91
MP3 - General / Re: Resurrecting/Preserving the Helix MP3 encoder
Last post by jaro1 -I lack the relevant knowledge to look for the error in the right place, but the presence of those segfaults as defined probably also means that the error will not be in the resampling process itself, but "to the left or right of it" (as you described in the edit), e.g. also because the upsampling is working and transparent, maybe..? ..and i understand "corrupted stuttery audio" as something more fatal than just distortion of a sonically non-transparent resampler.
Can code analysis in Lame be helpful here, how are these code "bindings" resolved there in case of downsampling activation?
Maintaining downsampling support seems to make sense only if this problem is fixed. If that won't happen, wouldn't changing the logic under which downsampling has to be used or by disabling downsampling support, be a workaround? But of course this could also have a negative consequence(s).
The article in the HA wikipedia on the Lame encoder, section "Resampling".. in the sense of the last sentence in this section I have a question.. If, after all, there was a bug in the code of the internal resampler, is it theoretically possible to perform a total replacement of internal resampler with e.g. a libretro resampler?