RVC-Project / RVC-Project/Retrieval-based-Voice-Conversion-WebUI
Weights / Voice Fusion : Too loud / Distortion in some cases
Nobody has claimed this yet.
- Dominant language
- Python
- Stars
- 38.4k
- Forks
- 5.3k
- PR merge metrics
- No merged PRs in 30d
Description
I trained 2 different datasets of the same voice, there is a bit of overlap between the 2 datasets. When I fusion the 2 weights at 0.5 for A and 0.5 for B, or even 0.7 for A 0.3 B, the new weight inference output is too loud and some of the loudest parts are distorted. I noticed this issue also when fusion 2 different voices. Changing the volume settings for inference doesn't change anything. It would be great if we could manually adjust the volume on the Ckpt processing for Voice fusion.
Contributor guide
No contributing guide indexed for this repository
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Start by locating the checkpoint or weight-fusion processing path and the inference volume handling. Reproduce the issue by fusing two overlapping voice datasets or two voices at the reported weights, then verify that manual volume adjustment changes the fused output without loud-part distortion.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python
- Domain
- audio-video-rtc, machine-learning
- Issue type
- Bug
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Stale
- Clarity
- Mostly clear
- Newbie friendliness
- 35/100