What's the attack/release time on the envelope follower?

[guyman]

Hello.

I'm using this envelope follower in a few of my creations (pretty sure it comes stocked by default). I see a coeff calculator if you go inside the module, but I just don't know how to pull a calculation from this. Anyone have an idea of how I can convert the 0-1 on that knob, or some value along the green line to milliseconds? I just want to know the value of the attack and release knob in milliseconds.

Env Follow.fsm

(33.69 KiB) Downloaded 771 times

Love & Peace


~ that guy

[Spogg]

Hi

The scaling is quite complex due to the DSP code, and also note that the response is exponential, so the actual time may be misleading in critical situations.

The schematic shows how to get the “nominal” time readout without any compensation for a -60dB point (which you might see for a reverb tail for example).

I bet there’s somebody here who can explain the maths properly…

Cheers

Spogg

[martinvicanek]

It is the nature of an exponential that it creeps forever towards its limit without actually reaching it. There is, however, a characteristic time when the goal is reached to a certain exent. Unfortunately there is no consensus about what precisely "a certain extent" is. In radioactive decay it is common to consider 50%, in other contexts you will find 1/e = 37%. Engineers seem to like 10% or even 1%, and in reverbs the 60 dB limit corresponds to 0.1% - pick your favorite threshold and stick to it for consistency. The good news is that whatever your choice, the numbers will not depend on it very much.

The envelope follower has difeerent reaction times to peaks and drops, repectively: usually the former is faster than the latter. It may be misleading to use "attack" and "decay" for these time constants, but that's what people got used to anyway.

[Spogg]

Thanks for this Martin.

Way back in the last century, if I wanted to create a defined time period with a Resistor Capacitor delay I calculated for 63%. So 1 megohm and 1 uF would achieve a change of 63% in 1 second. So if I had an “aiming voltage” of 10V, I set the threshold at 6.3 volts and I got about 1 second.
Of course, for accuracy, everything needed trimming due to component tolerances!

Happy days!

Cheers

Spogg