HEnquist
How does the quality of the FFT resampler compare to the sync resampler?
How does the quality of the FFT resampler compare to the sync resampler? I'm creating an audio file loading library, and I wonder if I should bother with adding multiple "quality" options when loading an audio file and resampling it to a target sample rate?
If the quality of the FFT resampler is already really good (like if it has a signal to noise ratio below the audible range), then I'd rather keep things simpler.
The FFT resampler gives similar quality as the best the async resampler can do. Both give artifacts far below the audible range. The async resampler is good if you need the asynchronous bit, if not it's just slower.
Thanks! That's what I suspected but I wasn't sure. You could probably mention that in the readme if you want.
I thought I had written something about this, and I have, but not here where people can find it. Instead it's in the readme for my dsp: https://github.com/HEnquist/camilladsp/blob/master/README.md#resampling I'll copy some over that into the rubato docs.