Fixes bug where random/uninitialzed audio values may appear in some circumstances. Also switches swr_convert_frame to swr_convert for a small optimization/improvement, no need to use AVFrame struct when we're only interested in samples.
Fixes bug where random/uninitialzed audio values may appear in some circumstances. Also switches swr_convert_frame to swr_convert for a small optimization/improvement, no need to use AVFrame struct when we're only interested in samples.