delay envelope follower

[gstopp at fibermux.com]

     I was thinking the same kind of thing - if your sample rate is in the 
     audio range, regardless of tap spacing, there will be notches in the 
     frequency response of the derived amplitude voltage. Yuk.
     
     That led me to the idea that you don't need audio-range response on an 
     envelope follower, maybe roll it off at a couple kHz and sample at 
     twice that. Maybe okay for an acoustic instrument, but not ideal.
     
     That led me to the idea of splitting the audio into frequency bands 
     like a vocoder, then tweaking each band for ideal response time, then 
     taking the highest output as the follower's output by diode summing.
     
     That led me to the idea of making the integrator time constant on a 
     normal rectifier-type envelope follower adaptive to the frequency of 
     the input signal, with some kind of crude period recovery function.
     
     That led me to the idea of using some kind of sample & hold, that 
     somehow knows to sample the top of each peak of the input signal.
     
     One of these ideas is probably the better than the others, but I'm not 
     sure which.
     
     - Gene
     gstopp at fibermux.com


______________________________ Reply Separator _________________________________
Subject: Re: delay envelope follower
Author:  Eli Brandt <eli at UX3.SP.CS.CMU.EDU> at ccrelayout
Date:    2/19/96 10:48 AM


Juergen Haible wrote:
> Clock rate -> highest frequency detected 
> Length -> lowest frequency detected
> Number of tabs -> ripple of VC
     
If I'm understanding the design, it seems like you'd want to have the 
taps irregularly spaced.  Otherwise a signal whose frequency is close 
to a multiple of the spacing will alias, causing a "throbbing" 
artifact in the output.