode
signal is recorded on a longitudinal track somewhere near the edge of
the tape. VITC is, however, not designed specifically for tape recorders,
but was designed purely for the video signal itself, regardless of where
that signal might be put into.
As technology progresses different types of video recorder are starting
to appear. The two types that should be considered are hard disk
recorders, and RAM recorders.
Hard disk recorders store video on a hard disk, or, more commonly, on a
hard disk array. Hard disk recorders generally have the disadvantage
that they cannot store anything like the amount of video that a tape
recorder can.
RAM recorders are even worse in this respect. Although very fast and
flexible, they generally store only a fraction of the amount of video a tape
recorder can.
However both hard disk and RAM recorders have no longitudinal track
and must therefore ‘fake’ an LTC from VITC. If tape recorders are ever
universally replaced by more advanced hard disk or RAM recorders, the
difference between LTC and VITC may become more confused.
Alternatively VITC may become the industry standard instead of LTC,
which may eventually be dropped altogether.
Sony Training Services 210

219. Broadcast Fundamentals
Part 18 SDI (serial digital interface)
Parallel digital television
During the late 70’s the television industry had begun to investigate the
idea of digitising television signals. At first such attempts were confined
to pieces of equipment such as standards converters, where the
analogue video signal was input to the unit, converted to a digital signal,
standards converted in the digital domain, converted to an analogue
signal again, and output from the unit.
The conversion to a digital video signal and back again was confined to
within the unit itself, and thus the methods of conversion tended to vary
from one manufacturer to another. It soon became obvious that by
maintaining the video signal in the digital form as it passed from one
piece of video equipment to another, the quality could be maintained at a
higher level than was normal for conventional analogue video signals. A
universal standard for digitising analogue video signals would thus be
very useful to the video industry.
In 1982 the CCIR met in Geneva. During their proceedings a
recommendation was made for digitising analogue component video
signals. The recommendation was CCIR 601 and, although only a
recommendation, very quickly became a de-facto standard throughout
the video industry.
CCIR 601 was a very basic document when it was published in 1982.
But it did promote thought throughout the video industry, and a few
manufacturers started to design equipment that conformed to it.
When the CCIR met again in 1986, CCIR 601 had been rewritten and
much improved. It had been altered in some places and added to in
others. A second recomendation was also drafted, called CCIR 656.
Although not a strict and absolute division, CCIR 601 described the
digital video signal itself and its conversion from an analogue signal, and
CCIR 656 described the physical interface, including the type of
connectors and cable.
CCIR 601/656
CCIR 601/656 involves the digitising of component analogue video
signals, and is commonly referred to as D1.
(A standard involving the digitising of composite video signals is
commonly referred to as D2. SDI is not based on composite video
signals or D2.)
There are two elements of CCIR 601/656 that need to be explained, the
quantisation levels and the sample structure.
Quantisation levels
The CCIR decided that component signals would be described at a
series of 8 bit binary numbers (words). This gives 256 possible
quantisation levels, 0 to 255.
211 Sony Broadcast & Professional Europe

220. Part 20 – SDI (serial digital interface)
H e x a d e c im a l
D e c im a l
B in a ry
255 FF 11111111
N ot used
P e a k w h i te 235 EB 1 1 1 01 01 1
2 2 0 q u a n ti s a t i o n l e v e l s
Y
B la c k 16 10 0001 0000
N ot used
0 00 00000000
S yncs not
d ig itis e d
255 FF 11111111
Peak + ve N ot used
240 F0 1 1 1 1 0000
c o lo u r le v e l
2 2 5 q u a n ti s a t i o n l e v e l s
R -Y
or 'B l a c k ' l e v e l 1 28 80 1 0000000
B -Y
P e a k -v e
c o lo u r le v e l 16 10 0001 0000
N ot used
0 00 00000000
Figure 99 CCIR-601 digitisation
In the case of the analogue Y signal, only the brightness part of the
signal is digitised, i.e. any sync pulses added to the Y signal are ignored.
Black level of the Y signal is set at 16 (10 hexadecimal or 00010000
binary). Peak white level is set at 235 (EB hexadecimal or 11101011
binary). The area between 0 and 15 is not used for digitising the Y
signal, and although CCIR 601 actually specifies that the Y signal may
occasionally be digitised beyond 235 for ‘super white’ signals, in almost
all cases the area between 236 and 255 is not used.
The samples resulting from the Y analogue component signal are
referred to as Y samples.
Sony Training Services 212

221. Broadcast Fundamentals
In the case of the analogue (R-Y) and (B-Y) signals, ‘black level’ or the
zero point is set at 128 (80 hexadecimal or 10000000 binary). The peak
positive excursion of each colour difference is set at 240
(F0 hexadecimal or 11110000 binary), and the peak negative excursion
is set at 16 (10 hexadecimal or 00010000 binary). As with the Y signal,
the area between 0 and 15 is not used for digitising the colour difference
signals, and the area between 240 and 255 is also not used.
The samples resulting from the (R-Y) analogue component colour
difference signal are referred to as Cr samples, the samples resulting
from the (B-Y) analogue component colour difference signal are referred
to as Cb samples.
Sample (word) structure
Samples are taken from the three analogue component signals at a rate
of 13.5 MHz. This frequency was chosen to give the greatest degree of
commonality between 625 and 525 line television standards.
Y Y Y Y Y Y Y Y Y Y
O rig in a l 1 3 .5 M H z s a m p le s
B -Y R -Y B -Y R -Y B -Y R -Y B -Y R -Y B -Y R -Y B -Y R -Y B -Y R -Y B -Y R -Y B -Y R -Y B -Y R -Y
F in a l 2 7 M H z s a m p le s C b Y C r Y C b Y C r Y C b Y C r Y C b Y C r Y C b Y C r Y
C o - s it e d tr i p l e t C o - s it e d tr i p l e t C o -s i te d tr i p l e t C o -s i te d t r i p l e t C o - s i te d tr i p l e t
S in g le Y S in g le Y S in g le Y S in g le Y S in g le Y
Figure 100 CCIR-601 sample structure
Each 13.5MHz ‘position’ therefore has a Y sample, a (B-Y) sample and a
(R-Y) sample. The first sample position of each line gives one Cb, Y and
Cr word. The Cb word is derived from the analogue (B-Y) component
signal. Likewise the Cr word is derived from the (R-Y) signal. Although
these three word occur sequentially in the CCIR 601/656 data stream, it
is important to remember that they originate from the same point on the
original image. They are therefore referred to as co-sited samples, or as
a co-sited triplet.
213 Sony Broadcast & Professional Europe

222. Part 20 – SDI (serial digital interface)
1 3 . 5 M H z s a m p l e p o s it i o n s
A n a lo g u e Y
c o m p o n e n t s ig n a l
A n a l o g u e ( B -Y )
c o m p o n e n t s ig n a l
A n a l o g u e ( R -Y )
c o m p o n e n t s ig n a l
F in a l 2 7 M H z s a m p le s C b Y C r Y C b Y C r Y C b Y C r Y C b Y C r Y C b Y C r Y
C o -s i t e d tr i p l e t C o - s i te d tr i p l e t C o - s i te d tr i p l e t C o - s i t e d tr i p l e t C o -s i t e d tr i p l e t
S in g le Y S in g l e Y S in g l e Y S in g le Y S in g le Y
Figure 101 Component analogue video and CCIR-601 sample comparison
The colour content of the second sample position is ignored, leaving a
single Y word.
The third sample position is treated as the first, the forth like the second,
as so on, giving a co-sited triplet, single Y, co-sited triplet, single Y, etc.
structure.
Thus CCIR 601/656 words occur at 27 Mhz, with the Y words at 13.5
MHz and each Cr and Cb word at 6.75 MHz.
CCIR 601/656 syncs
Up to this point the digital data stream does not contain any syncs. As
mentioned on the last page. the digital Y signal does not contain any
sync information as it is not digitised from the original analogue Y
component signal. Syncs need to be added somewhere to the
CCIR601/656 27MHz data stream so that the receiver can lock to the
incoming digital signal.
Sony Training Services 214

223. Broadcast Fundamentals
The CCIR recommended that a special sync signal would be added to
the beginning and end of every video line, even during the vertical
interval lines.
C b Y C r Y
C C IR 6 0 1 /6 5 6 C b Y C Y C b Y C Y C b Y C Y C b Y C Y
r r r r
d a ta s tr e a m
C o -s i te d t r i p l e t C o -s i te d tr i p l e t C o -s i te d tr i p l e t C o -s i te d t r i p l e t C o -s i te d t r i p l e t
S in g le Y S in g le Y S in g le Y S in g le Y S in g le Y
T
P re a m b le R
S
T i m in g R e fe r e n c e C o d e
Figure 102 CCIR-601 syncs
This signal is referred to as a Timing Reference Code (TRC). It consists
of four data words. The first three words are a preamble, a particular
sequence of data that will not occur anywhere else in the digital video
data stream. The forth word is referred to as a Timing Reference Signal
(TRS). This code will enable the receiver to ‘find its place’ in the digital
video signal.
As mentioned before the analogue Y signal is digitised between 16 and
235 and the colour signals are both digitised between 16 and 240. None
of the samples will ever reach 0 or 255 (00 or FF hex.). These values
are retained for the sync preamble words. The first preamble word if 255
(FF hex.), and the second and third preamble words are both 0 (00
hex.).
Timing Reference Signal structure
The TRS consists of 8 bits, as all other CCIR 601/656 words. The 8 bits
are specified as shown in the picture.
The most significant bit, bit 7, is always a ‘1’.Bit 6 is the F bit. It defines
which video field this particular TRS is in. A ‘0’ signifies field 1, and a ‘1’
signifies field 2.
Bit 5 is the V bit. It defines whether the TRS is in the active part of the
video field, or the vertical blanking interval. A ‘0’ signifes the active
portion, and a ‘1’ signifies the vertical blanking interval.
215 Sony Broadcast & Professional Europe

224. Part 20 – SDI (serial digital interface)
Bit 4 is the H bit. It defines whether the TRS is at the beginning or the
end of the line. A ‘0’ signifies the start of active video (SAV), and a ‘1’ the
end of active video (EAV).
M .S .B . b it 7 1 0 0 1 0 : F ie ld 1
F ie ld b it 1 : F ie ld 2
1 0 0 F
0 : A c ti v e fi e l d p o r ti o n
1 0 0 V V e rtic a l b it 1 : F ie ld b la n k in g p o rtio n
1 0 0 H 0 : S ta r t o f a c ti v e v i d e o ( S A V )
H o riz o n ta l b it 1 : E n d o f a c ti v e v i d e o ( E A V )
1 0 0 P3
1 0 0 P2
P r o t e c ti o n b i t s
1 0 0 P1
M .S .B . b it 0 1 0 0 P0
T i m i n g R e fe r e n c e S i g n a l
F irs t p re a m b le w o rd
S e c o n d p re a m b le w o rd
T h ird p re a m b le w o rd
Figure 103 CCIR-601 timing reference structure
Bits 3 to 0 are Hamming code protection bits P3 to P0. A different
combination of these four bits occurs for each combination of F, and H
bits. The Hamming distance between each combination means that it is
possible to detect and correct or one bit errors, and simply detect two bit
errors, in any TRS. The allocation of protection bits is shown in the
picture.
M .S .B . ( a lw a y s 1 ) 1 1 1 1 1 1 1 1
F 0 0 0 0 1 1 1 1
V 0 0 1 1 0 0 1 1
H 0 1 0 1 0 1 0 1
P3 0 1 1 0 0 1 1 0
P2 0 1 0 1 1 0 1 0
P1 0 0 1 1 1 1 0 0
L .S .B . P 0 0 1 1 0 1 0 0 1
Figure 104 CCIR-601 TRS protection bits
How a D1 reciever locks to an incoming D1 signal
Imagine a receiver is switched on and a D1 source is connected. The
first thing it will do is search for any words that have the value 255, i.e.
all 8 bits are ‘1’.
When the receiver finds this word it checks that the next two words are
0, i.e. all 8 bits of each word are ‘0’.
Sony Training Services 216

225. Broadcast Fundamentals
If these three words check out, the forth word is placed into a special
register which examines each bit individually. Bit 7 must be a ‘1’.
The combination of F’ V and H bits must also correspond to the
combination of four protection bits.
If any part of this process fails the receiver disregards these words and
starts the search again.
If all this checks out the receiver looks at the H bit. If it is a ‘0’ the
receiver knows it is at the beginning of the video line. If it is a ‘1’, it is at
the end of the line. The receiver is now said to be line locked.
The receiver now knows where each TRS will occur because it knows
how many clock cycles occur between each one. It now looks at the F bit
of each TRS until this bit changes state, i.e. either changes from a ‘0’ to
a ‘1’ or from a ‘1’ to a ‘0’.
A change from a ‘1’ to a ‘0’ signifies the beginning of field 1. A change
from a ‘0’ to a ‘1’ signifies the beginning of field 2. The receiver is now
frame or field locked, and the locking process is completed.
The V bit is not actually required for locking, but is used by some
systems to check the actual position of vertical blanking.
The receiver then simply checks that each TRS after this is valid. If there
is an error some systems ignore the complete TRC and check the next
one. Other systems are not so ‘clever’ and immediately fall out of lock.
The CCIR 601/656 interface driver
As shown already CCIR 601/656 signals are transmitted at 27 MHz. This
kind of frequency is far too high to transmit over any great distance with
‘normal’ logic circuits like TTL and CMOS drivers.
D i ff e r e n t ia l t r a n s m it t e r T w i s te d p a i r D i ffe r e n t i a l r e c e iv e r
In p u t O u tp u t
0 v o lts S y s te m g ro u n d C a b le s c re e n
Figure 105 CCIR601/656 ECL driver
ECL (emitter coupled logic) is capable of operating at higher frequencies
than TTL or CMOS circuits can. Differential ECL is capable of
transmitting over long distances with the proper sheilded cable.
Thus CCIR 601/656 signals require one differential ECL driver for each
bit and a further driver for the clock. Each driver has two output wires,
217 Sony Broadcast & Professional Europe

226. Part 20 – SDI (serial digital interface)
giving a total of 18 wires for 8 bit CCIR 601/656 video (16 for video and
2 for clock).
The CCIR 601/656 connector
CCIR 656 recommends a ‘D’ 25 connector be used to connect digital
video signals. (However, although the CCIR recommended slide locks
be used, most users prefer screw fitting D25 connectors, because they
tend to be more secure and easier to fix in place than slide locks.) This
pinout for this connector is shown in the picture.
P in 1
P in 1 4
P in 2 5
P in 1 3
P in P in
F u n c tio n F u n c tio n
num ber num ber
1 C lo c k + 14 C lo c k -
2 S y s te m g r o u n d 15 S y s te m g r o u n d
3 D a ta b i t 7 + 16 D a ta b i t 7 -
4 D a ta b i t 6 + 17 D a ta b i t 6 -
5 D a ta b i t 5 + 18 D a ta b i t 5 -
6 D a ta b i t 4 + 19 D a ta b i t 4 -
7 D a ta b i t 3 + 20 D a ta b i t 3 -
8 D a ta b i t 2 + 21 D a ta b i t 2 -
9 D a ta b i t 1 + 22 D a ta b i t 1 -
10 D a ta b i t 0 + 23 D a ta b i t 0 -
11 S p a re b it A + 24 S p a re b it A -
12 S p a re b it B + 25 S p a re b it B -
13 C h a s s i s g r o u n d ( s h ie l d )
E x a m p le o f c o n n e c to r
Figure 106 CCIR601-656 D25 parallel connector
Pins 1 and 14 are allocated to the 27 MHz clock positive and negative
differential ECL respectively.
The clock is separated from the data pins by two system ground pins 2
and 15. There are eight data pin pairs starting with pins 3 and 16 for
data
Sony Training Services 218

