I have run some subjective quality evaluation tests on the internal
resampling done in SDL’s audio subsystem. Since there is no filtering
applied after interpolation, there are a lot of artifacts. Rather than
propose that the internal resamplers be improved, as I understand that
this is a quality/speed tradeoff, I would like to ask whether we can
expose the ability to plug in a custom resampler? I see that this is
possible in the internals, but it is not exposed in the API. Is this
something that could potentially be done without too much trouble?