Envelope Self Modulation

[baconpaul]

ampeg can't modulate ampeg attack (or filteg attack for that matter) because of ModulationSource.h / canModulateModulators

So what bad thing happens if we lift that restriction? And how do we compensate for that thing?

Raised by slack user 'stamp'

[mkruselj]

@baconpaul I tested out lifting that restriction. Nothing terribly bad happens, but it doesn't seem to work properly either. So if I self-modulate filter EG attack with filter EG (digital mode for env, all stages linear), I don't hear any added curving in the envelope shapes, just the duration of the stage becomes longer or shorter depending on modulation amount.

The intention behind this FR is to be able to change the shape of the envelope curvature, and this is how it was done in Access Virus for example (through self-modulation). However, we don't need to do it like that, at all. The proper solution for this request is making attack/decay/release curve 3-position switches actual continuous controls that would reshape the curvature with more granularity (say, up to 10th power and 1/10th power in the other direction). Of course, this would make things a bit more CPU expensive, since we'd need to resort to using pow() rather than simple multiplication that is done currently, but the benefit is we could have finer-grained curvatures (there's a world of difference between linear phase and quadratic phase * phase in ADSREnvelope.h).

[mkruselj]

@baconpaul I think this is a "won't be done".

After some thinking I'm kinda not a fan of the solution proposed above (due to increase in CPU usage that would be incurred). I'd rather leave this sort of thing to MSEGs. So, fair to close by applying "Won't be Done" label?

[baconpaul]

yeah with mseg and (especially) with formula you can do literally what you want. agree.