227. Broadcast Fundamentals
bit 7, and finishing with pins 10 and 23 for data bit 0, positive and
negative differential ECL respectively.
Pin 13 is allocated as a chassis ground and can be connected to the
connector shell itself.
The increase to 10 bits
Pins 11 and 24, and pins 12 and 25 were originally allocated by CCIR as
two spare differential ECL pairs.
It took very little time for the video industry to start using these two extra
bits to extend the original 8 bit samples specified by CCIR to 10 bits.
However it is important to remember that these two bits are used as half
and quarter resolution bits. i.e. below the decimal point.
Thus Y samples now extend from 16.0 to 235.75 in 880 increments of
0.25 each. Cr and Cb samples now extend from 16.0 to 240.75 in 900
increments of 0.25 each.
219 Sony Broadcast & Professional Europe

228. Part 20 – SDI (serial digital interface)
Serial digital television
Parallel D1, based on CCIR 601/656, became widely used by the
professional video industry in the latter part of the 80’s. However there
was resistance to using D25 connectors. Compared to the BNC
connectors and co-ax cables used for analogue video, these connectors
were more expensive and unreliable, and the cable required was
expensive and heavy. It was also found that parallel D1 could not
reliably be made over long distances.
CCIR 656 specified a serial interface, but this was based on 8 bit
samples, and the industry had already moved ahead to 10 bit samples.
The CCIR 656 serial interface could not be used.
In the latter part of the 80’s Sony introduced two devices that helped set
an industry de-facto standard for serial D1 digital video standard, the
SBX1601A and the SBX1602A.
SBX1601A is a parallel to serial converter for 10 bit D1 signals, an
SBX1602A is a serial to parallel converter.
Sony Training Services 220

229. Broadcast Fundamentals
Serial digital audio
At the same time as digital video was developing, advancements were
made to audio as well. However because the frequencies of audio were
basically lower than those of video development was more rapid and a
number of parallel digital audio standards became popular comparatively
quickly. In the professional world a sampling frequency of 48 kHz
became popular and in audio CD’s a sampling frequency of 44.1 kHz is
used. 32 kHz was also used. Sample widths of 8 bits were used on
cheaper older systems, but in the professional arena 18, 20 or 24 bit
were becoming popular.
The AES and EBU collaborated to draw up a standard for transmitting
audio signals through a serial digital channel. With so many different
parallel audio standards already in use,<N>any serial standard had to
somehow be able to encompass all the popular professional parallel
standards, and have some method of informing the receiver which
parallel standard was being transmitted.
The standard, known as the AES/EBU/IEC 958 standard, or simply as
AES/EBU audio, contains two channels of audio, channel A & B, with a
maximum sample size of 24 bits, and a maximum sample frequency of
48kHz, neatly covering the most demanding of the popular professional
parallel standards.
Channel coding
The channel coding method chosen for AES/EBU audio was Bi-phase
mark, otherwise known as the Manchester 1 code. This is the same
channel coding method as is used for longitudinal timecode and for
Ethernet in computer networks.
As shown in the picture, Bi-phase mark places a transition at every bit
boundary, and a transition in the middle of each bit period for each ‘1’
bit.
Figure 107 Bi-phase Mark signal structure
Thus Bi-phase mark is not only polarity and direction independent, but is
also self clocking, i.e. if there are phase changes during transmission, all
an AES/EBU receiver has to do is to look for regular transitions
corresponding to the bit boundaries, and once locked to them, search for
any bit periods with a transition in the middle.
Any with no transitions are ‘0’s, and those with, are ‘1’s.
221 Sony Broadcast & Professional Europe

230. Part 20 – SDI (serial digital interface)
Signal structure
The standard organises data into serial blocks. Each block contains 192
frames. Each frame contains two sub-frames, one for channel A and one
for channel B. Every subframe contains one audio sample. This is
shown in following picture.
Figure 108 AES/EBU digital audio signal structure
Sync bits (4 bits)
These first four bits are used to define the beginning of a sub-frame.
They are different to normal data because they violate the laws of the bi-
phase mark channel coding system.
Normally in bi-phase mark there is always a transition at every clock
cycle, i.e. between one bit and the next. However syncs drops two of
these clock transitions at specific points, as shown in the picture.
There are three forms of syncs. Form X defines the start of sub-frame A,
form Y defines the start of sub-frame B, and form Z defines the start of
the block (which is also a sub-frame A).
Thus the receiver searches for these ‘illegal’ portions of the signal, and
decodes them to find out if it is a form X, Y or Z sync. From that it is able
to determine where it is in the AES/EBU signal structure.
Sony Training Services 222

231. Broadcast Fundamentals
Figure 109 AES/EBU audio syncs
Auxillary bits (4 bits)
These four bits serve two main purposes. They can be used to extend
the twenty bits of audio sample to 24 bits to give greater resolution to the
audio samples.
Alternatively they can be used to provide an extra channel of audio. By
combining three consecutive groups of auxiliary bits together you can
make an extra audio channel with only twelve bit resolution at a third the
sampling frequency.
Using the auxiliary bits for an extra audio channel for both sub-frames
gives a four channel audio transmission system with two high quality
channels and two low quality channels.
Audio data (20 bits)
The next twenty bit are the audio sample itself. The LSB is just after the
last auxiliary bit and the MSB is just before the V flag bit.
Flags (4 bits)
V flag - Validity - This flag indicates that the audio sample data is error
free.
U bit - User - This bit is not defines and can be used for any purpose.
C bit - Channel status - See the section below.
P flag - Parity - Parity for all bits in a sub-frame except the four sync bits.
Channel status bits
The third bit of the flags in each sub-frame is for channel status. Thus,
with 192 channel A sub-frames in a block there are 192 channel A status
bits. Channel B also has 192 channel status bits, one for each subframe
throughout the entire block.
The channel status bits define the type of audio being transmitted. The
table in the picture shows the allocation of the 192 channel status bits.
The same table can be used for either channel A or channel B.
223 Sony Broadcast & Professional Europe

232. Part 20 – SDI (serial digital interface)
Figure 110 AES/EBU channel status bits
Channel status for Digital Betacam
Digital Betacam uses a very specific type of AES/EBU. Staring at the top
of the table, Digital Betacam uses professional use of the channel status
block, thus bit 0 of byte 0 should be ‘1’.
Bits 2, 3 & 4 of byte 0 define if the analogue signal from which the digital
signal was sampled was emphasised, so the de-emphasis can occur
when the signal is converted beck to analogue.
Digital Betacam uses either no emphasis or CD emphasis. It does not
use CCITT emphasis.
Bits 6 & 7 of byte 0 define the sampling frequency of the AES/EBU audio
signal. This can be either 32, 44.1 or 48 kHz.
Digital Betacam uses a digital audio sampling frequency of 48 kHz only.
Sony Training Services 224

233. Broadcast Fundamentals
Bits 0, 1, 2 & 3 define how the two channels are related to one another.
Two channel mode means the two channel are not realated to one
another at all and are two independent channels. Single channel mode
means channel B is not used at all i.e. even though the sync and flag
bits are as they should be, all the audio sample data bits are ‘0’.
Primary/secondary channel mode means that channel A is copied to
channel B to increase the chance that the signal will get through without
error. Stereophonic mode means the two channels are related and
should not be separated. Any processing that is applied to one channel
A should also be applied to channel B. Channel A is taken as the left
channel.
Digital Betacam uses two channel mode. If the user chooses to make
the two channels a stereo pair Digital Betacam will still treat them an two
independent channels.
The last useful part of the table is bits 0, 1 & 2 of byte 2. These define
the maximum sample length. The sample can be defines as 20 bits, and
the auxiliary bits are simply not used and left as ‘0’. The sample can be
up to 24 bits, with the auxiliary bits being used to extend the normal 20
bits to 24 bits. The sample can also be 20 bits but using the auxiliary bit
as the coordination channel.
Digital Betacam uses AES/EBU as 20 bits, with the auxiliary bits not
used.
Relationship between the two channels
Both channels in AES/EBU audio must have the same sampling
frequency. This is a basic requirement.
If one channel is defined as a stereo pair channel then the other channel
must also be stereo. Further more if the two channels are a stereo pair
every other aspect of the two channels must be the same, i.e. Their
channel status bits must be identical.
If one channel is defined as a primary/secondary channel, the other
channel must also be, just as for stereo. However in this case the audio
data itself must also be identical.
If both channel are in two channel mode other aspects of the channel
status bits may also differ. For instance channel A may be emphasised,
channel B not. The sample length may also differ.
AES/EBU audio bit rate frequency
The final bit rate for AES/EBU audio depends on the original sample
frequecy. Taking the Digital Betacam particular use of AES/EBU audio
the sample frequency is 48 kHz, there are 32bits in each sample sub-
frame and two channels. Thus the bit rate can be calculated from the
following simple equation :-
48 x 1000 x 32 x 2 = 3.072 Mhz
225 Sony Broadcast & Professional Europe

234. Part 20 – SDI (serial digital interface)
SDI
SDI (Serial Digital Interface) is a particular implementation of serial
digital video based on CCIR 601 and CCIR 656, which also incorporates
two channels of AES/EBU digital audio (four audio channels altogether).
SDI has become very popular within Sony broadcast, professional and
industrial equipment as a method of carrying video and audio from one
piece of equipment to another.
As shown before the CCIR 601/656 digitises the whole analogue video
signal. This not only includes the active part of the signal, i.e. the picture
itself, but also the vertical and horizontal blanking intervals as well. This
part is always black and the digital data during this part of the signal is
always the same.
This represents a waste in terms of information density. The blanking
interval would be best used to contain useful information. This is where
embedded audio comes in.
Embedded audio
Given the bit rate frequency of AES/EBU audio as used in Digital
Betacam, 3.072 Mhz, and the frequency of serial digital video, 270 MHz,
there is actually enough space during the horizontal blanking interval of
a video signal to pack nearly twenty channels of audio.
Infact SDI uses only a portion of the available space during the
horizontal blanking interval to embed two AES/EBU audio channels, or
four audio channels.
Video index
The final addition to the SDI signal as far as Digital Betacam is
concerned id video index. This system replaces vertical interval
subcarrier (VISC) and colour frame identification (CFID) used in
Betacam SP.
Video index uses line 11 and 324 in 625 line video (PAL). These two
lines are within the vertical interval of the video signal.
Bit 2 of each colour sample is used so that video index is still transmitted
even in 8 bit digital video.
Sony Training Services 226

235. Broadcast Fundamentals
Part 19 Video compression
The reason for compression is space, or lack of it. In a transmission
system like an aerial transmitter, satellite link or cable television link,
there is a constant fight to get as much use out of the link as possible.
Television companies want to put as many different television channels
into one link as they can.
Traditional analogue signals
In traditional analogue transmission systems there is a severe limit to the
number of channels that can be squeezed into each link. A traditional
analogue satellite link can take just one channel. It is much the same for
cable. Aerial transmission systems can take a few analogue channels
but each one takes up a lot of the aerial’s capacity.
Compressing analogue signals
There are in fact various compression techniques used to compress
analogue signals and save bandwidth. The analogue colour difference
signals are compressed from the original (R-Y) and (B-Y) signals to U
and V signals before combining them with the luminance, Y, signal to
make a final composite signal. This is a form of compression.
The problem of analogue compression
An analogue signal is always susceptible to change and interference by
cross talk, noise and other unwanted additions. In analogue
transmission you try to make the analogue signal big enough or different
enough from the noise or cross talk to make it easy to differentiate them
at the receiving end.
Analogue compression tends to push the original signal down into the
noise, making it more difficult to separate later on.
Analogue to digital conversion
Analogue to digital conversion involves converting the television signal
(video and audio) from a continuously changing signal to a signal
comprising a series of defined numerical values.
Converting a television signal (video and audio) into a digital signal does
not save any space in the transmission system, in fact digital signals
demand greater bandwidth than analogue signals.
As digital signals use an entirely different method of encoding they are
less susceptible to interference by the kind of unwanted additions that
plague analogue signals.
Compressing digital signals
Compressing digital signals involves replacing the digital information
with a smaller amount of different digital information. The essential point
is that the original data and the compressed data are both digital data,
227 Sony Broadcast & Professional Europe

236. Part 21 – Video compression
and therefore just as resistant as each other to interference from
analogue noise.
Digital errors in transmission
Digital data is very resilient to the kind of interference that affects
analogue signals. However if analogue noise is too excessive it can alter
digital data as well.
If there is a break in the transmission path this can cause digital data to
be corrupted.
Compensating for digital errors
There are techniques for removing errors from transmitted digital data.
You can either use error correction techniques to replace the error data
with the original data, or, failing that, use concealment techniques to
replace the error data with data calculated to be similar to what the data
should have been.
The advantage of digital compression
Analogue compression tends to reduce the signal’s resolution and force
it into the noise. This reduces its quality.
Digital compression can be compressed as much as you like. The
compressed data is still digital data, and is still not affected by analogue
noise.
As mentioned, excessive analogue noise does alter digital data.
However the amount of noise required is greater than with analogue
signals.
Both analogue signals and digital signals can be corrupted by
intermittent breaks in the transmission link. However there are
techniques for correcting or concealing errors in digital data.
Entropy and redundancy
Any signal, data, or multimedia material of any sort, may be divided into
two basic types, entropy and redundancy.
Entropy
Entropy is another word for chaos. It is essentially unpredictable. In
many systems entropy is something bad. Something to be eliminated.
Entropy destroys order and introduces uncertainty.
However in multimedia signals entropy is the information we want to
keep. It represents the interesting parts of the data or signal.
Redundancy
Redundancy is something that can be dropped or eliminated. In
multimedia signals redundancy represents the parts of a signal that are
entirely predictable, and repetitious. If any part of a signal or data can be
Sony Training Services 228

237. Broadcast Fundamentals
predicted then it is unnecessary to include it in the signal or data at all. It
is, be definition, redundant.
Entropy and redundancy is video signals
If we assume a full quality digital video signal is the kind of signals
specified by CCIR601 and as used by SDI links, the total data rate is
270Mbps.
Thus digital video comprises a proportion of entropy and redundancy
totalling 270Mbps.
For the simplest possible video signal all of this 270Mbps is redundant
data. For the most complex all of it is entropy.
In practice video signals are never that simple or complex. However it is
possible to draw a very simple graph showing the proportion of entropy
and redundancy for video signals of all different types, from the simplest
to the most complex.
In reality video signals never become so complex that they comprise
entirely of entropy. There is always some redundancy somewhere in the
signal. It may be spatial redundancy, because each frame comprises
very little detail, or it may be temporal redundancy where each frame is
similar to the last.
Considering a 2:1 compression ratio
Imagine you wanted to compress the video signal to a half of its original
size, i.e. a 2:1 compression ratio. If it was reasonably simple it would be
easy to reduce the signal by the required amount by removing a
proportion of the redundancy. In many cases you may not even need to
remove all the redundancy.
229 Sony Broadcast & Professional Europe

