Archives, FRBR, and Digital Preservation

Blogging time was rather limited last week--so I hope to make up for that this week.

Just a quick rundown on last week's activities--

1) Archives- Worked on some listening exercises and tape configuration exercises. I really got my hands into open-reel tape (after washing my hands, of course). I attended a very interesting Brown Bag Luncheon with Jen Riley and Jon Dunn on digitizing and delivering audio and video. I knew a lot of the physics for audio that Jon was talking about, but the video stuff just floored me. To think it takes something like 150 gb/hour to deliver high-definition, uncompressed video (including sound) over the web. No wonder video preservation is nowhere close to where audio is now.

2) During work at WFIU this week, I couldn't help thinking that a lot of people just can't separate out in their minds the idea of a piece of music (work), a performance of music (expression or instantiation), and a physical object that might hold those things. We probably have over thirty recordings of Beethoven's 5th Symphony. In that list could be ten recordings with Claudio Abbado conducting (for example). Then maybe Abbado was conducting the Vienna Philharmonic on five of those. And of those five, three have the same recording date (let's say March 15, 1991 at the Musikverein)--meaning that that one performance is imprinted on five items that are issued by a record label (ex: Deutsche Grammophon). But it's doesn't stop there, because DG could publish it multiple times, and in various forms: 1) the entire work; 2) separate movements of that work (on a highlights album); 3) excerpts of those movements (like on a "Beethoven for Your Brain" album). And hopefully, this is all clear on the disc you're holding. More often that not, it's not clear--and major discographic searching ensues. It's even more problematic for folk music and church hymns, whose music is often "traditional" (i.e. we've forgotten who wrote it, or we assume it's PD by now) and whose lineage is lost because attribution was never given.

Of course this models assumes a master-servant hierarchy.

* That there is one ultimate work created by discrete people or associations of people.
* That there is such a thing as a work, or a group of works that are bound together.
* That there is one master recording from which all others are pressed. Or in the case of today, a master file from which derivative copies can be made. It also assumes all copies from this master file or recording will be . That's not true of most uncompressed files, which can be copied ad nauesam without loss of fidelity or resolution, but it is for compressed files which introduces artifacts the further you get away from the original.
* So then, what is the value or benefit of having a first-edition WAV file, as opposed to a 10th generation one? Nothing in theory--it should have no more value than the tenth. But is that file signal processed, equalized, or altered in any way? Hopefully, there is documentation that will tell. Preservation metadata to save the day!

3) I've been thinking about how I would structure digitization/preservation projects of my own, if I had the opportunity. There's quite a lot to think about in constructing workflows and creating objectives, maintaining best practices, etc. Plus, the effort of getting the funding for such endeavors. In terms of access, what are the needs of your users? Where are they? How they will they listen/view what you have online? What kind of software will they need? If scholars are using it, how will they gain access to the uncompressed audio, if they need to inspect it for research purposes? The more you show, the more they want. That made me think of how expensive it is to stream audio and video, and just how proprietary many streaming servers are (Real, Windows Media). How can we really give long-term access if our fates are tied to those of corporations?