Forums > MaxMSP

dynamic bandstop filter for guitar

December 14, 2008 | 9:38 pm

Hi there,

we are looking to implement a guitar effect which is meant to filter out the main fundamental frequency of a guitar sound, trying to only leave the partials.

When using fiddle to determine the fundamental frequency (highest peak) and applying filtergraph and biquad (with bandstop and Q=0), the resulting sound isn’t too much different. I realized that other filters-plugins (such as apEQ) provide a bigger bandwith limiting the fundamentals more heavily.

How could this be done ? Now that I think about it, a highpass with a offsetted (e.g. fundamental+100) center frequency would probably better suited.

any other ideas how to boost the timbre of the guitar sound ?

cheers, Nick


December 14, 2008 | 10:03 pm

Yeah, you’d be surprised at how little of any sound is the
fundamental. Check out the program SPEAR (freeware, Google it) and
you can play around with FFT fun. It’s a good way to experiment with
what happens when you add/subtract/amplify partials, as it shows them
all as sine tones. Once you find something you like, then you can try
to get the same effect (live) in Max, but the visual thing has helped
me a ton in moving around partials.

Also, you can make spectral remixes of anything.

On Sun, Dec 14, 2008 at 1:38 PM, Nick Laqua wrote:
>
> Hi there,
>
> we are looking to implement a guitar effect which is meant to filter out the main fundamental frequency of a guitar sound, trying to only leave the partials.
>
> When using fiddle to determine the fundamental frequency (highest peak) and applying filtergraph and biquad (with bandstop and Q=0), the resulting sound isn’t too much different. I realized that other filters-plugins (such as apEQ) provide a bigger bandwith limiting the fundamentals more heavily.
>
> How could this be done ? Now that I think about it, a highpass with a offsetted (e.g. fundamental+100) center frequency would probably better suited.
>
> any other ideas how to boost the timbre of the guitar sound ?
>
> cheers, Nick
>


December 15, 2008 | 3:30 pm

hello nick,

i am not sure, that searching for the highest peak is a reliable way of finding the fundamental.

however, when it is found already, then i’d recommend using reson~ to boost this frequency with a very high Q, invert the phase of the result and mix it with the original audio. this should cancel any found fundamental or lower range partial quite effectively.

your idea sounds very nice in any case. if you feel you’d like to share this approach, then i’d be happy to take a deeper look into it (i’m on 4.63 ..)

jrp


December 15, 2008 | 4:27 pm

"applying filtergraph and biquad (with bandstop and Q=0)"

Just to double-check — is this a typo? To filter anything out, you’ll want Q > 0. You’ll get fairly tight filtering with Q ca. 10, or an extremely tight notch filter with Q near 100.

Cheers,
EH


Viewing 4 posts - 1 through 4 (of 4 total)