Beyond Boundaries: Leveraging No-Code Solutions for Industry Innovation
DIC_video_coding_standards_07
1. Video Coding Standards
• H.120
• H.261
• MPEG-1 and MPEG-2/H.262
• H.263
• MPEG-4
Thomas Wiegand: Digital Image Communication Video Coding Standards 1
2. Video Coding Standards
MPEG-2
digital TV
2 -6 Mbps
ITU-R 601
166 Mbit/s H.261
ISDN 64 kbps
Picture
phone
H.263
PSTN < 28.8 kbps
picture
phone
Thomas Wiegand: Digital Image Communication Video Coding Standards 2
3. H.120: The First Digital Video Coding Standard
• ITU-T (ex-CCITT) Rec. H.120: The first digital
video coding standard (1984)
• v1 (1984) had conditional replenishment,
DPCM, scalar quantization, variable-length
coding, switch for quincunx sampling
• v2 (1988) added motion compensation and
background prediction
• Operated at 1544 (NTSC) and 2048 (PAL)
kbps
• Few units made, essentially not in use today
Thomas Wiegand: Digital Image Communication Video Coding Standards 3
4. H.261: The Basis of Modern Video Compression
• ITU-T (ex-CCITT) Rec. H.261: The first widespread
practical success
• First design (late ’80s) embodying typical structure
that dominates today: 16x16 macroblock motion
compensation, 8x8 DCT, scalar quantization, and
variable-length coding
• Other key aspects: loop filter, integer-pel motion
compensation accuracy, 2-D VLC for coefficients
• Operated at 64-2048 kbps
• Still in use, although mostly as a backward-
compatibility feature – overtaken by H.263
Thomas Wiegand: Digital Image Communication Video Coding Standards 4
5. H.261&3 Macroblock Structure
= luminance pixel
= chrominance pixel
(two chroma fields)
• Intra/Inter Decisions: 16x16
macroblocks
• DCT of 8x8 blocks
• H.261: 16x16 1-pel motion
• H.263: 16x16 1/2-pel
motion or
• H.263 AP mode: 8x8 1/2-
pel motion with overlapping
Thomas Wiegand: Digital Image Communication Video Coding Standards 5
6. MPEG-1
• Formally ISO/IEC 11172-2 (‘93), developed by
ISO/IEC JTC1 SC29 WG11 (MPEG) – use is
fairly widespread, but mostly overtaken by
MPEG-2
• Superior quality to H.261 when operated at
higher bit rates (≥ 1 Mbps for CIF 352x288
resolution)
• Can provide approximately VHS quality
between 1-2 Mbps using SIF 352x240/288
resolution
• Technical features: Adds bi-directional motion
prediction and half-pixel motion to H.261 design
Thomas Wiegand: Digital Image Communication Video Coding Standards 6
7. Predictive Coding with B-Pictures
I B P B P
Thomas Wiegand: Digital Image Communication Video Coding Standards 7
8. Hierarchical Syntax
• "Video Sequence“
• "Group of Pictures“ = “GOP“ (GOP structure is
very flexible)
I-Picture P-Picture P-Picture
B-Pictures
time
Thomas Wiegand: Digital Image Communication Video Coding Standards 8
9. MPEG: Coding of B-Pictures
• Motion compensated prediction from two
consecutive P- or I-pictures
either
only forward prediction (1 vector/macroblock)
or
only backward prediction (1 vector/macroblock)
or
average of forward and backward prediction =
bidirectional prediction = interpolation
(2 vectors/macroblock)
• Half-pel accuracy of motion compensation:
bilinear interpolation
• Coding of prediction error with 8x8-DCT, uniform
quantization, zig-zag-scan as in I-pictures, VLC
Thomas Wiegand: Digital Image Communication Video Coding Standards 9
10. MPEG-2/H.262
• Formally ISO/IEC 13818-2 & ITU-T H.262, developed
(‘94) jointly by ISO/IEC SC29 WG11 (MPEG) and
ITU-T
• Now in wide use for DVD standard and DTV
(the most commonly used video coding standard)
• Primary new technical features: support for interlaced-
scan pictures and scalability
• Essentially the same as MPEG-1 for progressive-scan
pictures, and MPEG-1 forward compatibility required
• Not especially useful below 2 Mbps (range of use
normally 2-20 Mbps)
Thomas Wiegand: Digital Image Communication Video Coding Standards 10
11. Interlaced Video
Vertical Vertical
Horizontal Temporal
Thomas Wiegand: Digital Image Communication Video Coding Standards 11
12. H.263: The Next Generation
• ITU-T Rec. H.263 (v1: 1995): The next
generation of video coding performance,
developed by ITU-T
• Has overtaken H.261 as dominant
