[LAU] Resonance

Fons Adriaensen fons at kokkinizita.net
Sun Dec 30 08:42:04 EST 2007


On Thu, Dec 27, 2007 at 05:16:48PM +0000, John Anderson wrote:

> > > The way to simulate this is not to use filters but
> > > delays with feedback.
> > 
> > I had to think about this for a while, but it does of course make
> > perfect sense. Using delays will give not only the fundamental but also
> > all the harmonics as well, just as a mechanical resonance would do :)
> 
> Ah, OK. A delay also implies a comb filter, right? So a multi-tap delay
> would allow you to simulate a space with different sizes in the 3
> different dimensions?

Not really. Better results can be had using a number of N different
delays and an N * N matrix for the feedback. This produces something
'similar' to a real room, bit still not the correct pattern of modes.
To do it correctly would require a 3-D lattice of delays, not at all
simple...

One way to get something close to real room with given dimensions
would be to synthesize the 3D impulse response in Ambisonic format,
using the standard equations to calculate the dominant modes, then
extract the signals you want virtual microphones, and storing the
result in a set of IR to be used with a convolution engine.

-- 
FA

Laboratorio di Acustica ed Elettroacustica
Parma, Italia

Lascia la spina, cogli la rosa.




More information about the Linux-audio-user mailing list