A model of itself

Yesterday’s post was a simple example of a tool to create procedural music. Of course there is much more one could do with this — such as the rich array of work by Toshio Iwai and many others.

But if your goal is to enable crowd-sourced procedural music, where a community collaborates on-line to build a kind of interactive musical world, then things need to be especially intuitive, and people who wish to contribute need to be able to understand quickly and easily the things that were made by other people.

This suggests that in some sense the thing needs to be a model of itself. You have to be able to look at it and say, right away, “oh, I see, this is the structure.”

Which means that the tool set needs to be very carefully chosen. A repetition should look like a repetition, a transposition should look like a transposition, and a theme and variations should look like a theme and variations– immediately, without any need for head scratching.

This is going to require something much more sophisticated than the little toy I posted yesterday. I suspect it will involve zooming in and out — so you can choose to tweak little details or to zoom out and rearrange the entire structure.

I’m not sure what such a community-built musical world construction kit will look like. But I’m pretty sure that when it finally starts to work properly, it will be great fun to play with!

Leave a Reply