Question regarding 4n butterworth filter - resonance peaks

Hi,

should a bandpass 4nbutterworth filter sound exactly the same as two 4n butter worth filter (hp + lp), if the borders / range is the same? I mean regarding produced resonance peaks?

Why not try it first ?

I could’t get them to sound the same , iow couldn’t do phase cancelation

seems that the bandpass has a bit more gain …then the hp—>serial lp

Here you go …

ps…its a redux file

https://app.box.com/s/kiljirwi5h7ttuf7qjniosj2xqd0sn0s

Ahh I mad a littlemistake the band pass should be set to 8n , because the high and low are 4th order in serial

Still no cancelation …

https://app.box.com/s/b0t5sbrp1u4uzqdltjbniazxphcs4p01

Hm verry funny ,seems that even bandpass filter don’t canceleach other out ( after phase inversion and summing )

Maybe the butterworth

introduces a few samples delay

Strange sometimes they .;sometimes they don’t …Gonna open a topic for this

Forget …this post …( oversampling was different ) stupid me

ah thanks, good idea with cancelation!

I asked this because for me, even if on the display the curves seems to be the same, the bandpass sounds different to me…

Would be interesting if the bandpass is internally just like a lowpass + high pass coded…

I tried to exactly match the frequencies and range as good as possible, also changed oversampling to 0 everywhere… It really seems that bp sounds different…

Attachment 5735 not found.

Hm verry funny ,seems that even bandpass filter don’t canceleach other out ( after phase inversion and summing )

Maybe the butterworth

introduces a few samples delay

Strange sometimes they .;sometimes they don’t …Gonna open a topic for this

Forget …this post …( oversampling was different ) stupid me

So were you able to cancel each chain out?

I tried to exactly match the frequencies and range as good as possible, also changed oversampling to 0 everywhere… It really seems that bp sounds different…

attachicon.gifhere you go 3ju.xrni

If you want to do phase cancelation you have to create a mono sum with one signal inverted , your last post removed the summing .

Anyway ,they don’t phase cancel , but me that ain’t really a problem …maybe the bp res increasing the gain a bit I don’t know

If you want to do phase cancelation you have to create a mono sum with one signal inverted , your last post removed the summing .

Anyway ,they don’t phase cancel , but me that ain’t really a problem …maybe the bp res increasing the gain a bit I don’t know

Hm, isn’t it enough to swap the stereo phase (L+R) of one chain, so they sum anyway even without a explicit summing chain?

EDIT: Yes of course this works. I thin k you overseen the inverser/gainer is now in the first chain and is now stereo.

A h yep , you’re right ., didn’t notice

Hey dblue ,

is it right to assume that the bandpass sounds different in comparison to lp+hp ? Or is it possible to cancel them out? Since to my ears, the lp+hp combination sounds better… (Building some mixing filter doofer, need to make decisions)

Hey dblue ,

is it right to assume that the bandpass sounds different in comparison to lp+hp ? Or is it possible to cancel them out? Since to my ears, the lp+hp combination sounds better… (Building some mixing filter doofer, need to make decisions)

Why dont you decide for yourself ?

I mean , if you think the lp+hp sounds beter then the bp , then use it .

Altough I am also curious why they don’t cancel each other out , that wouldn’t stop me from making decisions based on what I hear .

to my ears, the lp+hp combination sounds better

If it sounds good and works well for you, then simply use the damn thing :slight_smile:

I didn’t design or code the filters myself, nor am I much of a DSP ninja, so I can’t tell you precisely how each type differs at the low level, or how you might go about precisely matching the slope of the LP/HP modes to the Q of the BP mode, or if it’s even possible to do so in a way that they would perfectly cancel out. I somehow suspect that it’s not possible to do so, at least not without precisely setting all of the filter coefficients to the absolute perfect mathematical values, which is clearly impractical (perhaps even impossible) to do at the high level of messing around with the filter device sliders in Renoise.

So really, if one way or the other sounds better to you, then simply use that method and move on to more fun things like actually making music! :slight_smile:

Make it into a doofer and you will have your perfect bandpass filter. :slight_smile:

If it sounds good and works well for you, then simply use the damn thing :slight_smile:

So really, if one way or the other sounds better to you, then simply use that method and move on to more fun things like actually making music! :slight_smile:

ok :slight_smile: You are right.