Cherry Audio Forums

ColinP wrote: ↑Sat May 25, 2024 3:54 pm I've found the best way to think about what is a surprisingly complex area is to use finite-state machine diagrams.

I wrote this post a few years ago. I was getting exasperated with soft-synth developers who didn't understand how analog monosynth envelopes worked.

Monosynth Envelopes and Oscillators (and how to do them right)

Take a look at the SSM2055 envelope generator chip. My latest EGs are simply software versions of the diagram shown on the spec sheet. The whole EG is much less than one screen of code and is very efficient.

https://www.synfo.nl/datasheets/SSM2055-emu.pdf

Grant, your FSM diagrams are a lot neater than mine!

My slight frustration with EGs is that they are mostly designed with amplitude control in mind. ADSR is a good general purpose model but there are applications where one needs more unusual behaviour such as return to zero on retrigger, instant transitions, attack commit and non-zero termination.

The stock Super Envelope Generator is pretty flexible but still lacks features.

I'm also interested in better expressive interaction between EG shapes, velocity and aftertouch. Also modifications such as from damper, una corda and sostenuto pedals. A real pity I think is that key release velocity has been so neglected by synth designers as it's dirt cheap to implement and being able to control damping (and other things) by how fast one releases a key would add an extra dimension to expressivity.

My particular problem space at the moment is not in creating complex envelope shapes but in creating relatively simple but unusual ones. I'm working with a setup where every control is backed by a 16-step voltage controlled sequencer with cubic spline interpolation arranged in a circular pattern so there's through-zero C0, C1 and C2 continuity. Generating useful envelope control voltages to drive the interpolator is the goal.

If I drive the interpolator with a linear ramp we get the kind of curves shown below directly...


These fancy curves are easy to create as one can do real-time or slow-mo motion recording and non-real-time editing while seeing and hearing the effects. Plus there are randomization and smoothing facilities but what I'm trying to achieve is a relatively straightforward mechanism for bending that ramp so that for instance more detail is available in the attack phase, it's posssible to sustain and to add controlled amounts of aftertouch modulation in the sustain and release phases. Apart from the aftertouch potentially reducing the CV the general trend is rising - another thing that conventional EGs are not good at.

Being able to loop in the sustain phase would be a bonus as would a dual attack facility so that low velocities can produce slow swells while faster velocities take a more percussive path.

So my FSMs are still in a high degree of flux as I attempt to map a path through this minefield.

This is all only in the monophonic domain at the moment but the ideas could be extended.

MRBarton wrote: ↑Sun May 26, 2024 12:47 am Take a look at the SSM2055 envelope generator chip. My latest EGs are simply software versions of the diagram shown on the spec sheet. The whole EG is much less than one screen of code and is very efficient.

https://www.synfo.nl/datasheets/SSM2055-emu.pdf

I have mixed feelings about the CEM and SMM chips. In a way they were just too well designed. It's arguable that they caused innovation to stall for a while as so many synth designers used the off-the-shelf solutions they provided rather than pursuing their own potentially more characterful approaches.

I realise that many people will disagree strongly and will argue that the result was some kind of golden age but to my ears it was a nadir that we are only just beginning to recover from. Not because there was anything intrinsically wrong with these synths but because I think other avenues went unexplored for a few decades.

ColinP wrote: ↑Sun May 26, 2024 5:17 pm Grant, your FSM diagrams are a lot neater than mine!

You should see my VM development notebook! My handwriting and freeform drawing have always looked like a spider on crystal meth. If I need to communicate anything to someone else, I depend upon word processors, text editors, drawing packages, etc. I drew those diagrams in DIA, an open-source drawing package. Not as user-friendly as Visio, but good enough for my purposes.

My slight frustration with EGs is that they are mostly designed with amplitude control in mind. ADSR is a good general purpose model but there are applications where one needs more unusual behaviour such as return to zero on retrigger, instant transitions, attack commit and non-zero termination.

I asked Andrew Macaulay if he would consider making RTZ/NRTZ a switchable option on his Trapezoid and VC Envelope modules. The clever solution he came up with ("Retrigger Damping") exceeded my expectations.

utdgrant wrote: ↑Mon May 27, 2024 8:09 am You should see my VM development notebook! My handwriting and freeform drawing have always looked like a spider on crystal meth.

Take at look at spiders on caffeine...

https://en.wikipedia.org/wiki/Effect_of ... on_animals

utdgrant wrote: ↑Mon May 27, 2024 8:09 am If I need to communicate anything to someone else, I depend upon word processors, text editors, drawing packages, etc. I drew those diagrams in DIA, an open-source drawing package. Not as user-friendly as Visio, but good enough for my purposes.

I use draw.io which I suspect is very similar.

Andrew's modules are very comprehensive but even the https://store.cherryaudio.com/modules/v ... d-envelope can't produce the envelopes I'm after.

Having badly hijacked this thread already, I might return later with more envelope related scribbles...

The secret life of spiders! That explains the webs in my garden. An I just thought it was the uneven heights of the plants and flowers. This year I am finding yellow fennel that should be green or bronze. No idea if the spiders had anything to do with that!