Today’s post is unrelated to anything I’ve been talking about recently. Oddly enough, it starts with a link. For demonstration purposes I recommend that you open that in a new tab (because you’ve got tabs, right? everyone has tabs now!) and leave it running in the background.
Now, it must be admitted that my lack of understanding of music composition theory is only exceeding by my lack of understanding of intelligent algorithms. While I have a passing familiarity with both, I’m not even an amateur with either. However, lack of expertise has never stopped me from thinking weird thoughts before.
I had the above link open in the background for about forty minutes the other day and was struck with the way that it manages to come across as simultaneously chaotic and musical at the same time. Most music, when set to a loop, has very clear breakpoints. You can tell when a run ends and a new one begins, it’s generally obvious. This one, for whatever reason, blends almost completely seamlessly. Further, it also has a sort of weird uniformity such that I’m never sure if I’m hearing a given part of the music again on another loop around, or if it’s just something close in the basic piece.
Which got me thinking: is it possible to design an algorithm which knows enough about musical theory to compose an infinitely extensible piece of background music? One that feels like one unified piece, but which doesn’t actually have to be a loop. I imagine it’d be rather jazz-like, but I don’t really know.
Just one of those crazy thoughts.
Thomas
This is pretty easy to do actually. You should do some wikipedia research into recursive songs and such.
You should do some wikipedia research into recursive songs and such and provide me with a summary. I’m way to lazy to do anything useful you know.
Thomas
No, see, I’ve already DONE the research. My role is to motivate others into following my footsteps. ;p
I’ve got a cyberpunk story somewhere in my brain about a religion formed around a box that continually plays original music using some kind of procedural generation.