Background: There's an epidemic of "fake" .flac audio files floating around: lossy audio masquerading as true lossless-quality audio. I rewrote the fakeflac.py utility in Go, adding multithreading to greatly improve performance, and increasing portability by embedding an optional WASM build of ffmpeg.
Limitations: the original fakeflac algorithm is a great starting point, but it fails to find a cutoff point with some (obviously very low quality) audio files and produces a false score of 100%. I added CLI flags to tune the cutoff detection, but future versions definitely need some work (less normalization? entirely new detection algorithm?)