Question about Envelope Followers

[william]

When building an envelope follower for a gate, how does the envelope filter produce the gate.  It seems to me the rectified signal in the envelope filter should follow the amplitude of the source signal pretty closely.  Yet a gate has a point where it opens, and the full signal is allowed to pass.

I ask this because I am working on my own Gate design, but looking at the envelope followers for the MXR noise gate, the Boss NS-1 and NS-2 and a few synth VCA'a I'm unsure how to achive this.

Can someone maybe shed some light on this for me?

[Boofhead]

You are pretty close.

It seems to me the rectified signal in the envelope filter should follow the amplitude of the source signal pretty closely.

Yes that's usually true.

Yet a gate has a point where it opens, and the full signal is allowed to pass.

The rectified signal feeds into something which has a "jump" in the attenuation vs rectified signal level.  That point defines the threshold for the gate.

[Mark Hammer]

An envelope follower yield a fluctuating unipolar voltage (i.e., it only goes positive OR negative), where a gate is a constant-state voltage.  Your question is essentially how do you derive one from the other.

Set a threshold.  When the envelope voltage exceeds that threshold, the gate-voltage stays high, and when the envelope falls below that threshold, the gate voltage goes low.

Of course you can probably see a few issues with that immediately.

One is that the threshold has to be set very carefully, and the envelope signal provided needs to always be above the threshold when it should be and below the threshold when it should be.

Even when both the threshold and envelope sensitivity are set sublimely, the reaction time of the envelope follower results in some delay time for gate onset unless you do some tricks (see below).  As well, you tend to lose some of the note decay because, truth be told, parts of the trailing end of a note aren't a whole lot louder than the finger glisses you're hoping won't trigger a gate.  For those reasons, and a bunch of others, traditional analog synth-type processing of guitar signals was problematic.

One of the ways that can get around at least part of the gate-onset issue is by clever use of time delay.  Indeed, the E-H Attack-Delay used a BBD to delay the audio signal a bit so that it had enough time to accurately detect the onset of the note and THEN impose a slower attack envelope onto it.  Some of the older click-and-pop eliminators in the vinyl era also used a small amount of delay to permit higher-accuracy differentiation between signal and surface noise prior to gating out clicks and pops.

If you punch in "envelope follower" and "schematic" into Google, you'll likely be led to a bunch of analog modular synth circuits that nicely demonstrate the envelope-to-gate conversion process.  The older PAiA stuff that both Jack Orman and PAiA have posted shows such a circuit.