I am working on an ASR model with CTC loss function. I have a popular audio dataset and all its contents are in flac format. My doubt is, why do we need to convert flac file
