Mikelib

I have copied a model of a tin can from a pd patch and have noticed that there is some sound residue on the filters after the impulse finishes. This residue appears on the pd patch as well. Can anyone explain why this happens? How can I stop it.

Hi there,
        I have copied a model of a tin can from a pd patch and have noticed that there is some sound residue on the filters after the impulse finishes. This residue appears on the pd patch as well. Can anyone explain why this happens? How can I stop it.

filter-residue

I don't hear anything after the initial resonance in your patch (bad computer speakers though) but if you hook up a number~ object it shows it fluctuating between -0 and 0 which *might* be what your talking about.

I changed the patch to use reson~ instead of svf~ and it doesn't have the same problem. Maybe just use reson~ instead, as it sounds nearly the same as svf~ (to me on my bad comp speakers anyway). Also, from reading the reference files of both, it seems to me that reson~ is far less computationally complex, so is likely to be less cpu intensive - and is actually designed as a resonant bandpass filter.

Hopefully someone with more svf~ experience will chime in though!

I changed the patch to use reson~ instead of svf~ and it doesn't have the same problem.  Maybe just use reson~ instead, as it sounds nearly the same as svf~ (to me on my bad comp speakers anyway).  Also, from reading the reference files of both, it seems to me that reson~ is far less computationally complex, so is likely to be less cpu intensive - and is actually designed as a resonant bandpass filter.

also - cranking your headphones up to crazy levels could be seriously damaging to your hearing if you accidentally triggered some sound!

You could sort the problem with svf~ by using a dc offset filter. There is one in the percolate library, but I tried it with no luck. However, the one from tap.tools does work, so I'm not sure how big the differences between them are.

There was a post on this not long ago which suggested using [biquad~ 1. -1. 0. -0.9997 0.] Biquad is a two-pole filter, and I've read that dc blocking filters are usually one-pole, so it may not work quite as well.

I would suggest switching to use reson~ for this patch. It seems like less hassle!

You could sort the problem with svf~ by using a dc offset filter.  There is one in the percolate library, but I tried it with no luck.  However, the one from tap.tools does work, so I'm not sure how big the differences between them are.

There was a post on this not long ago which suggested using [biquad~ 1. -1. 0. -0.9997 0.]  Biquad is a two-pole filter, and I've read that dc blocking filters are usually one-pole, so it may not work quite as well.

I would suggest switching to use reson~ for this patch.  It seems like less hassle!


Filter residue