On Fri, Jan 18, 2008 at 02:25:07PM +0100, t_w_(a)freenet.de wrote: Or how about OSC? Just to play devil's advocate, note events are identified by a note number, and though there is a standard mapping of note numbers to frequency, there is nothing preventing anyone from interpreting them any other way. My drumkit performs MIDI IO quite hapily, and it's not even a pitched instrument. Aftertouch events can also refer to previous note on events. At least Roland drum controllers use this to encode events such as choking a cymbal. Ultimately, one must consider MIDI for what it was designed to be: a simple protocol for a keyboard to interface with a synthesiser. This it does quite well. I'm reminded of the story of the king and the toaster: http://www.cs.bgu.ac.il/~omri/Humor/SoftEng.html If we were to replace MIDI with something more general and complex, it's likely I wouldn't be able to build hardware to decode it with a PIC that costs less than a dollar. I think interoperability would suffer, because the functionality offered would be so vast that no device would implement it all. In fact, there are several "standards" for overcoming limitations such as the 7 bit resolution of control values. I believe the reason none of them have seen wide adoption is because there are so many ways it could be done. For example, do you want relative values (increase cutoff frequency by 5 Hz) or absolute values (set cutoff to 1028 Hz)? Different devices have good reasons doing it either way. For some aplications, 14 bits is enough. For others, 4 bits would do. MIDI's simple vocabulary is good enough only 80% of the time, but by being so simple, it also becomes universal, and thus ultimately useful.