videoconferencing codec
• Superior to H.261 at all bit rates
• Wins by a factor of two at very low rates
• Versions 2: H.263+ (late ’97/early ’98) and 3
(about now) later developed
Thomas Wiegand: Digital Image Communication Video Coding Standards 12
13. Technical Features of H.263
• “Baseline” Algorithm Features beat H.261
– Half-pel motion compensation
– 3-D variable length coding of DCT coefficients
– Median motion vector prediction
– More efficient coding pattern signaling
– Deletable GOB header overhead
• Optional Enhanced Modes
– Increased motion vector range with picture
extrapolation
– Advanced prediction mode (AP mode): Variable-
size and overlapped block motion compensation
(OBMC) with picture extrapolation
– PB-frames (bi-directional prediction)
– Arithmetic entropy coding
– Continuous-presence multipoint / video mux
Thomas Wiegand: Digital Image Communication Video Coding Standards 13
14. Motion Compensation H.261 vs. H.263
• H.261 (1990): integer-pel accuracy, loop filter,
1 motion vector per MB
• H.263 (1995): half-pel accuracy, no loop filter,
1 motion vector per MB (already in MPEG-1,2)
• H.263 option „Advanced Prediction Mode“
- overlapped block motion compensation (OBMC),
- switch between 1 or 4 motion vectors per MB
• H.263 option „PB frames“
Thomas Wiegand: Digital Image Communication Video Coding Standards 14
15. H.263 AP mode: OBMC
remote luminance Current
block
luminance
block (8x8)
M
A
C
remote luminance remote luminance R
block block O
B
L
O
remote luminance C
block
K
Thomas Wiegand: Digital Image Communication Video Coding Standards 15
17. Performance of H.263 AP Mode
36
35 1)
2)
34
33
PSNR [dB]
32
31
30
29 1) H.263 TMN-9 AP-mode
2) H.263 TMN-9 w/o options
28
27
26
0 32 64 128
rate [kbps]
Thomas Wiegand: Digital Image Communication Video Coding Standards 17
18. H.263: PB-frames
forward prediction
P B P
bidirectional bidirectional
prediction prediction
PB-Frame
Thomas Wiegand: Digital Image Communication Video Coding Standards 18
19. Performance of H.263 PB-Mode
36
1) 3a)
35 2)
34
33 3b)
32
31 1) H.263 TMN-9 w/o options (6.25 fps)
30 2) H.263 TMN-9 w/o options (12.5 fps)
3) H.263 TMN-9 PB-mode (12.5 fps)
29 a) P-frames
28 b) B-frames
27
26
0 32 64 128
rate [kbps]
Thomas Wiegand: Digital Image Communication Video Coding Standards 19
20. Performance of H.263 and H.261
1) 2) 3)
34
4)
5)
32
30
28 1) H.263 v1 TMN-9, all options
2) H.263 TMN-9 w/o options
26 3) H.261 PVRG codec
4) H.263 TMN-9 w/o options, integer-pel ME
24 5) H.261 PVRG codec w/o loop filter
32 64 128
rate [kbps]
Thomas Wiegand: Digital Image Communication Video Coding Standards 20
21. H.263+ Feature Categories
• Improved compression efficiency (e.g.,
15-25% overall improvement over
H.263v1)
• Error resilience (1st resilient video
standard)
• Custom and Flexible Video Formats
• Scalability for resilience and multipoint
• Supplemental enhancement information
Thomas Wiegand: Digital Image Communication Video Coding Standards 21
22. H.263+: Better Ways of Coding
• Efficiency & Perceptual Enhancement
– Advanced intra-coding
– Deblocking filter (in the loop)
– Alternate inter VLC for heavy motion
– Modified quantization/coefficient range
– Improved PB-frames (and B frames too)
– Tweaks of prior features (RC, UMV, AP, CPM)
• Dynamic Resolution Features
– Reference-picture resampling (also global motion)
– Reduced-resolution update
Thomas Wiegand: Digital Image Communication Video Coding Standards 22
23. H.263+: Improved Error Resilience
• Slice-structured coding (packetization, low
delay, object-oriented coding)
• Independent segment decoding
(packetization, encoder parallelization)
• Reference picture selection (with optional
back-channel)
• Scalability Features (see part 3)
Thomas Wiegand: Digital Image Communication Video Coding Standards 23
24. H.263+: New Kinds of Pictures
• Custom Video Source Formats
– Custom picture resolutions
– Custom pixel/picture aspect ratios
– Custom picture clock rates
• Scalability Pictures
– SNR scalability (fidelity enhancement layer)
– Temporal scalability (B pictures)
– Spatial scalability (resolution enhancement
layer)
