VogenSVS

An SVS (Singing Voice Synthesis) toolset. Some parts are still under construction.

Completion status:

• [x] (Vocoder) VocV1(VogenVoc) [Demo Code] [Demo Page]
  • Breathiness Generator Model [Train Code]
• [ ] (Acoustic Model)
• [ ] (Duration Model)

References

• R. Yang, Y. Peng and X. Hu, "A Fast High-Fidelity Source-Filter Vocoder With Lightweight Neural Modules," in IEEE/ACM Transactions on Audio, Speech, and Language Processing, vol. 31, pp. 3362-3373, 2023, doi: 10.1109/TASLP.2023.3321191.