[Table of Contents]


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

Re: arsclist Digital knowledge preservation



At 09:33 AM 6/18/2002 -0700, Richard L. Hess wrote:
We are having this precise discussion in the AMIA (Association of Moving Image Archivists) mailing list where it is heretical to suggest any form of lossy compression. On the other hand, we've had many analog systems that provide lossy storage to the original signal.

We have also had the discussion here that the CD Audio Red Book standard isn't good enough for archival purposes and we need higher resolution, although many of us are happily transferring sound to uncompressed, 44.1/16 CDs.

One thing to consider in selecting a format to store to is that it is truly as lossless as possible for the vast majority of material contained in the archive. On the other hand, the format has to be well-supported with the anticipation that players will be available long into the future.

There is lossless compression available for digital audio. Two options are MonkeysAudio (my choice) and Shorten. I verified that there were no losses by compressing a file, uncompressing it, and comparing the result with the original. They are identical at the bit level. However, the available compression is only of the order of a factor of two. Whether that is sufficient to be worthwhile and whether the code is likely to be available indefinitely or for different platforms are factors in the choice. For my purposes, lossless as well as lossy compression has its place.


I do not know the bitrates that are being coded to for the MP3 audio files mentioned in this post, but the difference between 192Mb/s MP3s and mono WAV files at 44.1/16 is about 1:3.7. I'm wondering if the savings of $7.00 and seven discs is worth the additional compression. These figures relate to making Red-Book audio CDs. The difference would be $3.00 and three discs if there were mono WAV files on the CDROM.

One important item to consider for audio transfers is that the cost of the final media is only a very small percentage of the final transfer cost. The biggest single cost is the staff to do the transfer as that staff needs to be skilled in making the transfer and in all of the arcane optimizations that must be done to (and the mechanical pre-processing that must be done before) the analog reproduction chain.

The cost of the medium per se is no longer a substantial issue. The cost of managing it - storing, indexing, accessing - is far greater in practice.


Note that in many contexts I refer to the MP3 discs as catalogues of the originals. Thought of in that light, their role is easier to recognize. They are (cross-)indexed representations of the archival material. Fidelity is sufficient for the purpose just as a catalogue of an art exhibit is sufficient for its purpose - and no one would confuse the representation in the catalogue with the original.

The good news, however, is that a large collection of items are being made accessible and being preserved in at least one form for future generations. I believe that our grandchildren will be happier if more is saved, albeit at SLIGHTLY lower quality than if only the pinnacle of quality is saved and many things were lost.

Accessibility is critical here, IMHO. Even the dedicated scholar will be at an advantage if he has a catalogue of the holdings of a repository sufficient to determine the value in the repository. The student or casual researcher may be content with the catalogue instead of visiting the repository and satisfying the conditions for access to the original.


There is a fallacy common to mathematics and geography: confusing the map with the territory. The catalogue is not the source material; it is a representation of the source sufficient for limited purposes and designed to serve those purposes. Even if the high-fidelity, lossless file is accessible, it may not be as useful for most purposes as a much smaller one. On historical material, a 32 Kbps MP3 may serve most users better than a 1.4 or 2.8 Mbps WAV. It will certainly be easier to distribute, hence to access.

Mike

mrichter@xxxxxxx
http://www.mrichter.com/


- For subscription instructions, see the ARSC home page http://www.arsc-audio.org/arsclist.html Copyright of individual posting is owned by the author of the posting and permission to re-transmit or publish a post must be secured from the author of the post.


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