Why does a MP3 encoder use a fast Fourier transform before applying the psychoacoustic model?8yr ⋅ phonolog-51448 ⋅ se/cs
