Inverting the phase response of a biquad all-pass filter


#1

Greetings!

I need an algorithm for inverting the phase response of a biquad all-pass filter. Thank you to everyone who can help with this issue.


#2

What do you mean with inverting ?


#3

Run your signal backwards through the same filter - everything else is doomed to fail because it requires exchanging poles/zeros, and one pole will end up outside the unit circle, rendering your inverse filter unstable.


#4

By inverting the phase of an all-pass filter, I mean changing what would be +90 degrees of phase shift to -90 degrees of phase shift. An all-pass filter alters phase according to algorithms that occur in nature. I want to reverse these algorithms that occur in nature, in order to cancel types of phase shift that occur in nature.


#5

Could you elaborate on those algorithms? However, your desired filter sounds quite noncausal to me


#6

My math skills are only in the 63rd percentile for the United States. This means that I know nothing about calculus.

It is my understanding that the peaks in the frequency response of rooms have the same phase shift characteristics as boosted biquad peaking filters.

Consider a two-way loudspeaker system with a 4th order crossover network. There will be 360 degrees of phase shift at the crossover frequency. I am exploring the idea of an all-pass filter that would change this to 0 degrees of phase shift after the correction has been applied.

Sometimes audio software developers tell audio engineers that the technology they are requesting is a bad idea. Am I being that kind of an audio engineer right now?