Feeding filters back into themselves turns dsp off?

May 10, 2013 at 1:30pm

Feeding filters back into themselves turns dsp off?

I was playing around today, attempting to recreate a crude version of the common moog filter emulation, when I found that this was impossible because trying to create the resonance was impossible. Funnily enough, when I hover my mouse over the cables I can see that data is still being processed, I just cant hear any of it. I also can’t hear anything else in the patch.

Am I going to have to get deep into some gen~ or faust code writing to play with filters? I have seen the “raw filters” externals kicking around too, but my grasp of how filter works is too minimal for me to be able to make use of this so far. Maybe it’s time to go out and get a copy of The Theory and Technique of Electronic Music

#68172
May 10, 2013 at 1:41pm

Unfortunately it appears that the same problem occurs with the raw filter externals… :(

#244753
May 10, 2013 at 2:04pm

care to share your patch ? sounds like it’s a matter of audio feedback loop, which you can’t do in the maxmsp world, and from what you want to achieve, it seems you should rather dive into gen~ or faustgen~ indeed. That said, there’s a moog filter emulation in faust examples iirc… though if you’re concerned by time and learning curve i’d rather suggest gen than a whole new faust language (which is awesome but relies on quite a different logic than MaxMSP or gen~).

#244754
May 10, 2013 at 2:05pm

Does it work the same way with send~ and receive~ instead of patchcords? Just curious.

#244755
May 10, 2013 at 9:43pm

You can feedback with send~ and receive~ but only because they introduce a one sample delay when used in this configuration

#244756
May 11, 2013 at 12:56am

Yeah, send~ and receive~ work but don’t behave as expected. It doesn’t really resonate very much, and only at certain frequencies. That might be explained by the added 1 sample delay.

#244757
May 11, 2013 at 1:55am

1 vector, actually.

you could try putting the object into a poly~, and setitng the poly to a vectorsize as small as possible without chrashing max.

#244758
May 11, 2013 at 2:24am

Interesting, can you elaborate? Just containing the feedback loop in a poly~?

#244759
May 11, 2013 at 3:14am

yes, and of course, setting the vector size of the runtime to something small (as the vectorsize of the poly patcher is relative to that.)

a vectorsize of 32 is what most audio software uses, and still should work with all hardware drivers, i would recommend that value for building sequencers – and for audio applications containing feedback loops.

if you shrink these 32 for /8 inside the poly you should see quite some difference with your filter, yet not perfection. at the cost of some CPU…

#244760
May 13, 2013 at 10:03pm

Interesting, thanks for your advice!

#249188

You must be logged in to reply to this topic.