Re: [LAU] Frequency-selective compressor?

On Fri, Sep 28, 2012 at 03:27:39PM +0200, Florian Paul Schmidt wrote:
> On Friday, September 28, 2012 09:15:21 AM Rick Green wrote:
> > A friend has discovered some historical recordings of his Grandfather in
> > the Smithsonian's collection, and would like us to clean them up a bit for
> > publication. One of the issues is that the triangle player is often off
> > the beat with the rest of the band, and he would like us to push the
> > triangle to the back of the mix.
> > Is there a plugin something like a compressor coupled to a notch or comb
> > filter that I could use to suppress the sound of the triangle within a
> > recording of a Cajun band?
>
> Hi,
>
> sadly it's not sooo simple to supress a single instrument from a mix. A
> triangle is pretty narrowbanded though, so one could maybe get away with a
> narrow notch filter. But the more problematic part is the attack phase of the
> triangle. It is rather wide spectrum (compared to the tail) and sadly the
> attack phase is what defines its rhythm.
>
> Maybe there's research out there though. If you had a model of the triangle
> sound there might be ways of a] detecting where a triangle note is played and
> b] basically subtract the model prediction from the audio signal at that
> point..
>
> Have fun,
>
> Flo
I agree with the above. I'll just add that it depends also a great deal
on what else is in the recording. Perhaps masking or compressing the
triangle wouldn't impede that much on other instruments, depending what
they are: if you've got a fiddle, you're pretty much out of luck.
Anyway, as to your main question, there are plenty of tools, if only you
can identify the frequency: multiband compression and parallel procesing
both come to mind.

Cheers,
S.M.
_______________________________________________
Linux-audio-user mailing list
Linux-audio-user@email-addr-hidden
http://lists.linuxaudio.org/listinfo/linux-audio-user
Received on Fri Sep 28 20:15:02 2012