in theory i guess. i dunno the theory behind synthesizing stringish sounds like that though. i quite liked it.

while i was at work today i was thinking about melody generation algorithms and i figured that one way you can see melodic variation is as permutations of note sequences. as an experiment, i entered the johnson-trotter algorithm for enumerating permutations of integer sequences of length 2, 3, and 4 where 1 = C, 2 = D#, 3 = F, 4 = F# into renoise by hand. i played the same sequence deeper and slower to make chords. what i got:

http://tindeck.com/listen/hiwi (with sampled rhodes and stock reverb) I think it would be better if i built a permutation generator in puredata since it was a pain to copy out all the data by hand. it's also obviously fairly boring and has virtually no rhythmic content.

i think that some interesting music generation stuff could be done with just subsets, permutations (

http://en.wikipedia.org/wiki/Permutations for reference) and partitions (

http://en.wikipedia.org/wiki/Partition_%28number_theory%29 ) where the partitions divide up time for rhythm, the subsets decide scales and chords, and the permutations represent melodic variation. If i get the time , I might build some machinery for this in pd. Since it can send midi control out, it can be used independently of the synthesis component. I'm not really sure how to program combinatorial objects in puredata though.