|
:: Friday, April 27, 2007 ::

Musical Analysis
The Machine's Got Rhythm: Computers are learning to understand music and join the band Interesting piece on advances in computer analysis of music [via dev.null]. Not any huge leap really as the same technique has been applied well to speech analysis over the last few years. And in acoustics has been used to design acoustic diffusers too. Ref:
As someone who has studied this field academically (my dissertation for my degree was on Algorithmic Composition and Synthesis) this is an interesting advance. Predictably the Musicians Union are upset about it. They do have a tendency to get the knickers in a twist about things they perceive as endangering musicians job yet never transpire to do so. Historically they have resisted many advancements in electronic music composition and production on grounds of job losses. (Here's one classic example.)
Although, they may have a point in this case as it is somebody's job to transcribe music to piano for publishing as a record of song writing for copyright reasons. However, these versions are often then transcribed back from piano to their original instruments when you buy books of printed music in shops, and are sometimes miles from what was originally written. If an automated process can help to reduce these errors I'm all for it.
What I'm must interested in is what it may contribute to the composition and production process. It's highly likely it will initially serve to further propagate more and more similar pre-packaged music. Like the music machines that churn out tunes for the proles in Orwell's 1984. You may be forgiven for thinking this happens already if you listen to the charts or most commercial radio stations. And of course it will be most profitably used in those music identification systems that listen to bits of music then sale you ringtones etc.
But what if you taught it to think differently? Introduced mutations to see what happens? Or to try to do something that is as far as possible from the norm? It could be made into just as powerful a tool for experimentation as it could for formulisation. That's what I'd want to hear.Labels: Academic, Acoustics, Computing, Electronic Music
:: Dan 27.4.07 [Arc]
[0 comments]

::
...
|
|