[sdiy] ADSR Sustain pot law?

[Magnus Danielson]

From: Neil Johnson <nej22 at hermes.cam.ac.uk>
Subject: [sdiy] ADSR Sustain pot law?
Date: Sat, 4 Jan 2003 12:12:15 +0000 (GMT)

> All,

Hi Neil,

> While putting together a list of replacement pots for my "Jenny" project,
> I was looking at the envelope generators and noticed that the two sustain
> pots are both linear.  Now, since in this synth everything is hardwired
> (rather than patchable) it strikes me that it would make better sense to
> replace them with log-law pots, which would lead to a more natural
> behaviour:
> 
> - VCA: sustain sets steady-state volume, so log-law would lead to natural
> "linear" setting (same as volume control);

If your VCA takes linear CV, yes.

> - VCF: sustain sets steady-state cutoff frequency: this is linear w.r.t.
> the control voltage, so for a "musical" control reponse would need a
> log-law curve, producing a 1-2-5-10-20-50-100, not 1-2-3-4-5-6-7,
> behaviour, so that the same rotation angle produces the same octave shift
> irrespective of where on the scale you are.

For a linear (Hz) CV input to the VCF cutoff frequency, you'd be better off
with an expo for the sustain.

There is an issue here with linear VCF cutoff frequencies, when you have a
decay and release phase, which is exponential decay to systain level or 0V,
you will experience a linear pitch change, not an exponential decay in pitch.
This may or may not be what you want.

I think that even an ugly quickhack of an expo-converter might be worth it if
you want to avoid the linear pitch change.

If you feel experiemental, toss in a switch to switch between linear and
exponential pitch change! The reasoning behind that is that well, theory isn't
everything, and you might want to leave an option in there if you feel like it.
There is however a backside to that... for the sustain would would have to
have a pot being both log and lin at the same time, or you need to convert the
sustain level and let that also go under the switch, which however also isn't
THAT hard to do.

However, there are simpler ways to acheive the effect by first assuming an expo
for the VCF and switch between linear and exponential decay and release by
modifying the charging of the ADSR cap. In such a system the sustain pot would
be linear for both cases.

> I think in general, for a fully patchable modular system, a linear sustain
> control makes sense so that, for example, controlling a pitch envelope
> would produce the right behaviour (comment: unless the CV input of the VCF
> or VCA is exponential, then the linear sustain control curve is now
> wrong!)
> 
> Anyway, I'mm going to try out log-law pots for both sustain controls and
> see how it feels.
> 
> Does this make any sense, or am I barking up a tree again?  *woof*

Well... I think you where *possibly* barking up the wrong tree, but for another
reason than you tought...

In the end, try it out and listen to it... does it sound OK with you, then go
for it.

Cheers,
Magnus