238. Part 21 – Video compression
However it the video signals became particularly complex, the amount of
entropy may exceed half of the original signal. In this case you would
have to throw away some of the entropy to reduce the signal by the
amount required.
The key point is that there is no loss except for the most complex video
signal. In many cases the video signal can be compressed and
decompressed with no loss of details at all.
Considering a 5:1 compression ratio
Now imaging you want to compress the same video signals by a 5:1
compression ratio, i.e. to a fifth of its originals capacity.
In this case the video signal needs to be that much simpler for there to
be no loss in compression. For anything but the most simple video
signals you would need to throw away some entropy in order to achieve
the required compression ratio.
The purpose of any compression scheme
The sole purpose of any compression scheme is to separate entropy
from redundancy, keep the entropy part and throw away the redundancy
part.
If the compression scheme is poor it will not be able to find all the
redundancy in the video signal and will not be able to compress the
signal enough without needlessly throwing away entropy.
A good compression scheme will be able to use a selection of tools and
techniques to find enough redundancy so that it is able to achieve the
required compression ratio without having to throw away any entropy.
Lossless and lossy compression
There are two basic types of compression systems used, lossless
compression and lossy compression. Both have advantages and
disadvantages, and are appropriate to the kind of data or signals they
are compressing.
Lossless compression
As the name suggests, lossless compression is a scheme where there is
no loss in the compression / decompression path. After decompression
you end up with exactly the same data or information as you started with
before you compressed.
Lossless compression is vital for computer data compression. If you are
compressing an executable, or peripheral driver file, you cannot accept
any loss during compression at all. If just one bit is wrong after
decompression the compression system has failed.
Lossless compression will remove as much redundancy as possible, but
will not remove any of the entropy.
The disadvantage of lossless compression is that you cannot specify a
compression ratio. It will depend on the kind of information that needs to
Sony Training Services 230

239. Broadcast Fundamentals
be compressed. A simple 1Mb file, like a bitmap image of a snowy scene
where most of the image is white, will compress much more than a
complex 1Mb file, like an executable file.
Examples of lossless compression schemes
A perfect example of a lossless compression scheme is PKZip or
WinZip. These programs are designed to compress computer files by as
much as possible without loss.
Lossy compression
The advantage of lossy compression is that you may specify the
compression ratio. This is vital in multimedia transmission systems,
satellite and cable television links, digital video and audio tape and disk
storage systems.
Examples of lossy compression schemes
The most popular lossy compression schemes for video signals is
MPEG and DV. For audio they are MP-3 and ATRAC.
There are others. For video tape there is Digital Betacam, Betacam SX,
IMX, and derivatives of DV like DVCAM and DVC Pro. For computers
there are higher compression ratio schemes like .AVI and Real Motion.
There are other lossy compression schemes intended for still images,
like the JPEG, GIF, and TIFF formats.
Inter-frame and Intra-frame
For video there are two basic methods of compression, inter-frame and
intra-frame compression.
Inter-frame compression
Inter-frame compression looks at the difference between frames. It is not
actually a compression scheme at all, but a way of processing the video
signal before compression takes place in order to achieve more efficient
compression.
There are three types of inter frame, the P (predicted) frame, the B
(between) frame and the R (reverse) frame.
The P frame
The P frame is a frame derived from a comparison between the frame in
question and the previous frame.
In most cases the difference between a frame and the previous frame is
small.
The B frame
The B frame is a frame derived from a comparison between the frame in
question and the average of the previous frame and the following frame.
231 Sony Broadcast & Professional Europe

240. Part 21 – Video compression
R frame
The P frame is a frame derived from a comparison between the frame in
question and the next frame.
Intra-frame compression
Intra-frame compression is a scheme for reducing the amount of data in
one video frame. The data is reorganised in order to separate
redundancy from entropy and the redundancy is discarded.
The toolbox
Intra-frame compression uses a series of tools. This set of tools is
sometimes referred to as a toolbox. The tools that are generally used
are DCT (discrete cosine transform), quantisation zig-zag scanning, run
length coding, variable length coding and data buffering.
What is DCT?
DCT (discrete cosine transform) is a method of describing data as a
discrete weighted set of cosines. Put simply the original data is
described not as its original samples or data words, but as how these
samples change.
The church organ
In understanding DCT a good place to start is to look at the church
organ.
Organ pipes produce a nice clean tone. The note they produce is pretty
close to a sine. Organists add the sound from other pipes to the note
they are actually playing to produce different tones. Organists call these
stops.
Turning the church organ upside down
It should be possible to analyse any of the many different sounds and
tones coming from a church organ and work out exactly which stops the
organist had pulled out.
The scientific equivalent
In fact it is possible to do this with any tone. You can break down any
note from any musical instrument into a series of sine waves. The lowest
frequency sine is the note itself, often called the fundamental. The rest
are all higher frequencies, normally multiples of the fundamental,
generally called the harmonics.
Taking things one stage further, it is actually possible to break any
repetitive or periodic signal into a series of sine waves or cosine waves,
or, to be more exact, a series of sine and cosine waves with associated
phases.
This process is commonly referred to as a Fourier analysis or a Fourier
transform.
Sony Training Services 232

241. Broadcast Fundamentals
The Fourier transform
A Fourier transform is a method of describing any periodic signal as a
series of sines and cosines.
When you do a Fourier transform of pure sine wave you get a single
spike at the frequency of the wave.
A violin creates something approaching a sawtooth waveform. A
sawtooth is made up from the fundamental and all the harmonics. The
amplitude of the fundamentals falls exponentially as the frequencies
increase.
A classic rich organ sound approaches a square wave. This wave is
made up from the odd harmonics of the fundamental. The amplitude of
these harmonics falls with increased frequency in the same way as it
does for the sawtooth wave.
233 Sony Broadcast & Professional Europe

242. Part 21 – Video compression
In both the pure sawtooth and square waveforms there are an infinite
number of harmonics. This is impossible. The high frequency harmonics
will always be lost no matter how small the loss is.Thus it is also
impossible to create an absolutely perfect sawtooth or square wave.
The mathematics of Fourier transforms
The mathematical formulae used for Fourier transforms looks hideous
but is actually quite simple. The basic Fourier transform expression is :-
∞
F ( s) = ∫
−∞
f (t ) exp − j 2πst dt
Where F(s) is the Fourier transform, f(t) is the original signal. The “exp”
is a neat mathematical shorthand for a particular addition of a sine and
cosine, which goes like this :-
exp − j 2πst = cos(2πst ) + j sin(2πst )
The reason for this is that although a signal is made up from a number of
pure sine waves, the amplitude and phase of each one may be different.
The expression above is a way of describing a sine wave at any
amplitude and phase by describing it in terms of a sine and cosine in the
real and imaginary plane.
Sony Training Services 234

243. Broadcast Fundamentals
The integral is simply an infinite sum of all the “f(t) exp” for every point in
time (t) from the beginning of time (-∞) to the end of time (∞).
The pros and cons of Fourier transforms
Fourier transforms work well for continuous periodic signals, and are
therefore very useful for things like analysing music, signals from deep
space, or vibration analysis on racing cars or aircraft.
However they do not work well for digital signals where there are a
series of discrete samples in time. For this we need a discrete version of
the Fourier transform.
The Discrete Fourier Transform (DFT)
DFT is a special form of Fourier transform for periodic signals made up
from discrete values, i.e. digital signals.
With analogue signals we use normal Fourier transforms which use
integrals. This is because analogue signals are continuous and any
summing analysis like Fourier transforms need to imagine the sum made
up from an infinite number of infinitely small time periods all along the
analogue signal we are studying.
We are not interesting in taking infinitely small points in time any more.
The samples now occur at specific points in time related to the sample
rate, or period. Therefore the integral in a normal Fourier transform can
be replaced by a simpler summing function ‘∑’.
A few other changes will take place. Rather than talking about a
continuous signal we are now looking at discrete samples. Thus let us
replace f(t) with f(k). Likewise the continuous Fourier function, F(s), will
be replaced by a discrete one, F(r).
The normal integral based Fourier transform changes in the case of the
DFT to :-
k =∞
F (r ) = ∑ f (k ) exp
k = −∞
− j 2πrk
Taking a set number of samples.
We have managed to get rid of the integral and replace it with something
more applicable to samples. However we are still stuck with this concept
of the signal stretching from the beginning of time to the end of time, i.e -
∞ to ∞.
We can adjust the expression so that we can find the DFT of a set
number of samples. To do this we imagine the DFT will repeat the
samples we are interested in again and again, and take the sum across
just the samples we want. We do not want to decide how many at the
moment so we will give the quantity of samples a letter. N would be
good.
This means that the DFT will change to :-
235 Sony Broadcast & Professional Europe

244. Part 21 – Video compression
k = N −1 − j 2πk
1
F (r ) =
N
∑
k =0
f (k ) exp N
It is conventional that the first sample is “0” and the last is (N-1). There is
no particular reason for this other than it makes the equations a little
simpler, but it means that if, for instance, you have 8 samples in your
selection “k” will go from “0” to “7”.
The judder problem of DFT
There is a problem with DFT over a set number of samples. We are
imagining the samples repeating again and again. However there is no
guarantee that the last sample will be anywhere near the same value as
the first sample.
This will produce a judder in the signals as it repeats. This judder is
energy and thus has its own harmonics that are nothing to do with the
original samples.
Sony Training Services 236

245. Broadcast Fundamentals
Discrete Cosine Transform (DCT) solution to judder
One neat way of removing the sudden jump between the last sample
and the first sample is to take twice as many samples, with half of them
being a mirror image.
Mirrored sines and cosines
An interesting thing happens to sines and cosines when they are
reflected about zero. The sines cancel out!
Put mathematically sin(-x) = -sin(x) and cos(-x) = cos(x)
237 Sony Broadcast & Professional Europe

246. Part 21 – Video compression
This can help to reduce the complexity of the DFT. We need only
consider the cosine parts of the transform.
Thus by taking twice the number of samples and considering only the
cosines the original expression drops to :-
2 k = N −1 (2k + 1)rπ
F (r ) = ∑
N k =0
f (k ) cos
2N
What does the result of DCT look like?
The original Fourier transform shows the frequencies that make up the
original signal. DCT does the same thing. The samples are replaced by
numbers that describe frequencies that make up the original sample
data. These numbers are called coefficients.
The samples are discrete, so the number of frequencies is also discrete.
In fact there are as many DCT coefficients as there are samples. If you
take more samples in your analysis, you get more coefficients.
The formula also places the lowest frequency coefficient first, in place of
the first sample. Remember that we are imagining that the set of
samples we are doing a DCT of is repeated again and again, because
DCT is based on Fourier transforms which only work on periodic signals
that go on for ever. The odd conclusion from thi is that the lowest
frequency coefficient is actually the average level of the samples,
sometimes called the DC coefficient. It is the second coefficient that is
the same as the fundamental frequency we looked at with church organs
right at the beginning.
DCT in video
Imagine a video picture made up of individual pixels. The pixels are laid
out in rows and columns.
As we have seen DCT operates on a group of samples. It cannot
operate on just one sample, because it is analysing how the samples are
changing, and what frequency coefficients the samples are made from.
Sony Training Services 238

247. Broadcast Fundamentals
The same thing applies to DCT as it is used in video. We need to
analyse a group of pixels in both the horizontal (x) and vertical (y)
directions.
Simple 2 by 2 DCT block
The smallest group of pixels DCT can work with is a 2 by 2 block. So let
us look at how this might work.
The whole picture is now split into 2 by 2 blocks. DCT will then remove
each block and replace it with 4 coefficients that describe how the
original pixels vary across the block.
The top left corner of the block will now hold the DC coefficient. The top
right pixel is replaced by a number describing how all 4 pixels are
changing in the horizontal direction. This number is called the horizontal
AC coefficient.
The bottom left pixel is replaced by a number describing how all 4 pixels
are changing in the vertical direction. The principles are the same as the
top right pixel, but for the vertical direction. This number is called the
vertical AC coefficient.
The bottom right pixel is replaced by a number describing how the 4
pixels are changing at a 45 degree angle from top left to bottom right.
This is called the diagonal AC coefficient.
Thus the 4 DCT coefficients describe the original pixels in 2 dimensions
in much the same way as a simple DCT operates on simple 1
dimensional samples. There is nothing lost in making this transform.
Describing the pixels in terms of their frequency coefficients, i.e how
they vary, is just as accurate a method as describing them as individual
pixels.
239 Sony Broadcast & Professional Europe

248. Part 21 – Video compression
A practical 8 by 8 DCT block
In virtually all practical compression schemes the simple 2 by 2 DCT
block is too small. There are benefits of selecting a larger block size.
The practical size most commonly used is an 8 by 8 DCT block,
transformed from 64 original pixels.
The basic principles are the same. The top left corner of the DCT block
is the DC coefficient, in the same basic way as it was in the simple 2 by
2 block.
The 8 coefficients along the top describe how the original 64 pixels vary
in the horizontal direction, in the same way as the top right coefficient did
in the simple 2 by 2 block. However there are now 8 possible
coefficients. Thus the left most of these coefficients describes the overall
low frequency horizontal variation of the 64 pixels. This is the
fundamental in the 1 dimensional DCT we looked at before.
The next coefficient to the right corresponds to the next highest
frequency change, and so on until the top right most coefficient
describes the highest frequency horizontal change in the original 64
pixels. These are the same as the harmonics in the 1 dimensional DCT.
The same principle applies to the coefficients down the left side for the
vertical direction, and likewise for the coefficients down the centre of the
block from top left to bottom right for the 45 degree diagonal direction.
However, with 64 original pixels there is now opportunity to provide
coefficients that describe how the original 64 pixels are varying at other
angles between vertical, 45 degrees and horizontal.
Sony Training Services 240

249. Broadcast Fundamentals
Thus the 8 by 8 group of DCT coefficients now describes a wide range
of frequency changes and angles. In fact the DCT block exactly
describes the original pixels, but in a different way.
The mathematics of DCT as used for video
The mathematical description of DCT so far has been 1 dimensional.
The expression we eventually concluded is :-
2 k = N −1
(2k + 1)rπ
F (r ) =
N
∑
k =0
f ( k ) cos
2N
241 Sony Broadcast & Professional Europe

250. Part 21 – Video compression
Now we need to replace the one dimensional DCT F(r) part with a two
dimensional part, F(u,v). The f(k) will be replaced by a two dimensional
f(x,y).
The cos part of the expression now needs to be done in two dimensions
as well, once for the x direction and another for the y direction to create
the u direction and v direction in the final DCT respectively. We therefore
have :-
2 x = N −1 y = M −1
(2 x + 1)uπ (2 y + 1)vπ
F (u , v) =
N M
∑ ∑
x =0 y =0
f ( x, y ) cos
2N
cos
2M
Where the original pixel block is N pixels wide by M pixels high.
This is not exactly correct. In practice we need a ‘fiddle factor’ for all the
coefficients across the top row or along the left column, i.e. when either
u=0 or v=0. This factor is √2. Thus the two dimensional DCT as used in
video is :-
2 x = N −1 y = M −1
(2 x + 1)uπ (2 y + 1)vπ
F (u , v) = Cu Cv
N M
∑ ∑x =0 y =0
f ( x, y ) cos
2N
cos
2M
Where :-
1
Cu = for u = 0 and C u = 1 for u = 1to N
2
1
Cv = for v = 0 and C v = 1 for v = 1to N
2
This expression looks pretty hideous, but you can see the vaious
elements in it, and where they come from. It is also worth bearing in
mind that it may look a lot simpler if you know more about the size of the
group of pixels you are looking at.
For instance, if the group is always square, we could say that the group
is N pixels by N pixels and eliminate the M and the annoying square
roots at the beginning, Thus :-
2 x = N −1 y = N −1
(2 x + 1)uπ (2 y + 1)vπ
F (u , v ) = C u C v
N
∑ ∑
x =0 y =0
f ( x, y ) cos
2N
cos
2N
If we also say that the pixel block is going to be 8 pixels by 8 pixels,
which it is for JPEG and all MPEG compression schemes, the
expression becomes even simpler, thus :-
1 x =7 y =7
(2 x + 1)uπ (2 y + 1)vπ
F (u , v) = Cu Cv ∑ ∑ f ( x, y) cos cos
4 x =0 y =0 16 16
Sony Training Services 242

