[sdiy] Good Expo curve for VCA use..

Scott Nordlund gsn10 at hotmail.com
Mon Oct 15 22:17:06 CEST 2012

>>> But note: the attack portion is an exponential charge curve as in your
>>> Expo1.
>>> This sounds good for sounds that start quickly, but is not useful for
>>> slowly fading in a sound in a constant fading manner.
>> Actually, I tend to think that the "inverted decay" curve (in analog
>> envelopes it's generally an inverted decay up to a preset threshold, at
>> which point it switches to the decay stage) is more desirable than an
>> exponential attack. An exponential attack (the DX7 being a good example of
>> this) tends to sound unnatural, like a sound played in reverse -
>> definitely not "constant fading". On a DX7 you generally need to use two
>> envelope stages to get a usable attack.
> Hey Scott, what you are referring when you say exponential or "inverted
> decay"? to exp1 or exp2 defined in the first post?

I mean exp1. It's still an exponential decay (as opposed to exponential growth),  but it's inverted (increasing) because the capacitor is charging rather than discharging. So it's 1 - e^(-t/T).

It's of course easy to justify an exponential decay, but I think it's a little harder to evaluate "naturalness" of attacks. Attacks in instruments are usually either a sudden onset, or articulated according to however the performer wants to play. But, for instance, the attack of a distant (i.e. no direct signal) pipe organ (the envelope being a unit step) in a highly reverberant cathedral (exponential decay with a long time constant) would be an inverted exponential decay, i.e. exp1. I don't think you generally hear things increasing in dB, anyway. It seems like it would be a good idea, but it sounds weird.

More information about the Synth-diy mailing list