Oscillator syncing

[dpwe]

Some analog synths (e.g. Prophet 5) implement oscillator syncing, where an oscillator waveform (typically a discontinuous one like sawtooth) is reset to phase 0 both by its own period and by zero-phase crossings of a second osc with a different frequency. This can give rich spectral content.

It's tricky to emulate in digital because of the intrinsic look-ahead needed for bandlimited waveforms, but maybe we can do it. I found the discussion in Brandt01 helpful.

[dpwe]

One issue this raises is that it is an instance of one audio-rate oscillator acting as input for a second. We have mod_osc, which marks a second osc as an input, but we don't want to reuse that since sync'd oscs might well want mod inputs too. The other precedent is the FM setup, where algorithm chooses one of the DX7 flow setups among up to 6 oscs. We could stick with this, then add another field (either per-osc or per-setup) that indicates how the audio-rate input controls the downstream oscillator - Yamaha FM (i.e., phase modulation), 'analog' FM, sync, ring modulation? Sync potentially has another parameter (acceptance angle for "soft sync"), but that could be another new field.

[dpwe]

Note that the sound people commonly associate with syncing isn't quite the simple model in Brandt01. Instead, it's most often the rather complex behavior of the original CEM3340 VCO chip - I have some notes and discussion in a doc.