auspicious3000/SpeechSplit

mel spectrogram normalization range

inconnu11 opened this issue · 1 comments

Hi, I observed that the range of spectrogram saved in npy file is -0.2 ~ 0.8. I am wondering why you normalize spectrogram into this range? For what reason?

The spectrogram should be between 0 and 1. Anyway, the fast vocoder is released. See README.