openvpi / DiffSinger

An advanced singing voice synthesis system with high fidelity, expressiveness, controllability and flexibility based on DiffSinger: Singing Voice Synthesis via Shallow Diffusion Mechanism
Apache License 2.0
2.73k stars 288 forks source link

New variance parameter: voicing #170

Closed yqzhishen closed 9 months ago

yqzhishen commented 9 months ago

This PR is based on #169.

This PR adds support for voicing, a new variance parameter.

Voicing is defined as RMS of the harmonic part, in dB presentation, which is similar to breathiness.

With this parameter, the harmonic part and the aperiodic part can be controlled separately. It is also useful to make some whisper-like styles. It should be noted that energy, breathiness and voicing should not be turned on together, because they are coupled with each other.