Re: [ARSCLIST] Digital Audio Preservation Question

[Don Cox <doncox@xxxxxxxxxxxxxx>]

On 07/04/08, Tom Fine wrote:
> Hi Brandon:
> 
> Can you cite a reasonable scenario where there will be an ability to
> read bits and bytes in the future but the whole idea of CODECs was
> forgotten? 

The idea may not be forgotten, but specific codecs and file format
decoders can be almost impossible to obtain.

There are already many image formats that are as good as lost.

> Especially the MP3 CODEC? I can't think of one, so that's
> not my reason for being anti-lossy schemes. My reason is simply audio
> quality. As for blowing up to WAV, as I said, it won't gain any
> quality but might (small probability) make the file slightly less
> vulnerable to being zapped by a drive-sector flaw due to less packing
> density (that might not be the right computer-engineer term, but what
> I'm saying is in WAV format, more space on the drive, more sectors are
> required for the same linear time in audio, so one bad sector might
> zap for instance a few measures that might be repeated elsewhere in
> the WAV file but it could completely destroy most of a
> lossy-compressed MP3 file and render it un-fixable). However, I think
> you run about the same probability of that sector cropping up in the
> file table or even the boot blocks, so the golden rule is many copies
> in many places.
> 
> -- Tom Fine
> 
> PS -- for these open-source volunteer-work CODECS, the idea that they
> might be unreadable in 50 years seems slightly more plausible. But MP3
> is an industry standard established by paid scientists and is widely
> licensed by for-profit companies and propagated.
> 
> ----- Original Message ----- 
> From: "Brandon Burke" <burke@xxxxxxxxxxxxxxxxxxx>
> To: <ARSCLIST@xxxxxxxxxxxxxxxx>
> Sent: Monday, April 07, 2008 3:50 AM
> Subject: Re: [ARSCLIST] Digital Audio Preservation Question
> 
> 
>> Tom,
>> 
>> Converting an accessioned archival MP3 file to WAV doesn't, in an of
>> itself,
>> mean creating
>> a 24/96 derivative. And it sounds to me like you're jumping to that
>> conclusion.
>> 
>> In this case (Martha's original point), the issue is readability in
>> the
>> future: 10, 20, 50 years down
>> the road. Of course, my guess is as good as anyone else's here but,
>> i think
>> conversion to WAV
>> (16/44.1) still sounds like a pretty good idea..
>> 
>> Brandon.
>> 
>> 
>> 
>> Right, Richard, I read the original post. I was commenting about
>> lossy
>>> formats in general.
>> 
>> As for the original poster's specific question, I don't see any
>> benefit in
>> blowing up an already inferior file, but I'd make several extra
>> copies in
>> different places on the theory that one bad sector could destroy an
>> MP3 file
>> whereas it might only cause a fixable glitch in a WAV file due to
>> the much
>> denser info-pack of the MP3 (ie what's left after the lossy
>> compression
>> packs more audio linear time into fewer hard drive sectors than if
>> the file
>> had been left full WAV). However, there's equal probability that
>> the sector
>> that will fail first will be the file table so the whole drive is
>> rendered
>> damaged, perhaps fatally. So it's a gray area. It all comes down to
>> many
>> copies in many places as far as digital storage but as far as audio
>> quality,
>> I believe there is no good argument that lossy compression is ever
>> a good
>> idea with archival versions of things.
>> 
>> -- Tom Fine
>> 
>> 
Regards
-- 
Don Cox
doncox@xxxxxxxxxxxxxx