ELL: Mime-Version: 1.0

[Brian Levy]

Dear Brian,
My name is Peter Copeland, and I am the "Conservation Manager" at
the British Library National Sound Archive. I'm slightly puzzled why
no-one has jumped in before me; but you have of course opened a very
big issue!
In principle, I support the strategy that analogue media should be
converted to a digital medium with error-correction, so that as errors
build up over time, it is possible to monitor the state of the
degradation and invoke a "cloning" operation before the failures become
catastrophic. My policy is also to digitise analogue audio *in
duplicate*, using two different types of carrier and storing them at
opposite ends of the country, to spread the burdens of self-destroying
media and World War III. I also make my staff and contractors use media
with greater "power-bandwidth product" so we do not lose anything, and
to use their ears (in the case of sound, of course!) to check the
digitised version sounds exactly the same as the original. This is
besides objective measurements of frequency-respose, speeds, noise,
etc.
However, there are two "lateral" issues against which I must warn
you.
One is the problem of "software." As I'm sure you know, the
computer industry is bedevilled with software incompatibilities, and
no matter how well the digital data is stored, if you cannot get the
recording into a state which human beings can appreciate, you have
failed as an archive. Unfortunately, computer software can be
copyrighted, and (in Britain) it lasts 70 years after the death of the
person who wrote it. Therefore you cannot copy the software to enable
an archival recovery programme to be monitored (let alone used). For
this reason, we currently honour patent law (which covers *hard*ware)
and lasts only eighteen years, and use formats covered by patents
rather than software.
My employer, the British Library, is currently planning to
circumvent this difficulty by storing the digitised artefacts
indefinitely. This involves a really massive digital store. The plan
is to store all our books, photographs, manuscripts, etc. in the
digital domain, using an IBM computer system with enough intelligence
capable of monitoring itself, doing its own cloning as required, and
storing the results in two separate locations. I don't think this is
feasible for an archive such as yours. Either prices will have to come
down or co-operative agreements will have to be invoked to make
everything work. Then, a century from now (when software copyright has
expired), one can get a representation back without breaking the law.
The other issue is getting analogue originals digitised in a way which
conserves all their properties. For audio this is comparatively
simple, we are only digitising a representation of two analogue
sound-pressures. (Anyone who thinks I'm oversimplifying, please shut
up for the moment). But for moving pictures, the problem is much more
complex. Different individuals have different tolerances to different
defects. I had to leave the television industry because my eyes were
predicted to fail, so I know I'm not normal; but I can see many
defects on DVD, for example, which other people evidently cannot. So
the digitisation of moving pictures is currently at a very primitive
state compared with audio.
On basic "information theory" arguments, I consider the format known as
DVC (with its extensions DVC-Pro or DVC-50, and DVC-HD or DVC-100, all of
which use the same chipsets - hardware) offers a route to the future. It is
not specific to a particular frame-rate (like film or video), and the (by
now) conventional metal-particle tape has been proven to last a decade at
least (unlike metal-evaporated, used by Super8). However, I do not think
there is yet enough hardware in the world for this format to be easily
readable in decades from now. But it is the nearest thing to an archival
format for video which has yet appeared.
Returning to audio: I am playing "devil's advocate" now. I respect the
strategy of the Library of Congress that analogue is more likely to be
playable in future. The trouble is that the power-bandwidth product of
analogue media always degrades with time, and this affects engineering
test tapes just as much as tapes of audio subject-matter. So you can never
know where you are as the sound degrades (at least, if you don't have
access to new and properly-made test tapes)! It is at least sensible to
consider digital alternatives, even if they may affect the analogue sound
(or image), on the principle "don't put all your eggs into one basket".
For audio, we use CD-Rs ourselves. Besides anything else, I support
the idea that the more CD players there are in the World Out There, the
more likely it is that we shall be able to play CDs in future years.
(There are now more than a billion, which is unprecedented).
You can see I am a very conservative philosopher, so I will end with
a very practical suggestion. Whilst I approve of Philips Red-Book
Standard CD-R media for storing most analogue audio (only high-end audio
and long running-times defeat it), I must urge you not to use "silver"
CD-Rs. Assuming you mean CD-Rs with the *metal* silver, these have been
tested to give longevity in a laboratory "standard atmosphere." But this
ignores the fact there is sulphur in the atmosphere of most areas of the
world, which can leach into the disc and form silver sulphate, so the
disc goes brown. The very earliest CDs made by Philips Dupont Optical
here in Britain suffered this fault, and we've been living with it ever
since. (I must, in all fairness, say that Philips have replaced all their
affected CDs honourably). But we conservative nutters do not use CD-Rs
which are silver *in colour*, to be certain! We only use ones with a gold
(metal, not colour) reflective layer.
I will stop now, I must give the rest of the world a chance to
comment.
Peter Copeland
<peter.copeland at bl.uk>


------------------------------------------
Brian Levy
Cultural Activist
Kiwat Hasinay Foundation:
Preserving Caddo Heritage
211 W. Colorado Ave.
Anadarko, OK  73005  USA
(1) 405-247-5840
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.linguistlist.org/pipermail/endangered-languages-l/attachments/20001020/f8ab138b/attachment.htm>