onepoll allpass in gen~

LiamCormacGould's icon

I've been messing around with gen~ just to see what it's all about but tbh a lot of it is over my head....

Max Patch
Copy patch and select New From Clipboard in Max.

Anyways I have been trying to implement 4 one-poll, all-pass filters in line. I guess to try and create a phaser effect. Here's what I have...

I have no idea if this is even close but it seems to break over 440 for the freq parameter. My math skills are very very lame so feed me gentle MSP fruit. Also just looking for pointers and not a full blown implementation as I like to work this stuff out a bit/puzzle through and see where I get and what I learn.

Cheers
Cormac

EDIT: I think I meant First order All Pass

LiamCormacGould's icon

ok so no one replied to that post :P but I soldiered on regardless and pretty much have a phaser (I think). Here it is... suggestions are more than welcome from anyone as to improvements to sound features etc.

Max Patch
Copy patch and select New From Clipboard in Max.

Also I am wondering how simple it would be to convert this to a second order all pass chain?

Peter McCulloch's icon

The spacing looks good; I know that for a phaser they shouldn't be evenly spaced, but I think this may be happening because you're not warping the frequency.

Cookbook coefficients for biquad: (also contains the two-pole version)
http://musicweb.ucsd.edu/~tre/biquad.pdf

It looks like there's an exponential operation when you calculate the frequency coefficients which accounts for the discrepancy. (set min and max freq to 3000, then compare to a 3000 Hz sinewave) It sounds fine, but your frequency range is inaccurate. (which may or may not matter)

HTH.

LiamCormacGould's icon

thanks for the response man.

Do you mean that warping the frequency would result in a different outcome i.e. different spacing or that the spacing would behave different over the various frequencies?

As for the frequency discrepancy I would like to repair this so at least I would know the difference but am not sure where to begin. Do I need a wholly different calculation?
I guess I can see that the gain/height of the output sinewave changes with the number of filters involved and also that this is not the same as a sinewave at the same frequency.

This is my first foray into this stuff tbh.

ps thanks for the link.