Has anyone tried to use the multidimensional transforms of FFTW to do multichannel convolution of real audio data? Is there a library around?
Thanks!
In don’t think multi-dimensions transforms will bring any benefits here, as you only need to transform the time-dimension to do convolution of audio data. Or is there a multi-channel transform in fftw speeding up things?
Unless you do some fancy spatial-domain transformations, the second dimension will just be transformed back and forth without any actual processing being done.