[Table of Contents]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [ARSCLIST] Digital Audio Preservation Question

At 07:28 PM 2008-04-06, Don Cox wrote:
On 06/04/08, Richard L. Hess wrote:

> There should be no additional loss converting the MP3 file to the WAV
> file. The information has already "taken the hit" when the MP3 file
> was made in the first place.
Not true, unfortunately. The decompression software also audibly affects
the sound quality.

Hi, Don,

Since the bits are thrown away during the compression cycle, are you saying that different decompression codecs sound different? I had always attributed the difference in sound between players to be the differences in the analog chain associated with the player.

I've always found MP3-WAV on the computer to be pretty transparent -- surprisingly so, actually if you don't compress too far. This is opposed to say a Palm as a listening device. For example, an MP3 bounced back to a WAV and played from a good portable CD player sounds MUCH better than the SAME MP3 played on a palm into the same listening chain.

What have you found?

Please note, my suggestion about archiving the MP3 was based on the fact that it was a DSPACE implementation and the SPECIFIC DSPACE implementation that I am aware of, U of Toronto's TSPACE, the last time I checked about two years ago was planning on migrating MP3 files but was offering no guarantee on WAV files. That policy may have changed partially due to my explaining why it was a good idea to at least commit to migrating BOTH types.

I agree with Don that keeping the original is important, and I agree with Brandon and George that in the GENERAL case it would be good to produce a WAV file, all other things being equal (which in my experience with one DSPACE implementation they were not).

I also agree with Tom that archival-quality material should have never been stored as an MP3 in the first place, but sometimes the world is not perfect and stuff happens--which was what I assume the original poster was asking about.

I did receive an off-list comment about why I normalized "preservation masters" and the reason for that is that the levels vary widely tape-to-tape and sometimes within each tape. I ingest at 24 bits, but we had agreed to save 16 bit files. I would think not normalizing in this context would be the poor way to go. If the 24 bit files were to be preserved, then no normalizing should be applied.

Also, the content of these tapes was very poorly recorded and most of the lectures (remember these are lectures) are now only available on about a 4th generation cassette. The cassettes limit useful data (like ORIGINAL bias for Jamie's magic process) that have already been provided, so at this point, the cassette noise and other 3.75 or 1.88 in/s 1/4-track intermediate generations also limit the useful information. In almost all instances going back to reels where available (and in almost all cases they are not originals, either) results in better sound than the cassettes I was originally going to transfer. Going back to the original reels was actually not in the scope of work I contracted for, and there was no more money in the budget, but I am doing it anyway -- parallel ingest is paying for it, essentially.

In a few hours, part of the client team will be here and I'll make a pitch to find more money to Plangent the tapes. The one-minute sample is a huge improvement, but not perfect, but this is wow that runs more than an octave plus or minus in some instances and different hum signatures cross over each other! (Thanks, Jamie).

There are LOTS of choices to be made in any archival stream and, unfortunately in my present project to which I was referring, many of the choices (including the available funds to pay for "in perpetuity" archiving) are limited.

We did manage to talk the client OUT (note this please, those who choose to brand me as "too CD Friendly") of only going to CDs. In fact, no CDs are being produced because of the
additional cost -- not only the CD production, but also the 8% sales tax here in Ontario that would be levied if I created tangible personal property. The fact that I can upload to UofT (approx 600 GB = approx 600 hrs) saves that 8% ON THE WHOLE PROJECT!!!!



Richard L. Hess email: richard@xxxxxxxxxxxxxxx
Aurora, Ontario, Canada (905) 713 6733 1-877-TAPE-FIX
Detailed contact information: http://www.richardhess.com/tape/contact.htm
Quality tape transfers -- even from hard-to-play tapes.

[Subject index] [Index for current month] [Table of Contents]