251. Broadcast Fundamentals
DCT in audio
Right at the beginning we looked at how audio signals can be broken
down into a fundamental and a group of harmonics.
DCT is very appropriate to breaking down digital audio signals. Audio is
a one dimensional stream of data, and there is no need to use the kind
of two dimensional DCT we use for video signals.
MPEG uses DCT in compressing audio signal. The modern trend for
MP-3 players are based on MPEG, which in turn uses DCT.
Basis pictures
Machines often do not go through the tedium of calculating DCT values
from scratch. They use pre-calculated values in a kind of look-up table.
These look up tables are called basis pictures. There are as many basis
pictures as there are samples. And each basis picture contains as many
numbers as there are samples.
Thus for an 8 by 8 group of video pixels there are 4096 basis picture
numbers.
Machines use simple matrix multiplication to perform DCT using basis
pictures.
Why bother?
DCT is used to rearrange the pixels in a video picture into frequency
coefficients because more video pictures have their energy in the low
frequency and DC coefficients.
243 Sony Broadcast & Professional Europe

252. Part 21 – Video compression
Therefore if you do a DCT of a complete video frame you will invarably
find all the big numbers in the top left corner. The bottom right corner
tends to be full of zeros.
Looking at it another way DCT allows us to separate the video frame’s
entropy and redundancy, with all the entropy in the top left corner of
each DCT block and the redunancy in the bottom right corner.
This helps in the variable length coding part of a compression system to
reduce the number of bits required.
Huffman codes are a particular type of variable length code. Huffman
codes are particularly efficient at reducing the number of bits required to
describe data.
While the Huffman coding principle is almost certainly not the best
variable length coding system for every eventuality, and there may be
new coding systems in the future that are more efficient than Huffman
coding, it remains the most common variable length coding system for
data compression.
Huffman’s three step process
Huffman coding is performed in three steps. The first is to analyse the
probability of the original data. The second is to use this analysis to
generate a series of Huffman codes. The third is to use these codes to
reduce the original data.
Step 1 - Analysis
In the analysis step, a probability is assigned to each possible number in
the original data. Depending on the chance of it occurring.
Let us assume we want to reduce the text on the following page.
There are 7462 words including spaces, and the quantity of letters,
numbers and symbols is as follows :-
Letter Quantity Letter Quaniity Letter Quantity
A 557 Q 5 7 1
B 114 R 377 8 0
C 167 S 367 9 1
D 266 T 520 0 4
E 653 U 144 ( 11
F 150 V 54 ) 11
G 142 W 125 . 43
H 287 X 3 , 69
I 382 Y 124 ‘ 44
J 3 Z 3 ! 1
Sony Training Services 244

253. Broadcast Fundamentals
K 67 1 0 ? 2
L 275 2 2 @ 1
M 133 3 0 - 29
N 422 4 0 Space 1336
O 473 5 1 & 0
P 94 6 0 % 0
245 Sony Broadcast & Professional Europe

254. Part 21 – Video compression
F9or many who live in Argyll, occasional visits to Glasgow are a part of life, especially if you live in the nearer mainland parts of
Argyll as I do. Unless you've had to head there in its rush hour, the city is less than an hour and a half from Strachur by the 'Rest
and be Thankful' pass and Loch Lomond-side. (In passing, why Rest and be Thankful? The original military road is steepest of all
near its very top and my guess is that your horse did the resting and you did the being thankful - that the poor animal hadn't
collapsed with the effort and let your cart, carriage or whatever run back over the edge - but I'd be interested in other versions,
especially the correct one if different from this. Doubly welcome if with sources).
But back to Glasgow. Why do we go? is a question I've often asked myself when traipsing around wet streets and sticky shops. I
suppose there are almost as many answers as visits, but apart, obviously, from the much greater number and variety of shops than
Argyll (population 90 000 or so including all its towns) can offer, there are also all sorts of entertainments and cultural and
commercial reasons for making the effort. Despite too much continuing poverty and some areas of awful housing, it's also a very
attractive place that's had a long, hard time getting past the 'no mean city' image in the minds of outsiders, amongst whom I have
to count myself. To-day we were mainly going to take Ewan out for a meal in the evening, but shops and a film were to be added
in.
At least as good as the shops and films, as far as I was concerned, would be the chance to explore another corner of the city on
foot and, since we had Tess the dog with us, I had a good excuse. Despite forecast warnings of blizzards on northern hills, they
certainly didn't apply here and now.
I've sometimes heard Glasgow referred to, tongue-in-cheek, as the 'dear green place' and I'd be at least as keen to know the
derivation of this one as of the 'Rest and be Thankful'. I do know that you can find yourself in several places within the city
boundaries where that title just doesn't sound ridiculous at all and I was trying to find myself a new one of these to-day. I was
heading for the north-west of the city, a little beyond Anniesland and between the road to Bearsden and Maryhill Road, home of
Partick Thistle FC. In this little corner, with the aid of a partially-remembered scrutiny of the A to Z, I reckoned to put together a
fairly peaceful 2 mile triangle from the end of 'Switchback Road': one side being the towpath of the Forth and Clyde Canal, a
second the banks of the River Kelvin (a tiny part of the 'Kelvin Way') and the third a crossing of Dawsholm Park and so it turned
out (the fact that the canal had been recently drained on this stretch was a bit of a surprise - but it's too early in the year for the
mud to be smelly and, as it's being restored (hooray!), it'll be full again in time).
I started near Lock 27, where the new-looking canalside pub of the same name had towpath tables and outdoor drinkers to go with
them - not bad, pre-Easter. An AA roadsign nearby indicated a microbrewery, which might repay following up (memo to non-
Brits, AA can be 'Automobile Association', as well as Alcoholics Anonymous). Having hardly started, I put aside thoughts of a
pint of real ale and headed east for a couple of enormous gasometers that weren't yet industrial archaeology but, like the mud,
were smell free. Just as well.
A little further on, where a road crossed the canal, I came across the first sign of canal restoration work in the form of a brand new
bridge bearing the carved legend 'Forth and Clyde Canal' and a relief of what I think must be the giant new wheel arrangement
near Falkirk. When completed, this will lift boats from the Forth and Clyde to the Union Canal so that they can sail on to
Edinburgh for the first time in decades. A man in the contractor's hut nearby reckoned that the work in this drained Glasgow
section would be ready in a handful of months, but that the restoration of the whole canal would take a couple more years. Time
enough yet to book your canalboat holiday across central Scotland.
Continuing on, with only a few people and a pair of mute swans for company (having passed the temporary earth dam holding
back the water from the canal's dry section) and a mile, now, from my starting point, I came all of a sudden on one of the canal's
biggest engineering achievements - the aqueduct spanning the steep-sided den of the River Kelvin Beyond the aqueduct, a flight of
locks rose to cross Maryhill Road. Between the locks seven pairs of brand-new, heavy wooden lock gates lay stacked around a
small basin waiting to be installed. The flight was further than I wanted to go, but I paused briefly on the aqueduct to admire a pair
of cormorants sitting somewhat grandly on top of an abandoned sandstone pier that once carried a railway across the Kelvin.
Bereft of its railway track and all connection with either bank, the pier looked for all the world like a sea-stack lost in the middle
of this most urban of rivers and the cormorants, therefore, seemed oddly 'at home'.
'Most urban' isn't very fair. Descending to the far bank of the Kelvin took me down through some fine broadleaved woods to a
quiet riverside and turning to go upstream soon revealed as fine a patch of primroses as I've seen this spring. I passed no-one at all
by the river (5 ish on a Saturday afternoon), though beyond a road bridge there was soon more nearby housing. In fact it was all
green and pleasant to the next railway bridge (West Highland line) and green beyond, past low sandstone cliffs, to a low-level road
bridge where it's possible to cross back to the right bank again and search west for a way into Dawsholm Park.
Crossing this bridge brought great news in the form of a man fishing. Added to the cormorants' presumed need for food it was
becoming clear that the Kelvin may be urban, but certainly isn't dead. The fisherman said it was greatly improved, claiming trout
and sea-trout for it, and even suggested Partick Mill (downstream) as a place to go between September and November to watch the
salmon passing over the weir. I think I shall. I heard recently that, down in England, the River Tame, which rises in the
ominously-named 'Black Country' before flowing through Birmingham and which I remember, from more than thirty years ago, as
possibly the dirtiest and deadest in the whole of industrial Britain, now also sees anglers on its banks. Without wanting to be
Pollyanna-ish, not all environmental news is black.
Leaving the angler to his rod, and wondering idly about the less-contemplative youth suggested by the scars on each of his cheeks,
I headed through some scrub by guesswork, crossed a group of all-weather football pitches and found the gates of Dawsholm
Park, again by guesswork. By more guesswork I climbed a pine-wooded ridge to be all of a sudden re-oriented by the middle-
distant gasworks and then by a grand view along the length of the industrial Clyde through Glasgow and out to Clydebank with its
shipyard cranes. Sure of my position again, a balcony of a path took me along a drumlin-crest above a pasture containing four very
hairy Highland Cattle (Glasgow is full of surprises) to a place where I could slip down easily to my car. Too late to seek out the
microbrewery, but another day.
It really is a dear green place if you look.
John Fisher
aboutargyll@compuserve.com
Sony Training Services 246

255. Broadcast Fundamentals
Now the letters are rearranged in order of the number of times they
occur in the text. The actual quantity is expressed as a probability with
regard to the total number of letters, 7462 :-
Letter Prob. Letter Prob. Letter Prob.
1 0 Z 0.000402 G 0.01903
3 0 0 0.000536 U 0.01930
4 0 Q 0.000670 F 0.02010
6 0 ( 0.001474 C 0.02238
8 0 ) 0.001474 D 0.03565
& 0 - 0.003886 L 0.03685
% 0 . 0.005762 H 0.03846
5 0.000134 ‘ 0.005896 S 0.04918
7 0.000134 V 0.007237 R 0.05052
9 0.000134 K 0.008979 I 0.05119
! 0.000134 , 0.009247 N 0.05655
@ 0.000134 P 0.012597 O 0.06339
2 0.000268 B 0.015277 T 0.06969
? 0.000268 Y 0.016617 A 0.07464
J 0.000402 W 0.016752 E 0.08751
X 0.000402 M 0.017824 Space 0.17904
Now the two lowest probabilities are summed together and the letters
are lumped together.
A new table is made in the same way with, and the two lowest
probabilities are summed with the two letters expressed together.
This process is repeated until there are just two letters left.
247 Sony Broadcast & Professional Europe

256. Part 21 – Video compression
Variable length codes are used predominantly in digital compression
schemes, and is the most effective method of reducing the amount of
data.
Entropy codes is another term used for variable length codes. Hamming
codes are a particularly efficient type of variable length, or entropy,
codes.
The principle behind variable length codes
The idea behind variable length codes is to know which number are
more likely to occur that others in your digital data, and replace these
with a special code that is smaller that the original number.
Numbers that are likely to occur a little less often are changed for slightly
larger codes. Numbers that and very unlikely to occur are replaced with
codes larger than the original number.
The hope is that there is a stronger chance that likely numbers will occur
in the data than unlikely numbers and there will be a lot more small
codes than big codes.
The results of discrete cosine transforms
DCT (discrete cosine transforms) replace original video pixel data with
numbers corresponding to how the data is changing. These numbers are
called coefficients.
DCT operated on a matrix of pixels. MPEG uses a matrix of 8 by 8
pixels. Other matrix sizes are used but 8 by 8 pixel matrices are the
most popular.
Conventionally the DC coefficient is placed in the top left corner of each
matrix, replacing the pixel data that original occupied this position. This
DC coefficient has a flat statistical probability curve. This means that is
just as likely that that it will contain any number.
AC coefficient bell curves
The rest of the coefficients are AC coefficients. The AC coefficients all
have bell shaped statistical probability curves. That is to say that there is
a greater chance that the number is somewhere in the middle.
The high frequency coefficients have a sharper bell curve than the low
frequency coefficients. That is to say there is a stronger chance that the
high frequency coefficients will contain a number closer to the middle
that with the low frequency AC coefficients.
Using bell curves for variable length coding
It is very useful that all the DCT AC coefficients have a bell curve
probability. It is possible to design variable length codes that takes
advantage of this bell curve by having small codes for numbers at the
peak of the curve and large codes for numbers at the outer edges of the
curve.
Sony Training Services 248

257. Broadcast Fundamentals
Numbering systems for the bell curves
The original video samples are all 10 bit values and have a range from 0
to 1023. The DCT AC coefficients use a negative numbering system and
therefore have a range from –512 to +511.
The peak of the bell curve is therefore centred about zero. In terms of
the original video the DCT AC coefficients are centred around the colour
grey.
Using a simple variable length coding system
Imagine a simple variable length coding system based on the bell curves
for the DCT AC coefficients.
DCT AC coefficient Variable length code
-9 11111111101
-8 1111111101
-7 111111101
-6 11111101
-5 1111101
-4 111101
-3 11101
-2 1101
-1 101
0 0
+1 100
+2 1100
+3 11100
+4 111100
+5 1111100
+6 11111100
As you can see the code for a zero is just one bit, a “0”. It is also easy to
see the pattern for negative and positive numbers. The number itself
indicates the number of “1”s with negative numbers ending in “01” and
positive numbers ending in “00”.
Let us consider a stream of 14 DCT AC coefficients thus :-
+1 –2 0 +4 –3 0 0 +6 –1 –3 +4 0 –1 +2
Converted into their simple variable length codes gives us :-
100110101111001110100111111001011110111110001011100
The original DCT AC coefficients were all 10 bit samples. Thus in this
stream there are 140 bits. The variable length codes for this stream have
51 bits.
249 Sony Broadcast & Professional Europe

258. Part 21 – Video compression
It is important to realise that the variable length codes are exactly the
same data as the original coefficients, they are just described in a more
efficient way.
Decoding variable length codes
If it is a little difficult to believe how variable length codes can be such an
efficient bit saver, decoding the original data from the apparent random
stream of variable length codes will seem a miracle.
However decoding variable length codes, while fiendishly clever, is
actually very simple.
Starting from the beginning you look to see if there is a code
corresponding to the first bit. If there is not then look to see if there is a
code for the first and second bit together. If not try the first, second and
third bits together.
In this case the first bit is a “1” which is not a code, The first and second
bits “10” is also not a code. However the first, second and third bits give
a good code “100”. The original coefficient for this code is “+1”.
Now discard these bits and start again, looking for the first time you see
a good code. The next good code is “1101”. Anything less than this is
not a good code. “1101” is the code for “-2”.
Cut this off and start again. The next bit is a “0”. This is a good code in
itself and gives a “0” as the coefficient data.
This same method can be used all the way through the variable code
stream replacing the good codes as you find them with the original
coefficients.
Disadvantages of variable length codes
Variable length codes can suffer from errors in the same way as any
other signal, both analogue and digital. Variable length codes can also
suffer when they are read from the middle, rather than from the
beginning.
However where variable length codes fail is when the original data does
not fall into the assumed statistical pattern, and when you want to
decode it starting in the middle.
Errors in variable length codes
Let us imagine that there is and error in the variable length codes thus :-
100110101110001110100111111001011110111110001011100
The eighth “1” has been received as a “0”. Decoding this data starts OK
but goes wrong when we reach the mistake thus :-
+1 –2 0 +3 0 –3 0 0 +6 –1 –3 +4 0 –1 +2
Instead of “+4” we now have “+3” “0”. A mistake in the variable length
codes can ripple down the data giving rise to a few incorrect coefficients.
Sony Training Services 250

