Midimutant update

Posted Nov. 24, 2018 by Dave Griffiths

An attempt at a rough explanation of what's going on with midimutant, and future plans etc. The executive summary is: no release yet, but the project is restarted with release aimed early next year.

Since Richard mentioned the midimutant project in an interview in 2017, and a subsequent publication in Raspberry Pi magazine, obviously there has been quite a lot of interest in whats happening with it. This is a bit of a side project, and more research and development type of thing than a product or commission with a deadline (so has to make way for our other work - which has been slightly over plentiful this year!).

There is a natural flow to this kinds of project, you kind of need to fiddle with it, hold it close to your chest for a bit to figure out the right path, or what it even actually is - before making it public, this is true even if the aim is to eventually open source it.

Reverse engineering Kyma

Tangent 1: Building an inhumanely complex FM synth designed for evolutionary strategies.

Tangent 2: Building a hardware synthesiser designed for evolutionary strategies.

This project in general led to an fanatical intensification of an existing interest in FM sound, and a fairly lengthy tangent into using midimutant on YM sound chips as well as the DX7. With the DX7 the limitiation (and fun really) comes from being restricted to the parameters in the sysex format and the design of the synth. The YM chips were a later budget development by Yamaha and were full of shortcuts and oddities. What would happen if we added hardware parameters for example overall clock speed to a genetic algorithm?