Inverting anomaly?

5 posts / 0 new
Last post
Aleqks's picture
Joined: September 1, 2009

A strange thing I discovered, is that when mixing a signal with the 180 degrees phase shifted version of it (figure 2), the two cancel out as expected... but when doing the same with effects in the chain (figure 1), phase cancellation does not occur... I'm wondering if I have missed something obvious here, if so please do tell as i'm quite stumped.. (I wanted to try a method of modulating a target signal with another by subtracting the difference) Perhaps the effects are not operating identically? I had the exact same parameter settings on each (it was a copy/paste)

Robson's picture
Joined: June 25, 2009

Exact same parameter settings are not enough. Most effects have internal state, like for example the contents of a delay line, or LFO phase, and even filters have tiny delay lines in them. The current state of the effect affects its output. However when you clone an effect its state is not cloned and the copy gets some default state.
Unfortunately AudioMulch doesn't have a command to reset state of all (or selected) effects.

I suppose that when you create your patch anew with audio turned off, and then start the audio, than both effects will start from the same state, and you will get cancellation.

Aleqks's picture
Joined: September 1, 2009

Ahh yes, I had forgotten about the variance in states of the effects; although in my experiments I constantly stopped and started/ turned on an off the audio (so in theory everything initialised together) with no avail: there was still no cancellation. Is there a way to synchronise the states of effects in Audiomulch?

Ross B.
Ross B.'s picture
Joined: April 11, 2009

>>>Is there a way to synchronise the states of effects in Audiomulch?

If you create and connect all of the effects with audio disabled (or load a document with audio disabled) and then enable audio, they should be synchronised (if they have the same parameters. This should work for things like Flanger and Phaser. It won't work for the granulators unless you set their DCor parameter to 0.

-- Ross

Shannon509's picture
Joined: December 19, 2009

Hmm.. Not sure I get the part where the signal goes through an effect but is not inverted..? It has to be inverted, though the waveform may not then be in sync with the one without the effect.

oh, both channels have effects. So the idea is that the effects are just changing the waveform enough that it doesn't cancel? That must be, because even if the effect uses a delay, it shouldn't invert it back to non-inversion (or what would you call it?) Though, the old Sdelay wet signal sounds a little inverted or something. There any kind of inversion going on there?