Unraveling Multimodality with Large Language Models.pdf
Digitisation (of audio, video and film): best practice and standards
1. Digitisation: Best Practice and
Standards
Sexto Seminario Internacional de Archivos Sonoros y Audiovisuales
MIÉRCOLES 25 DE JUNIO DE 2014
MESA 5: La digitalización de los acervos
Richard Wright
2. Best Practice? Examples
• Columbia University Library:
http://library.columbia.edu/services/preservatio
n/audio.html
“Audiotapes are preserved by re-recording
original recordings onto reel-to-reel tape,
following nationally recognized preservation
practices and guidelines; “
3. Columbia Univ. Library
• “Analog copies, carefully made and properly
stored, are considered nationally to be the
best medium for long-term preservation of
sound recordings.”
• “...into the future, when digital technology has
progressed to new levels of accuracy, we will
still have the option to create new digital
versions from the full continuum of sound.”
4. Columbia Univ. Library
• “It is clear that digital copies are the better
option for access and use. CDs suffer much
less from wear and tear during use than do
reel-to-reel or cassette tapes.”
• ... “Finally, high-quality CD-Rs are a stable
medium with a good life expectancy.”
5. BBC Archive
• 1997: vinegar syndrome in acetate magnetic
sound tracks (sepmags)
• Took decision to digitise to CD (with pulses to
mark sprocket holes)
• AND – to make new polyester analogue copies
• By 2001 – use of film viewing declined almost
to zero; polyester copies almost never used!
6. British Library Sound Archive
• 1980s- off-air recording of radio broadcasts
• Decided to ‘go digital’ – very few options
(before CD-R, in the days of 5 MB hard drives)
• Chose Sony PCM-F1 (betamax videotape)
• 15 years later- a struggle to migrate off this
format
7. British Film Institute
• Oral history recordings, mainly on cassette
• Copyied to BetaSP videotape (using only the
audio tracks) – because they had BetaSP !
8. BBC Archive
• Spent ten years copying analogue video tape
to Panasonic D3 format
• Because D3 was the agreed production format
in the BBC
• After 2004, D3 obsolete; tapes had to be
migrated (to files on LTO data tape)
• Problems finding enough equipment and
“head life”
9. INA (Paris)
• Late 1990’s: early adopter of datatape (before
LTO)
• Early adopter of Sony Petasite tape library and
SAIT helical-scan datatape
• Many problems, ended up migrating to a
different technology
• Eventually migrating again to LTO
MANY archives are on their 2nd or 3rd tape
library technology
10. Photographical Collections
• Early 1990s, many professional users of Kodak
Photo-CD, Picture CD formats
“Keep your memories safe for generations.
Preserving your favorite pictures is easy with a
KODAK Picture CD.”
http://wwwuk.kodak.com/global/en/consumer/k
iosk/kioskProducts.jhtml?pq-
path=2301249#tab=tab11
Proprietary encoding and storage formats on Photo-
CD; all have had to be migrated
11. Magneto-Optical Media
• Plasmon – UDO discs guaranteed for 50 years
• Drives were promised that would continue to
read the discs
• Company folded 2008, with 6000 customers
• But – assets bought by Alliance Storage
Technologies, Inc. in 2009
• My conclusion: money was lost, but data was
not !
12. Academy of Motion Picture
Arts and Sciences (AMPAS)
• 2007: “digital 12 times more expensive than
film”
• Quoted in NY Times – and just about
everywhere
• “the report’s startling bottom line: To store a
digital master record of a movie costs about
$12,514 a year, versus the $1,059 it costs to
keep a conventional film master.”
13.
14.
15. AMPAS
AMPAS estimation of digital costs for a century:
1) Find cost in 2005/6
2) Multiply by 100
AMPAS: $500 per terabyte per year for near-line
data tape storage
2014 cost for Amazon Glacier: $120/TB/yr (and
dropping)
16. Who can you trust?
• IASA, AMIA, ARSC
• Library of Congress
• PrestoCentre