Two Pole LP before Clipping

[petemoore]

  The popularity of two pole filters seems to migrate to the end of circuits.
 Most often I see them used as a cab sim, at or near the output.
 I wonder if the reason for this is more related to the sound [popularity] or just that they're a little less explored as an interstage option.
 Are there any circuits or discussions of introducing the sharper rolloff the 2-P filters provide earlier on in an effect such as before a clipping stage ?
 

[Cliff Schecht]

I guess it just depends on what you are trying to do. Typically you add distortion to a signal, clipping or otherwise, to add to the spectral content of a signal. With filtering, you are trying to remove or modify certain frequencies while not affecting others. Distorting and then filtering gives you the ability to add harmonics and then remove what you don't want, which makes the filtering requirements less stringent. Filtering before your distortion can remove stuff that you don't want emphasized, but if you are trying to prevent a certain frequency from getting boosted then you are going to need something more than a passive 1st order filter. This means that you need some sort of active filter to remove the frequency you're trying to kill and you'll probably have to boost the signal after filtering as well.

If you look at it another way, it usually makes more sense to add gain before you start taking away from the signal. I guess Friis formula could be applied here as well. This formula shows that the first stage in a system will contribute the most noise in a system and so the gains should be cascaded appropriately. Making the first stage a filtering stage (reducing the signal) and then boosting/distorting this filtered signal would add more noise than doing things the other way.

[Mark Hammer]

I look at distortion in terms of additive and subtractive synthesis.  Generating harmonics is the additiive part.  Tone-shaping prior to clip, or post-clip, is the subtractive part.  As a synth person, Cliff, you would understand that applying the identical 4-pole lowpass to a triangle or ramp wave would frequently nail you a different tone than applying it to a square wave.  So, there are cogent reasons for wanting to use filtering pre-clip, rather than making post-clip filtering do the heavy lifting.

I'm a big fan of multi-stage clipping, but an equally big fan of using pre-clip filtering in a deliberate tactical manner beforehand such that the harmonic content elicitted by multi-stage clipping does not result in undesirable fizz.  Does one need a deliberate overt 2-pole lowpass?  Not necessarily.  There may be many places where filtering is inserted that you don't think of.  For example, the EPFM version of the Anderton Tube Sound Fuzz (which transmuted into the Way Huge Red Llama) omitted the op-amp front end that Anderton had used in an earlier version of the circuit published in Guitar Player.  I prefer the latter rather than the former.  Why?  Because the op-amp stage lets me stick a feedback cap in that rolls off the top end before the signal hits the first CMOS invertor stage.  Then, i the first invertor stage, I use a second feedback cap to roll off the top end there as well.  Voila, two poles of lowpass filtering, without an obvious explicit 2-pole filter.