thanks all for reply,
the best performance I ever heard using musical pattern recognition was by
the trombone player george lewis http://en.wikipedia.org/wiki/George_Lewis_(trombonist)http://en.wikipedia.org/wiki/George_Lewis_%28trombonist%29 he uses a max patch called voyager. I have no idea how it works, maybe you find some documentation online. he worked on it for several years. and it was extremely good. marius.
thanks a lot, in this paperhttp://muse.jhu.edu/demo/leonardo_music_journal/v010/10.1lewis.html,George talk about the voyager programming in the FORMULA language, some interesting insights, thinking buy the cd http://www.amazon.com/gp/redirect.html%3FASIN=B00003JAI9%26tag=squidoox14-20%26lcode=xm2%26cID=2025%26ccmID=165953%26location=/o/ASIN/B00003JAI9%253FSubscriptionId=19BAZMZQFZJ6G2QYGCG2
Have begun to look Nick Collins work (bbcut), that looks to be very good with supercollider.
Looks that legato notes are a problem with fiddle~. Maybe a output mix of fiddle~ and fft~? just guessing.
I think that the length of the pattern to be recognized should be variable, but following the "common sense" concept of motif: 3 to 10 notes.
The motif is a rythmic/melodic pattern - i will think rythm and melody separately, but in some moment they will interact i still don't figure out how.
Aubio semms great, i will try a binary that i found herehttp://ftp.man.poznan.pl/pub/linux/debian/debian/pool/main/a/aubio/pd-aubio_0.3.2-2+b1_i386.deb.
Looks like i have a lot of work to do.
Thanks