Retrieval-based-Voice-Conversion-WebUI
Retrieval-based-Voice-Conversion-WebUI copied to clipboard
Questions about RVC
Hello there! I have a few questions!
- I saw somewhere in the code that the HuBERT model is used with a sampling rate of 16kHz!!!!! Does this mean it is downsampled and then upsampled back? Is it possible to train 48kHz HuBERT from scratch to prevent this?
- How does RVC work? Literally, there's zero information and papers about it. I don't want to inspect the code! I would like a nice, simple, and short yet full explanation.
- Are there other good new voice-to-voice models (OSS only)? Sing with RVC is crap. I have studio-quality recording but it is still bad (thank you, HuBERT)!
- Stereo audio when?
Thank you!