Page 1 of 1

Getting poles out of the Moog filter

PostPosted: Sat Dec 26, 2020 11:33 pm
by adamszabo
Hey guys, I am in a need of a bit of help. I am testing the stock Moog filter thats in FlowStone, and as I understand in a moog filter, the signal goes through several poles before going out of the filter. I would like to see how the filter behaves when it only goes through the first pole, the second and so on. In this schematic I only output pole 3 (red graph), and pole 4 (blue graph). The pole 3 is acting all strange when one moves the cutoff. Anyone know why it doesnt work? Does the code need to be changed in some form? Thank you!

Re: Getting poles out of the Moog filter

PostPosted: Tue Dec 29, 2020 6:05 pm
by martinvicanek
When you say the pole 3 [...] doesn't work, what would you expect instead?

The problem with this implementation is that there is an extra delay in the feedback loop, resulting in bad tuning and instability at high cutoffs. It is quite easily possible to overcome this drawback, refer to https://www.kvraudio.com/forum/viewtopic.php?t=350246

Re: Getting poles out of the Moog filter

PostPosted: Wed Dec 30, 2020 11:36 am
by adamszabo
martinvicanek wrote:When you say the pole 3 [...] doesn't work, what would you expect instead?


Thanks for the link, I will take a look!
What I expected is to simply get the output of the other poles like in your version. In that one it worked as expected.

Re: Getting poles out of the Moog filter

PostPosted: Thu Dec 31, 2020 5:46 pm
by martinvicanek
Ah, the filter in your second post is without an extra delay in the feedback loop (aka zero delay feedback). The loop gain is evaluated in advance and is distributed among the four lowpass stages. That's why you can choose an output terminal at each stage and get a normalized signal. Something similar happens with the second order state variable filter (in zero delay feedback implementation), where you have lowpass, bandpass and highpass output terminals.

You could probably apply a similar normalization to the filter in your first post, however the issues mentioned before remain.