I'm not sure how discussing of a high-level musical description for a database differs from a high-level musical description for a stream. Actually I thought that to actually put such a thing in a db you'd actually first analyse a stream. So I'm really puzzled, as to what is the distinction that is being made here, and why.
On the one hand we have lower level issues of simply how to analyze waveforms. From these methods you get raw data about the structure of the music itself. These are essentially represented by data structures of some sort. Given a large amount of this data, in a database perhaps, we come to the question of what one wants to do with it all. Thus we have the level of waveform analysis: getting the raw data from a waveform, and on the other hand we have programs that analyze the results of the waveform analysis. To have an 'artificially intelligent' program draw conclusions about music requires the data pulled from the waveform. But looking at the waveform could be done without the need for anything more, eg. bonk~ and fiddle~ which can be used quite simply within a patch. I wonder if that explains the neccesity for a distinction at all. Sorry if it doesn't, I'll try again.
No it's certainly not, because old compression schemes do not look at the whole file all of the time; they are highly local, as a way to save effort. E.g. maybe with Lempel-Ziv you'd use a 4k or 32k or 256k window.
Of course, my language was sloppily unspecific. By saying 'any old compression scheme' I was refering rather to the point that the compression scheme does not have to be tailored to this specific application. However, even a compression algorithm with a small window would yield some results that are useful, but I'm sure that you're right that something with a larger scope would generally be more useful. Again, there has been some research in to which kinds of algorithms work best for different applications. It may be useful to have a smaller window with some music--the similarities found then would be in harmonies, rather than larger scale properties (maybe thematic structure and repetition?) that may differ greatly within a genre. It all depends on what you are comparing, and what similarites you are looking for.
Art is in denial.
I'm sure that art is in denial. Hippie-dippie-you're-okay-I'm-okay critical theory is hogwash (IMHO), and often questions of aesthetics _are_ pushed off to other elements, as in lots of conceptual art. Personally, though, I don't care about aesthetics or critical theory (the two go hand and hand, don't they?). If I make art, that's that. While I can defend my position with critical arguments, I don't generally find this necessary. You may argue that this in itself is aesthetic, the aesthetic of the state of mind of the creator or something, the aesthetic of the enjoyment of the process maybe. This, perhaps, is the only way to have a meaningful discussion about art, because it provides a a definition for the discussion and a common vocabulary to conduct the discourse in. But, at least in my own case, which is an admittedly unimportant case, my interest is not in having a cool, beautiful, or otherwise aesthetic-based look, sound, concept, denial of concept, denial of aesthetic, whatever. Any discussion of such may as well be about another piece. I welcome such discussion and find it interesting, but it is more an excercise in strategies of debate than in coming to the heart of the work--which may just be the experience of creation. I would also argue that this applies to other artists than little ol' me: perhaps Merce Cunningham, Kazuo Ohno and Tatsumi Hijikata, others. I may be wrong however.
-Ian