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

Support OpenUtau yaml dictionary in diffsinger workflow #167

Open oxygen-dioxide opened 9 months ago

oxygen-dioxide commented 9 months ago

After this change, users will be able to use openutau yaml dictionaries in diffsinger workflow. OpenUtau yaml dictionary contains types definitions of each phoneme. Users can directly use dictionaries distributed in openutau yaml format in diffsinger training, and distribute them with their voicebanks, without converting them from/to openutau format (or writing the dictionary twice, once for training and once for openutau)