259. Broadcast Fundamentals
In practice the variable length codes and remarkably resilient and errors
dot not ‘travel’ too far down the variable length stream and only a few
coefficients are affected. Error correction schemes can be used to
correct these minor mistakes and replace the original coefficients again.
Decoding variable length codes from the middle
The second disadvantage of variable length codes is that you will get an
error if the variable length code stream is decoded from the middle.
To illustrate this assume that we start reading the original variable length
stream we made starting at the tenth bit.
Discard these 10 bits 100110101 and start reading here…
111001110100111111001011110111110001011100
The result of this is :-
+3 –3 0 0 +6 –1 –3 +4 0 –1 +2
So, even though the first coefficient is in error, the rest of the stream has
been decoded correctly.
Bad statistical patterns
Variable length codes can stand a few numbers that do not follow
statistical assumptions.
Assume that there are a few large numbers, both negative and positive,
in the stream mentioned above, thus :-
+1 –2 0 -50 –3 0 0 +6 –10 –3 +4 0 –1 +2
One of the coefficients has been replaced by –50 and another by –10.
The new stream is now thus :-
100110101111111111111111111111111111111111111111111111111
1011110100111111001111111111011110111110001011100
While this appears to be a massive increase in the number of bits we
had before it is still only 106 bits and still a saving on the 140 bit we had
originally.
However, what if the original stream were replaced by something like
this :-
+23 –12 +56 +1 +8 –3 -112 –31 +90 –121 +53 +27 +78 –94
This is still 14 coefficients. Each one is 10 bits, giving 140 bits
altogether. However the variable length stream corresponding to these
coefficients is :-
111111111111111111111110011111111111101111111111111111111
111111111111111111111111111111111111110010011111111001110
111111111111111111111111111111111111111111111111111111111
111111111111111111111111111111111111111111111111111111110
111111111111111111111111111111110111111111111111111111111
111111111111111111111111111111111111111111111111111111111
111111111100111111111111111111111111111111111111111111111
111111111111111111111111111111111111111111111111111111111
111111111111111111101111111111111111111111111111111111111
251 Sony Broadcast & Professional Europe

260. Part 21 – Video compression
111111111111111110011111111111111111111111111100111111111
111111111111111111111111111111111111111111111111111111111
111111111111001111111111111111111111111111111111111111111
11111111111111111111111111111111111111111111111111100
This is 737 bits! A huge increase on the original 140 bits. Thus it is very
important that variable length codes are as closely related to the
statistical pattern of the original data. This is the reason why variable
length codes cannot be used on the DCT DC coefficients, because there
is no guarantee that it will be close to zero at all.
Sony Training Services 252

261. Broadcast Fundamentals
The television station
The studio
The studio is the area where programmes are made. It may be small,
being not must more and a small room, to a large enclosure big enough
to fit a small group of houses.
Studios have sets and lighting. A set is a construction that provides a
background, or foreground, to the action taking place in the studio.
Lighting can be hung from the ceiling or placed on the floor. It not only
provides light to the set, but also adds mood and colour. Both set design
and lighting design require experience and skill to perfect.
Studios also contain video cameras. There are often fewer cameras in
news studios that those of drama, pop music shows, etc. However,
cameras in news studios can be very complex computer controlled
cameras, able to move between a few certain but very specific positions.
Studios can be used for news programmes, entertainment programmes
like pop music shows, chat shows and games shows, and for drama.
Drama, which includes everything from soaps to high profile period
drama productions, probably involve the most complex set and lighting
design in any studio. Indeed a soap will involve continuous use of a very
complex set design that operates like a well oiled machine.
The gallery
Studios often have an associated gallery. A gallery is a control room
generally placed next to the studio. It is also often set at a higher level to
the studio, so that gallery staff can look down on the studio.
Studios with galleries are generally used for live, or near live,
programme making. The gallery crew can direct the programme making
process to either create a programme that goes directly out to air, or a
programme recording that requires very little subsequent editing.
The post production studio
The post production studio is not really a studio at all. In fact it is more
like a studio gallery. Post production studios do not have sets or lighting.
They are designed specifically to take the recorded material from a
studio and edit it into a final programme (or series of programmes).
Post production studios are not as hectic and busy as galleries. The
work that is done in postproduction is not done to the clock, with all the
time critical nature of the work done in the live studio and gallery.
Post production may take many days or weeks for a single programme.
Although the equipment used in post production is similar to that used in
the gallery, there are a few obvious differences. For instance, special
effects machines in the gallery tend to be expensive pieces of hardware.
They can produce a few effects at high speed. In post production the
253 Sony Broadcast & Professional Europe

262. Part 22 – The television station
special effects machine will produce some very complex effects, and not
necessarily in real time.
The edit suite
The edit suite is really another name for the post production studio,
although there is probably the expectation that the post production
studio is more complex that the simple edit suite.
The linear edit suite
The linear edit suite generally used tape. All edits involve playing back
one or more tapes and recording to another tape. It is called linear
because edit have to be performed in a linear fashion, as they are
recorded to tape.
The non-linear edit suite
Non-linear edit suites are a recent invention compared to the linear edit
suite. Non-linear edit suites used to have a reputation for lower quality
results compared to the linear edit suite. They generally use computer
technology to allow ‘drag & drop’ timeline style editing.
Non-linear edit suites offer a very much more flexible way of working
than linear edit suites.
Non-linear editing is slowly increased in quality as the power of
computers has increased. It is now possible to perform non-linear edits
with the same high quality as conventional linear edits.
The news studio
The news studio is not really a single room, although there will be a
studio in the basic sense. A news studio will contain maybe three rooms.
There will be a studio proper, where the news programme will be made.
There will be an associated gallery as well.
The third room will be the news room itself. An integral part of the studio
complex, the news room is used to bring all the news items together,
with their associated script, video or film footage, and audio material.
The news room is a little like the opposite to a post production suite. The
work done in it is before the programme. There is some editing, although
it is not as complex as the editing equipment found in post production.
News editing is always simpler, and often done to very short and rigid
deadlines.
The outside broadcast vehicle
The outside broadcast vehicle, often called the outside broadcast truck,
or OB truck, is a small self contained transportable production facility. It
contains all the equipment to shoot, record, control, and edit a
programme. And when finished, transmit the result back to the television
station.
Sony Training Services 254

263. Broadcast Fundamentals
OB trucks are used for recording sports events, national celebration
events, important news events, etc. or any situation requiring production
facilites, where there are none normally.
OB truck are sometimes huge with separate rooms for video editing and
production, audio and camera control. They can just as easily be small,
with one room inside for everything.
OB trucks are often measured by the number of cameras they have. A 1
camera OB truck is small, an 8 camera OB truck is large.
255 Sony Broadcast & Professional Europe

264. Part 23– CCTV, security & surveillance
Part 20 CCTV, security & surveillance
What is CCTV?
CCTV stands for closed circuit television. It encompasses any television
systems that is not connected to any kind of transmitter, and is generally
a single connection between the source and destination. There would
normally be just one or two television screens at the destination,
CCTV also includes the use of microwave or infra-red links, video over
IP, and other such technologies. With these technologies the signal is
directed to a specific destination, rather than being broadcast to anyone.
As such, it is still closed and can therefore be regarded as CCTV.
CCTV does not include terrestrial broadcast television, satellite or cable
network television. It does not even include subscriber television. Such
services are only available to the customers who buy the subscription
and who have the relevant technology to receive the signal. Although
they could be regarded as ‘closed’ they do not fall into the description
‘CCTV’ because the number of receivers is relatively high.
CCTV privacy & evidence
CCTV has gained a reputation in many peoples minds as the instrument
of “big brother”, and a method for the authorities to pry on innocent
individuals. Sensible use of CCTV should never be an infringement of
privacy. In many cases CCTV is used for situations where personal
privacy is not an issue. Instrument monitoring, remote monitoring in
harzardous conditions, search and rescue, and many other applications
for CCTV do not involve personal privacy at all.
Data Protection Act 1998
Laws around the world are designed to protect people from incorrect use
of CCTV technology. In Great Britain the Data Protection Act 1998
includes 62 legally enforceable points to ensure correct use of CCTV
technology, and 30 suggested good practice points to improve public
perception of the technology. Details of the Data Protection Act 1998
can be found at http://www.dataprotection.gov.uk .
Continuity of evidence
CCTV equipment can be invaluable at collecting evidence as part of
legal proceedings. Recordings and images can all help build up a
convincing case. However evidence is useless if it has been tampered
with.
It is important that any images, video or sound material are not placed in
a position where they can easily be altered, or deleted, between the
point where they were recorded and where they are presented in court.
There may also be a necessity to guarantee that material is not
tampered with after it is presented in court, right up to the time it is
destroyed.
Sony Training Services 256

265. Broadcast Fundamentals
The process of ensuring that CCTV material is not altered between
recording, through its court appearance and its eventual destruction is
called continuity of evidence.
It is impossible to absolutely guarantee continuity of evidence. All that
can be done is to reduce the likelihood of tampering to such a level that
it is improbable.
Two methodologies can be used to ensure continuity of evidence,
trusted personnel and technology. Using both methods can provide very
convincing of evidence.
Trusted personnel
Recorded evidence can be placed in the hands of trusted personnel.
These people are trusted to prevent the material from being tampered
with either because it is their job or because their reputation depends on
it.
Trusted personnel include security guards, bonded store keepers etc.,
as well as notaries, like judges, doctors, police etc.. All trusted personnel
may be corrupted, but it is unlikely. It is this unlikelihood that provides
continuity of evidence.
Technology
Recorded CCTV material can be protected at every point from the
camera through the court room and eventually to destruction by using
technology.
Simple technology includes lock & key, safes, security doors etc.. These
are the kind of technologies that security guards and bonded store
keepers would use to back up their trusted personnel status.
CCTV material can also employ signal scrambling techniques and
electronic watermarking to make tampering difficult.
Making multiple copies of recordings, at separate remote sites can
improve security, and prove tampering.
CCTV use
CCTV’s primary use is in security & surveillance, and has gained a ‘big
brother’ reputation, with all the concerns of civil liberties and privacy.
However CCTV is important is increasing levels of safety in public areas,
offering better levels of monitoring and control for inspection, machine
operation, medical applications, and for work in remote or hazardous
environments.
Examples of CCTV usage
Town centre surveillance
Local councils and police are using CCTV increasingly to monitor city
and town centres. Cameras are mounted on buildings or posts at
strategic points. Many are fitted inside motorised environmental
257 Sony Broadcast & Professional Europe

266. Part 23– CCTV, security & surveillance
housings. The signals from these cameras are fed back to a central
control office, where staff can monitor activity in shopping areas, public
amenities, car parks, etc.. Although there are concerns about public
liberties, these systems have been very successful in reducing
robberies, vandalism and mugging in town centres.
On-the-spot views for sports events.
CCTV, and especially the use of miniature camera technology, is being
used increasingly to allow people to experience the thrill of sports events
by mounting cameras to racing cars, motorcycles and jockeys. Signals
can be fed back to the production studio where they can be fed into the
broadcast chain.
As well as allowing people at home to see what the racing driver or rider
is seeing while racing down the track, it also provides valuable
information to pit crews and officials.
Train crew assistance systems
Train platforms can be very long and often curved. It is often difficult for
drivers and guards to see the whole length of a platform as the train is
pulling out. CCTV is often used as a means of checking that doors are
closed and passengers are clear of the edge of the platform before
starting the train.
Cameras are mounted to the wall or ceiling at strategic points along the
length of the platform fed to two or three small monitors mounted just
outside the train window, so that the driver or guard can easily see them
without having to turn or stretch.
Biometric identification
At the cutting edge of CCTV is personal identification using biometrics,
Whilst not generally regarded as CCTV by many people, biometrics uses
the same basic systems as all other CCTV systems. Entry systems to
high security areas can involve specialist cameras linked to digitizers
and computers. These can be used to perform facial scans, fingerprint
scans or retinal scans to help identify people.
Search and rescue
CCTV is used extensively when searching for survivors in fires,
collapsed building, caves and pot-holes. Miniature cameras can be
pushed into places humans cannot get into. Small cameras can be fitted
to robot crawlers and sent into dangerous environments that are too
dangerous for humans.
Another important area or search and rescue takes advantage of the fact
that cameras can see part of the electromagnetic spectrum humans
cannot see, i.e. infra-red. Helicopters can search for people or animals in
open country, or sea, from their heat signature. Special cameras that are
sensitive to heat can make people or animals shine out like beacons
even at night.
Sony Training Services 258

267. Broadcast Fundamentals
Medical procedure monitoring
CCTV is now being used to monitor the progress of medical operations
and procedures. From endoscopes to remote control microsurgical
equipment. CCTV equipment is becoming increasingly important.
Linking CCTV equipment to shape recognition and 3D modelling
software, and systems can be built to assist medical teams in diagnosis.
Another important area of development is in remote consultation. By
using CCTV with video over IP, or streaming technology, it is possible
for top surgical consultant to assist difficult surgical procedures
anywhere in the world.
Microchip production and inspection
Many modern production processes operate at dimensions far too small
for humans to see with the naked eye. As dimensions become smaller
and smaller, it is also impossible see things using normal light. The
wavelength of light itself becomes a problem and other forms of radiation
must be used. CCTV sensitive to these wavelengths are used to allow
people to see these tiny dimensions.
Probably the most important of these is microchip production. CCTV is
used extensively in the production process, by coupling it to microscope
technology using ultra-violet and X-ray radiation.
CCTV terminology
CCTV technology has parallels with professional and broadcast video.
However some of the terms and jargon used in the CCTV industry is
entirely different to those used in broadcast video
Activity detection
The ability of a system to react to movement. A processing unit will
compare video frames to check for differences, i.e. movement. This can
be sent out as a signal to the systems control to switch to that camera.
Multiplexers can be adjusted to devote more, or all time to the camera
that has sensed movement. It is also possible to send zoom pan and tilt
control signals to the camera to make it zoom in closer on the detected
movement.
Alphanumeric video generator (AVG)
The CCTV equivalent to a character generator. Quality is generally not
as good as broadcast character generator, but the requirements for
character insertion in CCTV are not as stringent, and cost saving is an
issue.
CCD iris
A term used in CCTV to describe auto iris.
259 Sony Broadcast & Professional Europe

