Retrieval-based-Voice-Conversion-WebUI
Retrieval-based-Voice-Conversion-WebUI copied to clipboard
Proposed Enhancement: Input Optimization and Tuning
-
I tend to batch export my vocals (up to 30 tracks sometimes) from logic with a set duration and cycle range to make importing easier. This results in some files being mostly empty with only small parts having vocals in them. Automatically skipping the empty parts could improve performance.
-
If I understand it right the input audio gets analyzed for pitch in the first step of conversion. Could it be possible to "snap" this process to a scale to easily tune the audio perfectly? Think non-realtime autotune or melodyne.
Would be great if the following is possible:
- Automatically skip parts of input audio below a threshold (threshold slider in the UI)
- Tune the output audio to a set scale (scale selection dropdown and tuning strength slider in the UI)