Re: [SoX-users] Convolution, channel-selectiveness of effects and spectrograms

[Martin G. <mar...@gm...>]

On Fri, 28 Nov 2025 at 15:36, Peter P. <pet...@fa...> wrote:
> convolution can be achieved by fourier transforming
> both soundfiles into the frequency domain, multiplying the two spectra
> there and transforming the result back into the time domain.

Yes, and what SoX calls its convolution engine does it that way.

> brutefir is one widely available command line program for offline or
> realtime convolution

Thanks, it would be interesting to add that as an alternative to the
one that fir uses. Who knows which will win in speed? fir doesn't
use FFTW yet as far as I know, only boring 2^n-sixed DFTs though
I seem to remember its 2^n one is of a similar speed to FFTW's.

  M