Standalone Envelope Detector for Guitar Signal?

[Mr. Lime]

I'm looking for a circuit like the Copilot FX Autodialer pedal.

A simple envelope detector to connect with the CV inputs of other pedals.
Usually the CV inputs asks for a 0-5V range.


Does anybody have good experience with a certain circuit?


Zorg FX has a documentaion of their envelope filter which describes how the trigger works.
https://www.zorgeffects.com/images/Manuels/Love_Philter_Technical_Data_Eng.pdf

Looks like it's a good idea using a full wave rectifier to reduce ripple. I like the idea of having an extra input for an expression pedal but a simple circuit that tracks the guitar signal well would be sufficient.
Sensitivity, Attack and Decay controls might be useful but also a limiter (diode at output with 5V treshold?)to protect the other pedals.
As the Zorg effect is driving a vactrol, it's probably not optimised for the direct voltage feed.

Looking forward to suggestions! 

[StephenGiles]

An idea I had some years ago (possibly more like an obsession!!) was to use the EH Space Drum front end to drive a VCF, the reason being that only the initial peak of the CV is input level dependent, and as a result the CV is almost entirely ripple free. I never got it to work as I wanted unfortunately - thoughts?

[Mark Hammer]

What makes a "good" envelope follower depends on what you're using it to drive.  If the control element is a little sluggish, as can be true of LDRs, half-wave rectifiers can do a decent-enough job.  If the control element is fast and responsive, then full-wave rectification, and maybe even double rectification, is needed to eliminate audible envelope.

[ElectricDruid]

I know that some highly-regarded compression circuits use that "double compression" technique, but I've never understood conceptually how it's supposed to help. It feels a bit like "well, it worked the first time, so it should work better if we do it again, right?". But of course, that's not really true.
Full-wave rectification is better than half-wave because if the envelope of a signal is loud, then the negative peaks are likely to be loud as well as the positive peaks. Flipping negative peaks up then gives a signal which is easier to smooth and still provides a good overall conceptualisation of the envelope shape (of course, depending on the waveform, they may not be the same level, but it's still bound to be better than nothing).
Once you've done that, I don't see where the benefit is supposed to come from. Any further rectification is not taking *peaks* of the waveform, but rather flipping up what used to be the *zero crossings* of the waveform. That doesn't seem to me to add anything o the envelope information (since every waveform has zero crossings the same) so I don't understand it.
I'd be pleased to be educated as to why it's actually really useful, if anyone's got any ideas. Thanks.

Tom

[Mark Hammer]

I've got a whole folder of envelope-follower/rectifier schematics.  Many of them apply some lowpass filtering to the rectified result, in order to provide some more smoothing.  But of course, that raises the matter of "How fast and responsive do you need it to be?".  One may need something very fast for a limiter, but can tolerate something slower for a noise gate, or in the case of a number of bucket-brigade circuits (CE-1, A/DA Flanger) fading out wet path clock noise when one stops playing.

That said, it's not an impossible task to provide several parallel outputs of a basic FWR circuit with different degrees and rolloffs of lowpass filtering to suit different objectives.

[PRR]

Full-wave rectification is better than half-wave because if the envelope of a signal is loud, then the negative peaks are likely to be loud as well as the positive peaks.

Mostly + and - are same size; only advantage is reduced ripple. But several sounds, particularly male vocal sounds, are asymmetric enough to matter. Enough that AM transmitters (AM supports asymmetric modulation) have routinely supported phase-flip, as well as full-wave detection in limiters.

[Mr. Lime]

What makes a "good" envelope follower depends on what you're using it to drive.  If the control element is a little sluggish, as can be true of LDRs, half-wave rectifiers can do a decent-enough job.  If the control element is fast and responsive, then full-wave rectification, and maybe even double rectification, is needed to eliminate audible envelope.

I'd like to feed the detector with electric guitars, maybe even basses.
Thanks for the hint with double rectification. A circuit that looks promising I think is this (ditching the last compensation stage):






Looks like the Fullwave x2 output is quite smooth compared to the single retified output.


What efeects should be driven with it? Mainly I think of my Behringer (Mutron clone) Dual Phaser, my Threeway chorus or filter/wah wah pedals.

I also thought about feeding Tom's StOMPLFO, similar like my Drolo Moon Rabbit phaser does.
This would probably be the most advanced approach since the envelope detector could be used to control the frequency or the offset of the STOMPLFO.

Looking at the circuit, I think I would have to add an input buffer, a sensitivity pot and a gain stage to take use of the 5V headroom. I guess ~850mV is the maximum output swing (peak) of a typical bass/guitar pickup, but the average should be around 50-80mV.

So how would you guys calculate the needed amplification of the gain stage?
Take a minimum signal of 10mV and a stage with a gain of 500x?
Or is a gain of 200x sufficient?
Where would you place a diode to cut the high peaks?

[Mr. Lime]

What do you guys think about the Doepfer A 119 module?
https://doepfer.de/a119.htm

Would it suit the requirements?
I can't find information about the voltage output range but this thing would be a cheap alternative. The PSU seems wierd for the pedal board..

[Mark Hammer]

Harry Bissell's envelope follower, previously brought up on the forum several times, is pretty dang good (though apparently not perfect).

Here's a nice discussion.  https://lookmumnocomputer.discourse.group/t/bissell-envelope-follower/4666/3

Some 22 years back Harry came up from the Detroit area, to Ottawa, for a little DIY synth summit in my basement, prompted by Osamu Hoshuyama's year in residency here as visiting scholar at one of the local universities.  Among the many delights Harry brought was his self-designed guitar processor.  I call it "processor", rather than synth, because it used a hex pickup, but processed fuzzed versions of the individual strings, individually, with VCFs and VCAs for individual strings.  Fantastically responsive, with no audible lag, and precious little ripple.  More involved than most, but well worth it.

[Mr. Lime]

In the Modwiggler forum a guy compared the Doepfer with the Bissel circuit and it looks like the Doepfer offers the smoother envelope:
Link

i've done a quick comparison between the Bissell and a modded Doepfer A-119 (lowered the 100nF filter cap to 47nF).

source material is a TR-606 snare drum, as seen on the first image.

the second picture shows both the Bissel (top) and A-119 (bottom) envelopes. Bissel is nice and snappy while the Doepfer looks more like a camel's hump... keep in mind that an unmodded A-119 would smooth out even more of the attack portion!

[ElectricDruid]

That looks to me like the Doepfer module just uses much heavier filtering on the envelope. So yes, it'll be smoother, but you also lose more detail and have much more lag.

[Mark Hammer]

Bear in mind that Harry's search for an optimal envelope extractor was principally for use with guitar. Doepfer's would be a more general-purpose circuit.  And while it is always possible to strap on a slew limiter to "smooth out" an envelope, an unsmoothed envelope can only be as fast as it can be.

I really have to build one of those.

[Mr. Lime]

Some comments from Harry himself can be found in this thread: Making a generic envelope follower

Looks like PCB, Kit or preassembled Eurorack can be purchased from Australia here:
Elby Designs

A open source project with SMD parts:
http://wiki.midibox.org/doku.php?id=envelope-follower-bissel-based