So I’ve spent almost a week on this now. I’m still not quite there with the voicecards, there were a few bugs that I fixed (that I had introduced myself while refactoring), but it still doesn’t sound good.
If I play a fixed note, set the oscaillator mix to 0 (all OSC1), and set OSC1 waveform to be a wavetable synthesis algorithm (or some others, but notably not Wavequence or the cz ones), the pitch of OSC1 is incorrect, and actually wavers if I change the selected waveform for OSC2, or the crossmod function…
But I can’t think of anything else I can do except for optimisation to reduce cycles, since compiling with -O2 sounds better than -Os. I think I need a break for the moment, may come back to it in a few weeks.
Anyway my code is up there at https://github.com/machfour/ambika. I’ve been a bit unapologetic with refactoring so I don’t recommend trying to patch my commits on top of anyone else’s forks.