On Tuesday, December 17, 2013, Fons Adriaensen wrote: > On Tue, Dec 17, 2013 at 10:53:23AM +1100, Alexis Shaw wrote: > > > > 1. That's giant overkill, completely useless. Whatever you need > > > as filtering in a decoder can be done easily with much shorter > > > impulse response. Or even with some simple IIR filters. > > > > It is overkill for anything except room correction where you often need > 1/2 > > second of filtering > > At 386khz that leads to a 128k fir filter > > * 386 kHz ?? Even 192 is completely useless. > > * Half a second of room correction would be useful only at very low > frequencies (unless you want to correct your bathroom). And if and > when it is necesssary, that part of the frequency range can be > processed at a much lower sample rate and require only a short filter. > > * And then, room correction would be done on the speaker signals, > not on the complete matrix. The latter *is* possible of course, > and would be ideal, *iff* you can derive the filters. Which would > require measuring the room response in higher order Ambisonic format, > not on option unless you have a EigenMike, and even then dubious. > > My other project is making an eigenmike type microphone using 1000 mems microphones I am intending to do that....
> > > > 2. The method you propose (DFT, matrix, IDFT) is wrong, you'd need > > > linear convolution which is not the same thing. > > > > > > No you are wrong here, convolusion in. The time domain is equivelent to > > multiplication in the > > Fourier domain. That is simple sampling theory. Look up the overlap and > add > > method of fir filter implementation > > You don't have to tell me that (as the author of zita-convolver). > But DFT, multiply, IDFT without overlap (which is what the OP > described) won't do it. Btw I am the op. > > Ciao, > > -- > FA > > A world of exhaustive, reliable metadata would be an utopia. > It's also a pipe-dream, founded on self-delusion, nerd hubris > and hysterically inflated market opportunities. (Cory Doctorow) > > _______________________________________________ > Sursound mailing list > Sursound@music.vt.edu <javascript:;> > https://mail.music.vt.edu/mailman/listinfo/sursound > -------------- next part -------------- An HTML attachment was scrubbed... URL: <https://mail.music.vt.edu/mailman/private/sursound/attachments/20131217/b800bf3d/attachment.html> _______________________________________________ Sursound mailing list Sursound@music.vt.edu https://mail.music.vt.edu/mailman/listinfo/sursound