268. Part 23– CCTV, security & surveillance
C mount
Cine mount. The original lens mounting system for CCTV cameras. 1”
(25.4mm) diameter with 32 threads/inch. Back flange to CCD distance
standardised at 0.69” (17.526mm).
Conditional refresh
A system used to save transmission bandwidth by transmitting video
frames only when a change is detected.
CS mount
Cine short mount. An more contemporary lens mount to the C mount
standard offering cheaper smaller lenses. CS mount has exactly the
same dimensions as C mount, but the back flange to CCD distance is
reduced to (12.5mm)
Back light compensation (BCL)
A camera feature that automatically compensates for strong background
lighting to improve detail in darker areas of the image that would appear
as just a black shape.
Dense waveband division multiplexing (DWDM)
A technology that places a large number of video channels onto a fibre
optic cable.
Dome camera
Any CCTV camera installed in a dome. Modern dome cameras a pre
built as a complete assembly, often with pan, tilt and zoom capability,
and are sometimes referred to a PTZ cameras. Some dome cameras
have a network output rather than a conventional video output, so that
the video signal can be sent out on a computer network as a
compressed data stream.
Duplex
Used in CCTV to describe a multiplexer that can perform more than one
function at one time, like displaying and recording multiple images.
Dwell time
The time a multiplexer stays on one camera in its rotation.
Hi-Z
A common term in the CCTV industry to denote an unterminated
analogue video cable connection. The line impedance of video cable is
75ohm. Many coaxial analogue video connectors have a switch that can
be switched between Hi-Z and Lo-Z or 75ohm.
Equipment can be daisy chained on the same video connection. All
equipment except for the first and last are set to Hi-Z (high impedance –
Sony Training Services 260

269. Broadcast Fundamentals
terminator switched off) The first and last pieces of equipment are set to
75ohms impedance, either by switching the equipment’s terminator
switch to 75ohms or by fitting a BNC terminator to the cable end.
Kangaroo lens
A lens with two fixed iris positions – fully open and partially closed.
Designed to be used with cameras that have electronic (sensor) auto-iris
capability as a way of providing 2 ‘gears’ for the auto-iris.
Lambert radiator
A primary source of light that is designed to be imperfectly diffused.
Lambert reflector
Like a Lambert radiator but for a secondary (reflected) light source.
Minimum object distance (MOD)
The closest distance a particular CCTV lens can focus to, measured
from the front of the lens to the object.
Multiplexer
A unit that multiplexes a number of video signals.
A multiplexer can be designed to show more than one video image on
one screen by down converting the incoming video images and placing
these smaller images into one outgoing signal. Can be refered to as
spacial multiplexing. Each image has poorer resolution that the original
but is running at real time. Probably the most popular of these is the
Quad.
A time division multiplexer divides each video frame, or series of frames,
between a number of inputs. The output effectively chops between the
inputs on a rotational basis. Each video input has the same resolution
but is not as smooth. The output of a time division multiplexer is not easy
to view because the image flickers from one image to another. Time
division multiplexers are generally used as a way of recording multiple
video signals to one tape. Timecode is also recorded and is used by the
multiplexer during playback to demultiplex the recorded signal.
Pan & tilt head (P/T head)
A motorised camera mount that allows the camera to be panned (move
round) and tilted (moved up and down) remotely. Pan & tilt heads are
often combined with a zoom camera to give a pan tilt zoom assembly
commonly called a PTZ camera. Many dome cameras are also PTZ
cameras
261 Sony Broadcast & Professional Europe

270. Part 23– CCTV, security & surveillance
Pre-position lens
A CCTV lens which outputs signals for zoom and focus positions, so that
hey can be stored in the control station and thus allow pre-set positions
to be called up by the controller quickly.
PTZ camera
See Pan & tilt head.
Quad
A unit that spacially multiplexes four video signals into one signal; to
show all four images on one monitor. The resolution of each image is a
quarter of the original but is running at real time.
Repeater
A unit that can be placed part way along a very long transmission path to
amplify the signal back to full amplitude again. Cable repeaters can be
used to re-amplify video and audio signals. Microwave repeaters do the
same thing for video and audio signals on microwave links.
Analogue repeaters also amplify noise, so there is a limit to the number
of analogue repeaters that can be used before the noise level becomes
so great that it swamps the original video or audio signal
Retained image
A term used in CCTV to describe an image that remains on the camera
sensor after the object has gone. Retained image is a temporary artifact
as a result of a delay in the camera sensor, but is sometimes used to
describe the burnt-in image because a CCTV camera is looking at the
same thing all the time.
Vari-focus lens
A manual zoom lens. All other industries regard all lenses with a variable
focal length as zoom lenses. The CCTV industry need to be able to
differentiate manual zoom lenses from motorised ones, because many
CCTV cameras are operated remotely. The focal length of a vari-focus
lens can be altered during installation, then must be considered set from
the point of view of the operator in the control room. (See zoom lens.)
Z
Zoom lens
A lens with a remotely controlled motorised variable focal length. The
CCTV industry differentiates between manual and motorised zoom
lenses. The manual variety are called vari-focus lenses. (See vari-focus
lens.)
Sony Training Services 262

271. Broadcast Fundamentals
The typical CCTV chain
A typical CCTV chain consists of a number of devices communicating
with one another. Video and audio signals go in one direction, from the
camera or microphone to the monitor or speaker. Control signals go in
the opposite direction, from the control point to the camera or
microphone.
The camera
All CCTV systems start with the camera. This is the main input to the
whole system. There may be just one, or many cameras. Cameras may
be monochrome or colour, and can vary in size and quality. Some
CCTV cameras are fitted into environmental housings to protect them
from weather or harzardous conditions. Some are hidden behind
screens or domes to make them more discrete. Many cameras may
have controls for zoom and iris, as well as motors for pan and tilt.
The microphone
Many CCTV systems are video only systems. Some have associated
audio. In many cases microphones are fitted to the cameras themselves.
Increasingly microphones are fitted separately from the camera. This
allows it to be strategically sited to pick up the best sound.
The transmission chain
The transmission chain relays video and sound back to the control or
processing station. In most CCTV systems the transmission system
consists of simple video cables routed directly from each camera to the
control station, and audio cables routed directly from the microphones to
the control station.
The transmission chain may consist of switch gear, routers, signal
compressors and decompressors, IP packetisers, microwave or infra-red
links.
The control station
The control station consists of a control panel to either switch between
incoming video and audio signals directly, or to send control signal to a
remote switcher or router to perform the switching elsewhere.
Video signals are switched between cameras and recorders and the
monitors, and between the microphones and monitors and the speakers
and headphones
The control station can also send signals to cameras to perform zoom,
pan and tilt movements, as well as iris control and filters, depending on
the camera’s capability. Outside cameras may have wipers to clear
water and dust from the camera housing window and heaters inside the
housing to heat the camera if the temperature falls below freezing.
These could be automatic or controlled manually from the control
station.
263 Sony Broadcast & Professional Europe

272. Part 23– CCTV, security & surveillance
The processing station
In some CCTV systems the control station is replaced by a processing
station. This is popular in automated CCTV systems for applications like
automatic recognition systems. The images from the cameras are not
viewed by an operator but are processed by computer and either placed
in a database of compared to a database.
The speakers & headphones
Audio signals are routed from the control station’s panel into either
speakers or headphone, so that the operator can hear the sound. In
most cases this will be a mono, rather than a stereo feed.
The video recorder
Many CCTV systems have recording capability. This allows the
controller to review past events, and acts as a good source of evidence.
Video recorders can be simple VHS machines or more expensive digital
machine like those based on the DV tape format.
If the CCTV system has audio capability this is recorded on the same
tape as the video signal.
In many cases the video recorder records one camera feed. In some
cases the video recorder is able to record more that one camera feed on
the same tape.
The monitor
The monitor is the final destination for the video signals. They can be fed
from any of the cameras or from one of the video recorders.
Sony Training Services 264

273. 265
Figure 111
C a m e ra & M o n ito r s & s p e a k e r s
m ic r o p h o n e
M o to r is e d R o u tin g m a tr ix
c a m e r a in
h o u s in g &
Broadcast Fundamentals
m ic r o p h o n e
S w itc h e r
M o to r is e d
c a m e r a in
h o u s in g
M ic r o w a v e lin k
C o n tr o l s ta tio n
C a m e ra &
m ic r o p h o n e
In te r n e t
M in i c a m e r a
V id e o r e c o r d e r s
C o m p re s s o r D e c o m p re s s o r
/p a c k e tis e r /d e p a c k e tis e r
Typical CCTV system
Sony Broadcast & Professional Europe

274. Part 23– CCTV, security & surveillance
CCTV cameras
CCTV cameras are essentially the same as those used for broadcast,
although there tends to be less emphasis on quality.
General purpose cameras
Most CCTV cameras are designed for general indoor and outdoor use.
They vary in size from about 10x3x3cm to 30x10x10cm. General
purpose cameras normally sell without a lens, and one needs to be
bought and fitted. Both the older C and newer CS lens mounts are
popular.
General
purpose
cameras
normally
have an
analogue composite output. Some have an analogue component, and a
few have digital video outputs.
General purpose cameras can be fitted into weather housings. This
housing can be fitted with a heater to ensure it still operate in cold
weather, and a wiper to ensure the housing window is kept clear of rain
drops and dust.
General purpose cameras can also be fitted to pan and tile mechanisms.
This allows then to be repositioned remotely.
The control for these cameras is through separate RS-232, RS-422 and
RS-485 connection with proprietary control protocols.
Net cameras and web cameras
This is a relatively new type of camera. It incorporates all the
functionality of a standard camera but with a single network connection.
This connection can be used for both the video signal from the camera
and the control signals to the camera. Software is installed on a
computer which allows
the camera to be
controlled.
Network connected
cameras are easy to
install. Images can be
sent to any location on
Sony Training Services 266

275. Broadcast Fundamentals
a company network, or out over the Internet. Control can, likewise, be
sent from anywhere on the network, or the Internet, with the appropriate
software.
Net cameras use some form of image compression to reduce the
amount of data sent on the network link, while keeping image quality as
high as possible. JPEG is a common compression format. Some use the
more complex MPEG compression to improve the compression/quality
ratio.
Some net and web cameras have pan & tilt as well as zoom capability
built in. small motors in the camera assembly allow for this kind of
control, which is achieved by sending control signals over the network
link, from a controlling computer.
Dome cameras
This type of CCTV
camera is becoming
more popular. They are
not discrete, as most
people recognise these
domes for what they
are. Older dome
cameras were really
housings for a number
of cameras, with each
one pointing in a
different direction.
Modern dome cameras
now have pan and tilt
built in. Many also have zoom control as well. These are commonly
called PTZ cameras.
Domes are generally made from some kind of high impact plastic, to
give some protection from vandalism. They are generally tinted. This
hides the actual orientation of the camera inside, but reduces the
sensitivity of the camera.
Many dome cameras have standard analogue composite video outputs.
The control for pan and tilt for these cameras is through proprietary
RS-232, RS-422 or RS-485 connection, just like their general purpose
camera equivalent.
An increasing amount of dome cameras have network connections.
These cameras offer the same advantages as network and web
cameras, but in a protective dome.
Night vision cameras and dual condition cameras
Night vision cameras are
designed to operate after
dark. Two methods are
267 Sony Broadcast & Professional Europe

276. Part 23– CCTV, security & surveillance
used. The first is through intensification, and the second through the use
of infra-red.
Intensification cameras cannot work in complete darkness. They use
techniques to intensify the sensor, allowing then to pick up objects with
just the slightest amount of light. Images tend to be lower quality than
normal because the noise is intensified as well.
Infra red cameras have extended sensitivity beyond the normal electro-
magnetic spectrum into the infra-red. They can pick up heat, and build
up an image from a combination of what little light there is and the
temperature of the objects in the scene. Infra-red night vision CCTV
cameras sometimes have infra-red lamps mounted in the same
construction as the camera itself. This provides a good picture through
the camera, while remaining completely dark to the human eye.
Night vision cameras are all monochrome because they are looking for
basic form and shape with whatever light or heat is available. Their
colourimetry is wrong.
Dual condition cameras are able to switch between normal daylight
colour operation and monochrome low light operation. This can either be
achieved by switching on and off the sensitivity to infra-red, or switching
on and off intensification.
Wireless CCTV cameras
This kind
A e r ia l
of camera
C a m e ra has no
cable
Pow er
A e r ia l R e c e iv e r
Pow er
V id e o
o u tp u t
connection, other than power. It has an in-built wireless transmitter
operating in the GHz range of frequencies, and can transmit short
distances to a receiving station. Wireless CCTV are easy to install, and
reposition, and are ideal for temporary installations, and installations
were cameras may need to be moved on occasions.
However wireless cameras can be easy to tap into. All you need is the
same receiver on the same frequency.
Sony Training Services 268

277. Broadcast Fundamentals
Pin-hole & bullet cameras
Pin hole cameras used to refer to basic cameras consisting of a box with
a pin hole in the front. The name has been ‘stolen’ and now also refers
to a classification of sub-minature camera with a very small lens at the
front.
These types of CCTV camera remain at the fringe of main stream
CCTV, with all the concerns of privacy, spying, and discrete surveillance.
The largest of these is the bullet camera, sometimes called the lipstick
camera. The processing electronics is designed into a separate unit, and
the camera
head is
nothing more
than the lens
and the
sensor. This
makes it as
small as
possible, like a
bullet or
lipstick tube
(hence the
name). The
separate unit
is often called
the camera
control unit (CCU), although, in truth, it contains as much of the
electronics that can be removed from the camera head itself.
These cameras offer a good compromise between reasonable quality
and a discrete camera.
The smallest cameras are all single unit, single CCD or CMOS sensor
type. The lens, sensor and electronics are all integrated onto the same
small circuit board. There is no electronic control, and very little iris or
focus control. What there is, is always manual.
This type of CCTV camera is popular for fitting into clock faces, wall
mounted electrical sockets, light fittings, etc.. Image quality tends to
suffer because of the restricted space for electronics and the small lens.
True pin-hole CCTV cameras, sometimes called SWAT cameras, have a
thin shaft or flexible tube, protruding from the front of the camera head
with a very small lens mounted at the front. The shaft itself is a light
guide. These cameras provide for the smallest identifiable intrusion in to
a room space and are the most discrete of all CCTV cameras.
A standard camera can be fitted with a pin-hole lens. This lens is often
long and thin, and comes to a point. The whole camera can be fitted
behind a wall with the only visible part being the tiny front element of the
lens.
Pin-hole cameras tend to be wide angle and the image is often greatly
distorted.
269 Sony Broadcast & Professional Europe

278. Part 23– CCTV, security & surveillance
Biometrics cameras
Biometric cameras are
specifically designed to scan
specific human physical
features. This includes hand
geometry, face, iris, & retina.
Biometrics can also be used
for signature recognition.
Biometric camera outputs can
either have conventional video
composite or component
outputs or direct computer
connections. These outputs
are connected directly into a
computer.
Video connections need to be fed into a plug in board with an
appropriate video input. The board performs the image capture. The
computer then analyses the image.
Computer connections include RS-232 and RS-422 connections, USB
and 10-BaseT connections. This type of connection is becoming more
popular because it is easier to fit. The image capture is performed by the
camera into an internal frame store. The computer output is a digital
download of the frame store.
Some biometric cameras can now perform some of the image analysis,
internally in dedicated hardware, searching the image for relevant
information, discarding the rest, and even breaking the relevant data
down into a digital code. This greatly reduces the computer’s workload,
and speeds up transmission by only sending relevant data to the
computer rather than the whole image.
Once in the computer the image, or code, is compared to a database of
known patterns to recognise the person.
Instrumentation cameras
This is a loose definition of CCTV camera. Instrumentation cameras are
similar to other CCTV camera types in many respects, and cameras
designed for other purposes can be used as instrumentation cameras.
Instrumentation cameras are specifically designed to be fitted to
precision machines and instruments. They are intended for monitoring of
inaccessible areas, or looking on very low light level areas or in special
lighting conditions, like microscopes.
Many instrumentation cameras use the same design techniques as
some pin-hole cameras. Some have small camera heads linked via light
guides. Many have separate camera controllers to reduce weight and
size on the instrument itself.
Instrumentation cameras are often able to operate in very low light
levels. However this feature should not be confused with night vision
cameras. Night vision cameras often use sensors sensitive to infra red,
and illuminate the scene with an infra-red lamp. Intensification night
Sony Training Services 270

