I'm working on a vocoder and it sounds like it's important to measure the strength of the modulator in each frequency band properly. The example I started working from used avg~, but today I tried env~ and I think I hear a difference but am not sure. So I was wondering--what's the difference? This test patch shows a greater difference as the input signal gets louder, but what's odd is that avg~ is less than the rms value of env~.

I expected the opposite from reading the help files. Can anyone explain the differences and when one might be more appropriate than the other?
env~ vs avg~.pd

@jameslo the differences are explained in the help files. Maybe what you expected from cyclone/avg~ is "peak to peak" amplitude?
I'd suggest not to use either of the objects and use signal objects instead:

envfollow.pd
the averaging is handled by the lowpass filter. [lop~ 1] averages over 1 second, [lop~ 10] one tenth of a second and so on.

Thanks @solipp, yes I was thinking avg~ measured pk-pk. Am I right to think it's closer to your abs avg example? And except for your method of averaging and env~'s hanning windowing (which I assume means it is examining overlapping windows for each datum output), can I assume that your rms example is equivalent to env~? Also, in the kinds of processors that measure the amplitude of a modulating signal in order to modify a carrier (e.g. vocoders, timbre stampers, spectral noise reduction), do you have any general guidance as to which measurement is most appropriate? Earlier on this forum I asked a similar question about the FFT examples I.04 - I.06, which all seem to use different amplitude measurements for reasons I don't understand.

@jameslo said:

Am I right to think it's closer to your abs avg example?

yes.

can I assume that your rms example is equivalent to env~?

also yes, except that env~ output is in db units.

do you have any general guidance as to which measurement is most appropriate?

not really. Try all of them and see what you like most. The peak method might be too cpu expensive if you have a lot of filter bands.
Katjas blog is a good read: http://www.katjaas.nl/compander/compander.html

Earlier on this forum I asked a similar question about the FFT examples I.04 - I.06, which all seem to use different amplitude measurements for reasons I don't understand.

couldn't find this post. Actually they are all the same...

@solipp RE I.04 - I.06, sorry, that was poorly worded. What I meant was that I.04 appears to operate on the power of each FFT band, whereas I.05 and I.06 operate on the magnitude. Why is that?

I think the assumption underlying my whole line of questioning is that there is theory underlying these choices, and not just taste. Oh, and thanks for referring me to Katjas blog, really good stuff.

@jameslo the I.04 example operates on the magnitude as well. The only difference is that the powers are filtered by the noise-mask before the root.

yes, there is theory. In this case (amplitude measurement) it's plain signal theory. But you can think about it in less technical terms. e.g. the RMS thing is just a signal that is ring modulated by it self. So you end up with 2 frequency bands, one at double the original frequency and one at 0hz! The lowpass filters the high frequency band out and what's left is the 0hz signal or dc-offset or amplitude...
Now there is something like the hilbert~ allpass, which magically shifts the phase of your signal by 90 degrees, roughly speaking. That's very handy because if you multiply these signals by themselves, the phases of the high frequency bands are 180 degrees apart and if you add them together, they cancel each other out and you got the amplitude instantly, without the need for a lowpass filter. (the magnitude calculation in the fft examples works similar)
I think there are no rules if you have a basic understanding of what you're doing in DSP, it's all about artistic decisions and compromises. Do you need instantaneous amplitude detection i.e. fast attacks of your envelope follower in a vocoder, even if it's very cpu expensive? Maybe it sounds cool. But maybe avg~ is just good enough.