From gstopp@fibermux.com Fri Mar 1 07:45:21 1996 1 Mar 96 10:45:08 +0500 27 Feb 96 15:38:55 +0500 Tue, 27 Feb 1996 14:57:16 -0500 (EST) ID ; Tue, 27 Feb 1996 14:54:27 -0500 (EST) From: gstopp@fibermux.com Date: Tue, 27 Feb 96 11:54:27 PDT Encoding: 85 Text Subject: Re: Envelope Follower Vince from Squishy writes: > >Can anyone suggest a good envelope follower? > Here's my windy $0.02: As is usual with most circuits that are intended to produce some electronic function based on some aspect of a "live" sound, the word "good" is loaded with caveats. In this case, the "live" sound aspect to be captured is amplitude, and the caveats have to do with response time. The traditional description of an envelope follower is a device that has an audio input and a voltage output. The voltage at the output is proportional to the amplitude (volume, loudness, energy content) of the audio signal at the input. Therefore an envelope follower performs the reverse function of an envelope generator paired with a VCA - you give it a sound and it gives you the envelope of that sound. The simplest form of envelope follower is the peak detector, like the one you see in simple AM radio circuits in school textbooks. This consists of a diode and a capacitor. The signal goes through the diode onto the cap, which charges up to the highest voltage that the input waveform gets to, and then stays there. Now as-is this is pretty lame as an envelope follower, because when the signal goes back to zero the cap is still charged up, so the response time is terrible. We need to put a resistor across the cap to discharge it back down to zero when the input goes away. The problem is, as we make this "bleeder" resistor smaller to improve the response time, we increase the ripple that is passed through for low frequency inputs. In other words, if you play a real low note into the thing, like 50 hertz, the supposedly constant DC level output of the envelope follower will actually have a huge honkin' 50 hertz ripple on it. This is no good, we want a steady DC output that is the same as the PEAK of the 50 hertz signal. So in order to fill in the gaps between the 50 hertz waveform peaks, we make the rectifier better and go to a full-wave rectifier. Now the negative transitions of the input signal are not just tossed out the window as in the peak detector (which is a half-wave rectifier), they are inverted and become additional positive peaks. Now we can halve our bleeder resistor value and have acceptible ripple. Most envelope followers are of this design, and work pretty good, but if you tweak the lag time so that the 50 hertz ripple is a hundred millivolts or so (not a big percentage of the total envelope), the output of the envelope follower still takes about a tenth of a second to fall back to zero if the input is going at full blast and then goes away suddenly. This may work for guitars and voice, but for drums it sucks. The next trick is to run about 3 or 4 full-wave rectifiers in a row, with A.C. coupling before every rectifier. I've read that this works a little better, but that there are some strange-looking waveforms and assymetries involved. An idea for an ideal solution I've seen proposed is to use a tapped analog delay line, with a delay time equal to slightly more than the period of the lowest input frequency of interest, with all taps summed together with diodes. With this design, the highest voltage on any of the taps will be the output of the device, and no ripple filter is required. On paper this looks like the perfect envelope follower, with the fastest possible response time. Anybody out there ever done this? The full-wave rectifier designs (single and multiple) are included in the Electronotes Preferred Circuit Collection. Somehow a couple got left out when I made copies for the AH list but now I have the missing pages, if anybody's interested. - Gene gstopp@fibermux.com Subject: Envelope Follower Author: squishy@bga.com (SquishMeister) at ccrelayout Date: 2/18/96 10:36 AM Can anyone suggest a good envelope follower? Thanks. Vince. Squishy Records