Filter problems

#1

Test1.axp (4.8 KB)

So here is a test project. Its a simple lfo, that feeds a vcf. Throught the Vcf i have a osc/saw running. It all ends in a vca (volume control) and then goes to the audio out.
If you load up the project and dont touch anything, you will hear a near closed filter with a sawtooth running through. The filter freq is modulated by a LFO. The filter res is set high (61). Dont touch anything, just listen. Why do i get a rythmic bubbly thing going on on top of the sawtooth? And try and increase the res to max and watch/hear the signal get REALLY distorted. On a normal analog or digital synth, this wont happen. Its like the self osc peak is really loud compared to the rest of the signal.

Can anybody explain to me why this is and what i can do to not have this issue?

This simple patch is a part of a bigger patch, and im just trying to solve the problem.
I already searched the forum, but couldnt find any answers.
Thanks:)

#2

Weird, i get the same issue here. However i'd suggest to put a *c object before the output: the filter does amplify the signal when resonance is high enough, therefore you can get the output to clip pretty easily. A correct gain staging solves a lot of of problems, sometimes.
Also these filters don't seem to work well with fast modulations. The bubbling sound might just be an aliasing artifact.

However i noticed some weirder behaviour: for example in a setup like this Test1.axp (5.7 KB)
If you crank the resonance in the vcf filter up you can get the filters to distort beyond recovery. I wonder what's causing this @johannes

#3

I'm not really surprised to hear this bubbly thing, it's the lfo and oscillator frequencies beating against each other.
Are you sure other synths (hard or soft) give you a stable tone when heavily modulating a filter?

#4

Yup. Seen the same behaviour. And also thought that it could be aliasing:) Edit: This to sputnik

#5

It might just be the raw digital nature/sound of axoloti, but i have never heard this kind of sound in any digital/analog synth before. But i'll test in reaktor and get back.

#6

From a hardware electronics standpoint, filters do get extremely loud when in self oscillation. Also, generally the volume drops as resonance increases, right up to when oscillation starts. When it doesn't its because of some fancy compensation circuitry. Hope this helps

#7

Ok. It might be the "fancy" circuitry thats needed in all this:) I dont get the rythimc bubbles in reaktor. I mean, i do. They just dont sound as "exposive". But reaktors filters might have the res fanciness built right in:)

A nice analog-style filter is definently on my wish list;)

#8

I think many synths will prevent a filter to sweep at 30Hz and 64 semitones of depth.

Yes you get it into unstable territory.

I some situations the explosive character may be preferred over a more tamed filter. The difference could also come from the sawtooth you're feeding in.

I invite contributions!
Some literature:
https://www.native-instruments.com/fileadmin/ni_media/downloads/pdf/VAFilterDesign_1.1.1.pdf

#9

Hehe. I couldn't program a single thing even if my life depended on it. But I guess I could do some reading about it:)

#10

Yepp, it starts to self-oscillate. However the problem i get is that the filters lp1m after the vcf3 start distorting and they don't stop even if you take the resonance down again. I think it's a problem of the lp1m inlet not being saturated, therefore causing this trouble