@ingox @porres Hi, just stumbled across your Markov remarks. A long time ago I wrote a markov external for pure data (c code, no abstraction) which can handle markov chains of any order (realtime adjustable) and has the option for internal/external feedback changing probabilities and such which is still functional. As it uses binary search trees it is extremely fast and has (somewhat limited) support for lists of integers as elements (like dtime, keynum, velocity or chords).
It is quite flexible as it also allows to define deviations in the values to still being considered as being equal and more.
It's open source and I can dust it off and send it in case you're interested. I'm never on this forum and just registered for letting you know. Don't know if I receive/see any answer from you.
-
MIDI into [seq] and Markov chains
-
@Orm that sounds great. Would be nice to try your external.
-
I just applied recently with my midifilemarkovgenerator which is based on this abstraction for the MIDI innovation awards:
https://www.midi.org/component/zoo/item/midifilemarkovgenerator
Not sure, if it makes sense. But you can vote for it or something else until the 14th of may, if you like it...