279. Broadcast Fundamentals
vision cameras are very sensitive, but are also designed to be
reasonably robust. Either they only employ mild amounts of
intensification, or the camera will not be damaged by exposing it to
bright light.
However instrumentation cameras designed for very low light levels are
designed for only for low light levels, not for a different kind of light, like
infra-red. They are not at all robust, and can often be damaged by
exposing them to normal light levels. They require expert setting up,
care and maintenance.
Intensified CDD (ICCD) cameras use an intensifier before a CCD
sensor. This boosts the amount of signal produced by the light before
the sensor reads it.
Reading CCTV camera specifications
All CCTV camera manufacturers produce specifications, making the
figures and details look as appealing as possible. Therefore, it is worth
investigating exactly how some of these specifications are found, and
things one should bear in mind when reading them.
Camera format
CCTV cameras are designed in a variety of formats depending on the
size of their sensor. All sensors have a 4:3 aspect ration, in common
with standard domestic television.
It is a common misconception that the camera format is the same as the
distance from one corner of the CCD sensor to the opposite corner, i.e.
a ½” sensor is ½” across its diagonal. This is not so. Sensor diagonals
are about 0.6 times the format size. The reason for this goes back to the
days of tube cameras where the sensitive area of the old 1” tube was
only about 0.6 of the overall tube diameter.
271 Sony Broadcast & Professional Europe

280. Part 23– CCTV, security & surveillance
T h e tr a d itio n a l 1 " c a m e r a tu b e
1"
S e n s o r d ia g o n a l a p p r o x
6 0 % o f tu b e d ia m e te r
1" 2 /3 " 1 /2 " 1 /3 " 1 /4 "
CCD
s e n s o rs
1 2 .8 x 9 .6 8 .8 x 6 .6 6 .4 x 4 .8 4 .8 x 3 .6 3 .2 x 2 .4
A ll d im e n s io n s
1 5 .9 d ia g o n a l 1 1 d ia g o n a l 8 d ia g o n a l 6 d ia g o n a l 4 d ia g o n a l
in m m u n le s s
s p e c ifie d
Figure 112 CCTV format sensor sizes
The table below shows sensor dimensions for various camera formats
and the ratio differences between the format and the sensors.
1” 2/3” ½” 1/3” ¼”
Sensor horizontal 12.8 8.8 6.4 4.8 3.2
Sensor vertical 9.6 6.6 4.8 3.6 2.4
Sensor diagonal 15.9 11 8 6 4
Sensor ratio 1:1.6 1:1.53 1:1.59 1:1.41 1:1.59
The camera format has an effect on the kind of lens that can be fitted,
and how it will behave. This is covered in more detail in the section on
CCTV lenses.
Resolution
Resolution is a measure of the resolving power of the camera.
All CCTV cameras, colour or monochrome, are the single sensor type.
The sensor pixels in colour CCTV cameras are divided between the
three primary colours. Thus, for the same sensor density, there is a
difference in resolution between monochrome and colour CCTV
cameras. Monochrome CCTV cameras will therefore tend to have a
higher resolution than colour CCTV cameras.
Sony Training Services 272

281. Broadcast Fundamentals
Still cameras often use the number of pixels in the sensor as a measure
of resolution. However this is not a good method of defining resolution in
video cameras. Sensor resolution will give a basic figure for the sensor
itself, not of the eventual output signal. In many cases only a proportion
of the pixels are actually used in the picture. If specifications mention
‘active pixels’ or ‘effective pixels’ rather than simply ‘pixels’, this will give
greater assurance that all these pixels are part of the picture.
The camera’s circuitry will also affect the sensor’s resolution. Badly built
circuitry will have a poor bandwidth that will reduce the resolution
provided by the sensor by the time the signal reaches the output. Having
a good sensor and bad circuitry is a waste. CCTV camera resolution
figures should always be related to the final output signal.
Resolution figures are sometimes given as vertical resolution. This is the
number of active lines in the picture. All PAL based CCTV cameras are
built around the PAL television system with 625 lines per frame. Of this
575 lines are active. All PAL based CCTV cameras should be able to
achieve a vertical resolution of 575 lines.
Most resolution figures normally define the horizontal resolution. This is
a measure of the number of individual pixels per line the camera is able
to resolve, and is measured in vertical lines. Horizontal resolution can
never be higher than the sensor’s horizontal resolution, and is often
lower, due to bandwidth limitations of the circuitry between the sensor
and the output.
Horizontal resolution and bandwidth are related by the equation :-
 1 
Bandwidth =  
 Period 
Each horizontal line lasts about 50uS long (exactly 52uS). The pixels, or
vertical lines, are divided up into this 50uS. The period is one clock
cycle, producing two vertical lines, one black, one white.
Therefore :-
50 × 10 −6
Period =
 Lines 
 
 2 
1 × 10 −4
=
Lines
Therefore the bandwidth can be found by combining these two
equations :-
 
 
 1 
Bandwidth =  −4 
  1 × 10  
 
  Lines  
  
273 Sony Broadcast & Professional Europe

282. Part 23– CCTV, security & surveillance
= Lines × 10000
These equations boil down to a very simple rule. If the number of lines or
pixels is measured in hundreds, and the bandwidth in MHz, the two are
equal, i.e. 400 vertical lines = 4MHz bandwidth, 600 lines = 6MHz
bandwidth. This is a rough approximation as the exact PAL line duration
is 52uS not 50uS.
Bandwidth, probably more than any other parameter, is the figure that is
more difficult to achieve. Bandwidth costs money and separates the
good cameras from the bad ones. For square pixels the horizontal
resolution would need to be 768 vertical lines, or pixels, which gives
almost 8MHz bandwidth! No CCTV camera can achieve this. Cameras
achieving 600 vertical lines are considered good quality.
SNR
A camera’s SNR is found by comparing the amount of video signal to the
amount of noise, in decibels, with the equation :-
 video 
20 log dB
 noise 
As a guide, an SNR of about 20dB is poor and is probably not viewable.
30dB will give a barely distinguishable image. 50dB is acceptable and
60dB good.
As a ratio of video signal to noise, 20dB is 10:1, and 60dB is 1000:1.
Sensitivity
Sensitivity is a measurement of how much signal the camera produces
for a certain amount of light.
Sensitivity can be measured as the minimum amount of light that will
give a recognisable picture, and is sometimes called ‘minimum
illumination’. Figures of below 10 lux should be possible for standard
CCTV cameras. However although this method provides an easy guide
to CCTV planners and installers, it is a highly subjective measurement.
What is a recognisable picture to one person may be unrecognisable to
another.
Professional and broadcast cameras use a different, more quantifiable
method for measuring sensitivity. The camera is pointed towards a
known light source. This is often a 2000 lux source at 3200K light
temperature (colour). The iris is then closed until the output is exactly
700mV.
Thus a reasonable sensitive camera may be f11 at 2000lux, whereas a
less sensitive camera may be f8 at 2000lux.
CCTV camera specifications are often not so consistent. Different lux
levels are specified. In the case of low light and night cameras normal
colour temperatures are meaningless because the camera is not
designed to be lit with standard 3200K light! These cameras often
Sony Training Services 274

283. Broadcast Fundamentals
specify the minimum illumination sensitivity, and should quote figures
very much less than 1.
Dome camera manufacturers specify sensitivity with the dome removed,
because the figure is better than with it fitted, Some give figures with the
dome fitted as well. The camera would normally be used with the dome
fitted. This factor needs to be remembered. Dome cameras need to be
more sensitive than other cameras, if they are to overcome the losses
through the dome itself.
Cameras with AGC
CCTV cameras with automatic gain control (AGC) add another
complication to the specifications. Manufacturers will quote sensitivity
figures with AGC switched on. However they will generally quote SNR
figures with the AGC switched off. The reasons for this are obvious. It
makes the figures look better!
Output formats
CCTV cameras use many different video output formats, from the simple
analogue composite output fitted to most cameras, through the analogue
Y-C output format, digital formats of one kind or another, and direct
computer network outputs used by some of the latest cameras.
Specifications always show the SNR, sensitivity, etc. from the best
output. The most common output connection people use is the analogue
composite output. Many cameras have it fitted and it is a simple
connection. However it is also the worst quality output. Some cameras
have a component output, so called the Y-C output. This provides higher
quality but is more difficult to connect.
Some newer CCTV cameras have computer network connections,
These cameras convert the video into a compressed data stream which
can be sent down a network cable. Many use the JPEG format, some
use the MPEG format, which can be modified and set up to give good
quality at low data rates.
Camera mounting
Most general purpose cameras have a screw hole underneath them to
secure them to a tripod or bracket. This is the same as is used by many
professional still cameras and is based on the ¼” Whitworth thread, with
20 thread per inch.
Enclosure types
Most CCTV enclosures quote conformance to the National Electrical
Manufacturers Association (NEMA) standards. These are American
standards but are often quoted in manufacturers specifications.
IEC Publication 60529 also specifies enclosure types, and are
sometimes used in specifications referred to simply as IP numbers.
Type Purpose Usage IP No
1 Indoor. General. Accidental contact prevention. Falling dust/dirt. 10
2 Indoor. Light water As Type 1 with protection from falling & light 11
275 Sony Broadcast & Professional Europe

284. Part 23– CCTV, security & surveillance
protection splashing non-corrosive liquid.
3 Indoor/outdoor. Light water As Type 2 with protection from sleet & ice. Dust & 54
protection. rain tight.
3R Indoor/outdoor. Light water As Type 3 with further protection from ice build-up 14
& ice protection
3S Indoor/outdoor. Light water As Type 3S but can still operate with heavy ice 54
& heavy ice protection build-up.
4 Indoor/outdoor. Heavy As Type 3 with protection from falling, splashing 56
water protection. and hose fed water, and condensation.
4X Indoor/outdoor. Heavy As Type 4 with corrosion protection. 56
water protection.
Corrosion resistant
5 Indoor. Light industrial. Protection from lint dust & dirt, light splashing, 52
dripping, seepage & condensation of non
corrosive liquids.
6 Indoor/outdoor. Light As 3R with protection from limited water 67
submersion. submersion.
6P Indoor/outdoor. Prolonged As 3R with protection from prolonged water 67
submersion. submersion.
7 Indoor. Hazardous Protection from light explosions, hazardous dust, -
conditions. pressure differentials, acetylene, hydrogen,
various hydro-carbons.
8 Indoor/outdoor. Hazardous Protection from light explosions, hazardous dust, -
conditions. pressure differentials, acetylene, hydrogen,
various hydro-carbons.
9 Indoor. Hazardous Protection from light explosions, hazardous dust, -
conditions. pressure differentials, metal dust, carbon dust,
grain dust, fibres.
10 Indoor/outdoor. Hazardous Mine safety. Health administration. Protection -
conditions. from methane and coal dust.
12 Indoor. Light industrial. As 5 but with oil protection & no knock-outs. 52
12K Indoor. Light industrial. As 5 but with oil protection & no knock-outs. 52
13 Indoor. Heavy industrial As 12 with heavier protection. 54
NEMA Type 4 is a popular enclosure type for many outdoor CCTV
cameras. Some attain NEMA Type 6 or 6P.
Sony Training Services 276

285. Broadcast Fundamentals
CCTV lenses
CCTV lenses are highly functional, and very much built to purpose. They
are simpler than professional and broadcast video, and still, camera
lenses. These lenses are on a quality equal with domestic and consumer
cameras and camcorders.
General purpose CCTV cameras were traditionally based on the 1”
video tube. Lenses were mounted to the camera with a C mount. (Cine
mount. 1” (25.4mm) diameter, 32 threads/inch. Back flange to sensor :
0.69” (17.526mm).)
Later general purpose CCTV cameras have the more compact CS lens
mount. This mount is exactly the same as the C mount but with a lens to
sensor distance reduced to 12.5mm.
Many new CCTV cameras are now supplied with fixed lenses. This
reduces cost, is simpler for the system designer to work with and install,
and allows the camera manufacturer to integrate the lens and camera
more closely, adding features to the complete camera that can only be
added if the lens and camera are fixed together.
Some smaller CCTV cameras are just too small to allow the lens to be
removed. Pin-hole cameras and web cameras often only have a simple
single convex lens.
Choosing CCTV lenses
Assuming you have a general purpose CCTV camera, that has been
delivered without a lens. What lens do you fit to it? What criteria do you
have to bear in mind when choosing a lens?
Focal length
CCTV lenses are available as either fixed focal length lenses or lenses
where the focal length can be varied. Fixed focal length lenses,
sometimes called prime lenses, are available in a number of different
focal lengths from fish eye and wide angle, through normal, to telephoto
and super telephoto.
Wide angle lenses
Wide angle lenses see a wider angle of view than the human eye.
Although it will cover a greater area, it is difficult to make out detail. The
image will also look distorted, with near objects appearing to be very
near and far objects very far. Fish eye lenses are super wide angle. The
angle of view may be over 180 degrees, and image distortion is so great
that it becomes circular.
Normal lenses
A normal, or standard, lens is a lens that will produce about the same
scene on the monitor as the human eye. Geometry and angle of view
are similar to the human eye, although the human eye actually has a
very odd view of the world that the brain sorts out for us. The normal
277 Sony Broadcast & Professional Europe

286. Part 23– CCTV, security & surveillance
focal length is about the same as the diagonal distance across the
sensor, so will vary from one camera format to another.
Telephoto lenses
Telephoto lenses have a narrower angle of view than the human eye.
They magnify the scene. They also shorten depth, with near objects and
far objects appearing closer to each other.
Super telephoto lenses allow you to see detail from a long distance. The
angle of view is very small, and they are difficult to set up and maintain
in the correct position. The slightest movement can push them off target.
The table below shows the focal lengths and angles of view for various
lenses, for all 5 common CCTV formats.
Angle ¼” 1/3” ½” 2/3” 1”
Fish eye > 100 < 1.5 <2 < 2.5 <3 <5
Wide angle 40 - 100 1.5 - 5 2-7 2.5 - 9 3 - 12 5 - 17
Normal or standard @ 40 @5 @7 @9 @ 12 @ 17
Telephoto 8 - 40 5- 24 7 - 34 9 - 45 12 - 62 17 - 90
Super telephoto <8 > 24 > 34 > 45 > 62 > 90
Zoom and vari-focus lenses
Variable focal length lenses fall into two groups, The motorised lenses
are commonly referred to as zoom lenses, just as all variable focal
length lenses are in still photography, domestic and professional
camcorders.
Manual variable focal length CCTV lenses are commonly referred to as
vari-focal lenses, to differentiate them from the motorised ones. The
focal length can be set up during installation. Once set up they
effectively become fixed focal length lenses to the operator, because
they cannot be altered remotely.
Format
The next thing to consider is the lens and camera format. Older CCTV
cameras were based on the 1” tube. All lenses were therefore matched
to these sensors. As tubes were replaced by CCD sensors, so camera
designs became smaller. These first sensors copied the 1” tubes,
allowing the same lenses to be used. Later, more compact, CCTV
cameras were designed with smaller sensors, which required new
lenses to match to. We now have 5 different CCTV lens and camera
formats. It is important to match each camera to the correct lens.
Sony Training Services 278