Thomas Wiegand: Digital Image Communication Video Coding Standards 24
25. Spatial Scalable Coding
EI EP EP EP EP
I P P P P
Thomas Wiegand: Digital Image Communication Video Coding Standards 25
26. What New Applications Does H.263+ Enable?
• Error/packet-loss resilient video
• Scalable multiple bit rate video at low bit rates
• Flexible-format low bit rate video
• High frame rate video (e.g., 72 frames / sec)
• Progressive picture quality refinement
• Dynamic video picture resolution
• Graphics/still-picture snapshot coding
• Object-layered video using chroma key
• Special effects / global motion
Thomas Wiegand: Digital Image Communication Video Coding Standards 26
27. MPEG-4: H.263 + Additions + Variable Shape Coding
• MPEG-4 (v1: early 1999), formally ISO/IEC
14496-2: Roughly follows H.263 design and
adds all prior features and various “trick
modes” and (most important) shape coding
• Includes zero-tree wavelet coding of still
pictures, segmented coding of shapes,
coding of synthetic content
• v2 (early 2000) developed, v3 and v4 in
progress
Thomas Wiegand: Digital Image Communication Video Coding Standards 27
28. MPEG-4 v1: Simple Profile
• H.263v1 Baseline (exact compatibility to H.263, plus
custom picture format with a different picture header)
• Variable block-size and picture-extrapolating MC
(parts of H.263v1 Annexes D & F & H.263+ Annex J)
• Spatial-Predictive Intra Coding (similar to H.263
Annex I)
• Altered step size for DC coefficients
• Relative placement of rectangular pictures (sort of
H.263+ Annex R)
• Slice-structured coding (like H.263+ Annex K)
• Data partitioning and reversible VLC (not in H.263
until v3)
Thomas Wiegand: Digital Image Communication Video Coding Standards 28
29. MPEG-4 v1: Core Profile
• Binary shape coding (shape coding is the
most unique new feature of MPEG-4 –
different than H.263+ Annex L chroma
keying)
• B-picture temporal scalability (a subset of
H.263+ Annex O)
• P-picture temporal scalability (a subset of
H.263+ Annex N)
• MPEG-2-style inverse quantization (not in
H.263)
Thomas Wiegand: Digital Image Communication Video Coding Standards 29
30. MPEG-4 v1: Main Profile
• Grey-scale shape coding (soft representation
of shapes)
• Interlace (backward compatibility for 1940’s-
era analog compression technique)
• Static sprite coding (warpable pictures,
different than H.263+ Annex P)
• Scalable still pictures
Thomas Wiegand: Digital Image Communication Video Coding Standards 30
31. MPEG-4 v1: Other Technical Features
• Temporal & spatial scalability (a subset of
H.263+ Annex O)
• Overlapped block motion compensation (part
of H.263v1 Annex F)
• 12-bit video
• Dynamic 2D mesh coding
• Face animation modeling
Thomas Wiegand: Digital Image Communication Video Coding Standards 31
32. MPEG-4 v2
• Fidelity
– Quarter-pel Motion Compensation
– Global MC (somewhat different than
H.263+ Annex P)
– Shape-Adaptive DCT (for shape-based
coding)
– Reduced-Resolution Update (H.263+
Annex Q)
• Error Resilience
– Reference Picture Selection (H.263+
Annex N)
Thomas Wiegand: Digital Image Communication Video Coding Standards 32
33. MPEG-4 v2
• Object based spatial scalability (spatial scalability for
arbitrary shaped coding)
• Multiple auxiliary components (specialized apps)
• Mesh coding for Body Mesh and 3-D Mesh (synthetic
and semi-synthetic content)
• Still-Picture Coding
– Wavelet tiling for still picture (random access
within a still picture)
– Error resilience for still pictures (recovery from
errors)
– Scalable arbitrary shape for still picture coding
(scalability with shape coding for still pictures)
Thomas Wiegand: Digital Image Communication Video Coding Standards 33
34. H.263++ New Version 3 Features, I
• Annex U: Fidelity enhancement by
macroblock and block-level reference picture
selection
- Improved compression performance
- Improved error resilience
• Annex V: Packet Loss & Error Resilience
using data partitioning with reversible VLCs
(roughly similar to MPEG-4 data partitioning,
but improved by using reversible coding of
motion vectors rather than coefficients)
Thomas Wiegand: Digital Image Communication Video Coding Standards 34
35. H.263++ New Version 3 Features, II
• Annex W: Additional Supplemental Enhancement
Information
– IDCT Mismatch Elimination (specific fixed-point
fast IDCT)
– Arbitrary binary user data
– Text messages (arbitrary, copyright, caption, video
description, and URI)
– Error Resilience:
• Picture header repetition (current, previous,
next+TR, next-TR)
• Spare reference pictures for error concealment
– Interlaced field indications (top & bottom)
Thomas Wiegand: Digital Image Communication Video Coding Standards 35