there is much progress. the engine remains in python (with a small C++ loop injection, excuse my french i am not a programmer)

but now, it has weird esoteric sample analysis based on genetics(?)

the formulas coming out of this analysis are in no way resembling the sources.

but they are pretty interesting:

they contain filtering, modulation, formant movement, wild envelopes.

i am thinking of putting this online open source it and call for beta testers and maybe collaboration.

the GUI is not ready yet, but the audio engine has come a long way.

on a mac mini m4 with my digitakt ii + overbridge i achieve 64 polyphony easy. with minimum latency.

tell me how we could do this. i don't like github. i am thinking of hosting the thing my self.