DiffSinger is a voice synthesizer using artificial intelligence to produce the voice from scratch,  mimics the qualities of the voice provider completely

Voicebank counting expressions and recorded time:

– Standard: 2:20:00
– Calm: 0:20:00
– Growl: 0:15:00
– Falsetto: 0:18:00
– Idol: 0:19:00
– Cute: 0:20:00
– Female: 0:15:00
– Screamo: 0:14:00
– Whisper: 0:18:00
– Nasal: 0:13:00
– Solid: 0:23:00
– Mature: 0:20:00