Is it me or the audio reference is broken?
#2
by Vali - opened
I am trying to use it but the output is always horrible no matter the settings. For example when the audio reference used is female the output is somewhat male but doesn't really sound human. So is it broken for new or what?
Sorry! I found a bug. The reference audio was being incorrectly resampled to 44.1kHz when it should have been 16kHz before being fed into the codec. This has now been corrected.
Somewhat related for the breathing, dynamic, and nsfw lines with the reference audio provided it just long "aaaaaa" "oooooo" sounds, instead of (breath) {text}