287. Broadcast Fundamentals
Sensor
R e s u ltin g im a g e
Lens
1 " le n s o n a 1 " c a m e r a
Sensor
R e s u ltin g im a g e
Lens
1 /3 " le n s o n a 1 /3 " c a m e r a
Figure 113 Good camera and lens combinations
Each sensor should have a matched lens fitted. If a 1/3” lens is used on
a 1” camera the lens will try to focus the whole image onto a small part
of the 1” sensor. The rest of the sensor will pick up nothing. Whether you
see a rectangular or circular image in the centre of the picture will
depend on if the lens itself has an internal rectangular mask.
Sensor R e s u ltin g im a g e
Lens
1 /3 " le n s o n a 1 " c a m e r a
Sensor R e s u ltin g im a g e
Lens
1 " le n s o n a 1 /3 " c a m e r a
Figure 114 Poor camera and lens combinations
However, if a 1” format lens is fitted to a 1/3” format camera it will still
work, although the optics will be incorrect. The lens will be trying to
279 Sony Broadcast & Professional Europe

288. Part 23– CCTV, security & surveillance
project the image onto a 1” sensor. Although the image will be in focus,
the sensor will only pick up the central portion on the image. It will
appear that you fitted a lens with a larger focal length. Put another way,
it will be as if you fitted a telephoto lens onto the camera
Lens mounts
An important parameter to consider is the lens mount. Most CCTV
cameras with removable lenses have a C or CS mount. C mount (cine
mount) was the original mount for CCTV cameras. CS (cine small) is a
newer mount intended for more compact design.
Both the C and CS mount are a 1” diameter, 32 threads per inch screw.
The only difference is that the distance from the back of the lens to the
sensor, the back flange to sensor distance, is 17.526mm on the C mount
and 12.5mm on the CS mount, i.e. about 5mm.
G o o d m o u n ts (in fo c u s ) B a d m o u n ts (o u t o f fo c u s )
1 2 .5 m m 1 7 .5 2 6 m m
C S le n s o n a C S c a m e r a C S le n s o n a C c a m e r a
1 7 .5 2 6 m m 1 2 .5 m m
C le n s o n a C c a m e r a C le n s o n a C S c a m e r a
5m m
1 2 .5 m m
C le n s & a d a p to r o n a C S c a m e r a
Figure 115 C & CS lens mount combinations
If the camera is a C mount you can only fit C mount lenses. If it is a CS
mount, you can fit either a C or CS lens. However if you fit a C lens to a
CS mount camera you must remember to fit a C-CS adaptor. This is
Sony Training Services 280

289. Broadcast Fundamentals
really nothing more than a threaded ring that pushes the CS lens away
from the camera by 5mm so that the back flange to sensor distance is
the same as that of a C mount.
It is important to remember that some lenses protrude from the back of
the back flange. It is therefore possible to damage either the lens or
camera if you try to screw a C mount lens to a CS mount camera without
an adaptor.
Ultra miniature lens mounts
Mounts used on smaller CCTV cameras include the M10.5 0.5mm and
M15.5 0.5mm threaded lenses. These offer a very small mount and are
popular in instrumentation CCTV and “lipstick” cameras.
Iris and aperture control
What is the iris?
The iris is a mechanical device that varies the size of a hole somewhere
inside the lens. The hole itself is called the aperture. The iris allows you
to control the amount of light through the lens.
The reason for an iris is that camera sensors have a certain range of
sensitivity. Too little light and detail in the shadows and gloomy areas is
lost. This whole area of the image will turn black. Too much light and
detail in light areas of the image tend to be burnt out. This whole area
will turn white and may spread into other areas of the image.
It is important to adjust the iris so that there is reasonable detail across
the whole image. (Of course, if you can also control the lighting this will
help.)
Aperture and depth of field
The iris also adjusts the depth of field. This is an important and often
forgotten aspect of the iris. Installers often open the iris as far as
possible, to create as bright an image as possible, and wonder why it is
so difficult to focus. A wide aperture gives a bright image and a narrow
depth of field. Focussing is more difficult. A narrow aperture gives a dark
image and a wide depth of field. Focussing is easier.
f stop numbers
The aperture is defined as an f number. The lower the number the larger
the hole in the lens and the more light gets through. The numbers are
standardised as “stop” numbers. The standard numbers start at 1 and
each stop is 1.4 times the last. Thus 1.0,1.4, 2, 2.8, 4, 5.6, 8, 11, 16, 22,
32, 45 and so on.
Each f stop lets twice the amount of light through as the next f stop up
the scale. The difference betweenf1, f1.4 and f2 is much greater than it
is for high f numbers. Therefore ½ lens stops are often used close to f1,
like f1.2 and f1.8.
281 Sony Broadcast & Professional Europe

290. Part 23– CCTV, security & surveillance
All lenses are defined by the lowest f number possible. The higher the
quality, the lower this number is. Hence a 9mm f1.2 lens is a better, and
probably more expensive, lens than a 9mm f1.8 lens. The f1.2 lens will
have larger lens elements in it to allow more light to get through when
the iris is fully opened.
The f number is also mathematically a function of the focal length of the
lens. Hence longer focal length lenses have higher f number ranges than
shorter focal lengths. Put another way, it is more difficult to get a lot of
light through telephoto and super telephoto lenses.
In broadcast and still cameras irises close to apertures of about f32. This
represents a hold only about 2-3mm diameter. CCTV cameras often
specify minimum apertures far smaller than this. f360 or even higher that
f1000 may be used. These lenses are necessary for ultra sensitive night
vision cameras, if they are also used for daylight operation.
CCTV lens iris control
A CCTV lens iris may be manual or automatic. Manual irises are set
during installation and cannot be altered by the operator afterwards.
They are good for indoor use where there is little change in lighting
conditions during the day (or night).
Automatic irises are motorised. There are two type of automatic iris
control, video servo and DC servo. With video servo iris control the video
signal is sent to the lens. Circuitry in the lens measures the video signal
and adjusts the iris so that the video signal is standard 1 volt. With DC
servo iris control the video signal is measured in the camera, and a
simple DC signal is sent to the lens to control the iris. This type of control
is sometimes called galvo control and the lenses called galvanometric
lenses.
Video servo makes the lens a little more expensive than DC servo, and
visa-versa for the camera, although, in practice, most cameras have
both video and DC servo outputs.
There has been some confusion in the
past about iris control connection between 1 3
the camera and lens. For a while there
were many different connector types.
Camera manufacturers would often
provide a plug that installers could fit onto
the end of the lens cable, before fitting it 2 4
to the camera. Some camera
manufacturers opted to fit simple screw D C s e rv o V id e o s e r v o
terminals to make it as easy as possible to 1 C o n tro l - +9v pow er
fit any lens. 2 C o n tro l + -
3 D r iv e + V id e o
However an increasing number of camera 4 D riv e - (G n d ) G nd
and lens manufacturers are opting for a
standard 4 pin square plug for iris control.
The connector is often called the Panasonic connector, or Hi-Rose
connector from the Japanese Hirose Electric company.
Sony Training Services 282

291. Broadcast Fundamentals
Camera sensor auto-iris
Some cameras have sensors that can assist the lens iris. The sensor
has an electronic shutter which can be used just like an iris. This is
explained in more detail on page 113.
Lens filters
CCTV lenses are sometimes fitted with a filter. These are used to correct
for colour imbalance, or protect the camera from possible damage. They
include neutral density, and neutral density spot filters, coloured filters
and polarising filters.
Special effect filters are not used in CCTV cameras. These would
remove from the clarity of the image and almost certainly go against the
purpose of the camera.
Filters are covered on page 91.
CCTV switchers and control stations
A CCTV system may be designed with just one camera and one
monitor. This is popular for instrumentation, machine control and remote
monitoring in harzardous environments.
However many CCTV systems are devised for security and surveillance.
In these scenarios there are likely to be many cameras involved. The
central control room could have one monitor fitted for each camera.
Some systems are designed this way. It allows an operator to view every
camera all the time.
However if there are a lot of cameras a single operator will find it more
difficult to keep and eye on all the monitors at once. If the site being
surveyed has little activity, there may be little need to have a monitor on
all the time, for every camera. Not to mention cost, which becomes a
problem if many monitors need to be purchased and maintained.
The solution is to view many cameras on just one monitor. This can be
done in two ways. Either you can switch the monitor to show a different
camera, or you can squash the output from many cameras and fit them
all onto one monitor screen as a mosaic.
Camera switching systems
By far the most common way of sharing one monitor amongst many
cameras is by using a series of switches to select the camera you want
to look at.
Switching monitor
The simplest way of doing this is to have a simple switch in the monitor
itself. The monitor has a row of push button switches on its front panel.
Pressing one of these buttons connects that camera to the monitor
screen. Switching monitors can be used for up to about 8 cameras.
283 Sony Broadcast & Professional Europe

292. Part 23– CCTV, security & surveillance
Simple video switch
This system can be improved by putting the switch in a separate box.
The switch box has many inputs, one for each camera, and a single
output for the monitor. Operation is similar to the switching monitor, but
this modular arrangement allows for a system to be built with more
cameras, and lends itself better to improvement and expansion later on.
Remote video control
Increasing the complexity a little more, systems are available that
separate the video signals from the control box. Video signals do not go
through the controller’s switches themselves, of indeed through the
controller’s push button panel at all. The push button panel simply sends
a control signal to a separate box that has all the video connections and
the video switchgear. This method removes the video from the
controllers panel, making it easier to route cables and making the
controllers area much tidier. Quality tends to be better because the video
signals are screened better.
Computer controlled switching
Extending this idea to its logical conclusion, the control link between the
controllers panel and the video switching box becomes a computer
network link. Both the panel and the box have network addresses. The
control panel itself is replaced by a computer. The push buttons become
virtual buttons on the computer monitor.
This approach provides the ultimate flexibility. Now that the control is
over a standard computer network, the control computer can be a very
long distance from the video switching box. The system can also be
designed with many control computers, and each computer may be
given different levels of access to the system.
Camera control systems
CCTV systems often have cameras with remote zoom, pan and tilt
capability. Control signals must be sent from the control station to all
those cameras that need them.
Zoom, pan and tilt are normally controlled by a joystick on the control
panel. The controller could have one joystick for every camera.
However, if there are many cameras with zoom, pan and tilt capability,
this could result in a control panel covered in joysticks. So it is logical to
provide one joystick that controls all the cameras. It is also logical to only
control the camera that the operator has selected to look through.
Linking camera motion control selection to camera viewing selection
Control signals are not sent to all the cameras at the same time. It is
logical to only control the camera that the controller is actually looking
through.
Switchers characteristics
CCTV switchers have a number of important characteristics that define
their quality and suitability.
Sony Training Services 284

293. Broadcast Fundamentals
Bandwidth
The quality of CCTV equipment is often defined by its bandwidth. ‘Width’
suggests a difference between a lower frequency and an upper
frequency. However video equipment is easily able to operate to very
low frequencies. We assume that the lower frequency is in fact zero. The
only interesting frequency is the upper limiting frequency. The upper
frequency limit is the same as the bandwidth.
All pieces of video equipment are able to transmit a certain range of
frequencies. Relatively low frequencies up to about 1MHz are easy to
process and transmit. Indeed it should be possible for video equipment
to handle frequencies of several MHz.
However above about 5MHz the higher the frequency the greater the
losses. There is no sudden loss with increasing frequency. The signal
power gradually drops and the frequency increases. So at what point
you do decide enough is enough?
A CCTV switcher’s bandwidth is defined as the frequency at which the
signal power has dropped by half it level. This is the same as a 3dB drop
(or a –3dB gain) in power. Some specifications quote the bandwidth or
frequency response as the 3dB point.
Pow er
-3 d B
B a n d w id t h F re q u e n c y
Figure 116 Bandwidth
Some specifications quote specific drops in power at specific
frequencies. This is not an ideal method as it makes it more difficult to
compare with other specifications. Indeed this may be done specifically
to hide a poor bandwidth.
Signal to noise ratio (SNR)
This is another important characteristic of any CCTV processing
equipment. There is always a certain amount of noise contained within
video signals.
285 Sony Broadcast & Professional Europe

294. Part 23– CCTV, security & surveillance
The SNR is defined as a ratio of the noise power to the video power. If
the video signal is properly set up at 0.7V, it is easy to find out what the
actual noise level is.
CCTV over IP
Character and shape recognition
Sony Training Services 286

295. Broadcast Fundamentals
Part 21 Numbers & equations
Decibels
A measure of relative power. First used to measure audio power as
Bels. The Bel is now seldom used, the decibel is far more common. 1B =
10dB. Now used to measure signal power in many application areas.
Bel is a logarithmic ratio defined as :-
 P1 
B = log 10  
 P2 
Where P1 and P2 are the 2 power levels being measured.
Therefore decibels can be found by the equation :-
 P1 
dB = 10 log 10  
 P2 
Decibels & absolute sound levels
Decibels are used as a measure of absolute sound levels. However the
decibel is a ratio, not an absolute quantity. It needs some reference
level. The threshold of hearing, the lowest level of sound that the human
ear can hear is used as P2 in the equation above.
The table below shows audio levels in dB’s.
dB Sound level
150-160 Eardrum perforation. Space shuttle taking off.
140-150 Jet fighter taking off
130-140 Threshold of pain.
120-130 Sheet metal rivet gun.
110-120 Rock cancert, on stage. Close thunder clap.
100-110 Busy motorway underpass.
90-100 Middle of orchestra playing 1812 overture.
80-90 Busy street traffic or motorway hard shoulder. Vacuum cleaner
70-80 Average factory.
60-70 Department store. Normal close conversation.
50-60 Average office.
40-50 Quiet street. Average household. Mosquito.
30-40 Soft music. Average fridge.
20-30 Country garden. Babbling brook.
10-20 Rusting leaves. Quiet whisper.
0-10 Rustling leaves.
0 Threshold of hearing. Perceived silence.
287 Sony Broadcast & Professional Europe

296. Part 24 – Numbers and equations
Decibel as a direct ratio
It is sometimes easy to forget exactly what signal ratio gives a specific
decibel level. The diagram below shows direct power ratios compared to
decibel quantities.
R a t io 1 0 0 0 0 = 4 0 d B
dB
30 R a tio 1 0 0 0 = 3 0 d B
20
R a tio 1 0 0
= 20dB
10 R a t io 1 0
= 10dB
R a tio 1
= 0dB
-1 0 10 20 30 40 50 60 70 80 90 100
R a tio 0 .5
= -3 .0 1 d B D ir e c t r a tio
-1 0
-2 0
Figure 117 Decibel to direct ratio relationship
You can see that a ratio of 1 is exactly 0dB, as one would expect. A ratio
of 10 is 10dB, and a ratio of ½ is about –3dB. Above 10dB the ratio to
dB relationship climbs at a dramatic rate. At 20dB there is a 100 direct
ratio, and at 40dB there is a 1000 direct ratio.
CD is 96dB signal to noise ratio. As a direct ratio this is about
4,000,000,000!
Decibels are used because the human ear is logarithmic, i.e. we are
very sensitive to the slightest sound, but can still handle relatively loud
sounds without damaging our ears.
Signal to noise ratio
The decibel can be used as a measure of signal to noise ratio. Using the
equation above P1 is the signal power and P2 is the noise power, thus :-
 Signal 
log 10  
 Noise 
Sony Training Services 288

297. Broadcast Fundamentals
Part 22 Things to do
Find depth of field picture
Finish the filter table
Rewrite the Dichroic block chapter
Sort out the CCD sensor chapter
Sort out the VTR chapter
Sort out timecode chapter
289 Sony Broadcast & Professional Europe