Fix for #703. Adds an optional sampleRate parameter to the decodeAudioToFloat32() function. The caption transcriber passes in a value of 16000 to prepare audio in the format Whisper expects.
A Fortis Scientia site