music: Beatles- Revolver
This has been going on in geek circles for much longer, probably, but this is the first I’ve heard of it. We now have a computer program that can predict and produce pop songs. I’ve heard rumors of something along the same lines but for classical music (analyze a composer’s body of work and produce a symphony that he could have written) that has fooled the experts, but this program does more than creates music in the style of a certain composer. It claims to be able to take data along twenty dimensions and create original music that they say will be commercially successful.
I suppose things like aesthetic popularity can be modeled mathematically considering it works off principles related to normal distributions. This can sit somewhat comfortably with me; I’d be a little more concerned if there was a computer program that tried to model a more objective evaluation of the arts. Richard Powers tells of the unforunate end to such a project in his book Galatea 2.2.
The notion is intruiging that pop music could be reduced to a formula. Yes, it can be described as ‘formulaic,’ but does the comparison carry that completely? I think that there would need to be a fourth axis to these kinds of calculations, one that would incorporate time. Maybe they have already built something like that into the program. But fads change. The popular music of the early ‘90’s was grunge. Now it is electronica-laced teenaged vocalists. Thirty years ago it was disco. Forty years ago it was the Beatles. Sixty years ago it was big band swing. Who is to guess what it will be in the next couple of years? Were a computer program to extrapolate that, or even claim to do so, it might even dictate the next fad instead of predicting it. Sadly, this is exactly what the suits that produce and distribute pop music are looking for.
Posted by davidtaus at April 9, 2004 12:20 PM