Singing Voice Conversion

A simple singing voice conversion demo and inference workflow.

Installation

Install the required dependencies:

pip install -r requirements.txt

Model

Download the model from the following link:

models.pth

Place the downloaded file in the following directory:

lib/vocal_remover/5_HP-Karaoke-UVR.pth

Usage

Main entry file:

demo_svc.py

Main Parameters

• sid: speaker ID, int, range [0, 18]
• f0_up_key: pitch shift, int, range [-12, 12]
  • Use 0 for male-to-male or female-to-female conversion
  • Use +12 for male-to-female conversion
  • Use -12 for female-to-male conversion
• index_rate: voice similarity, float, default 0.7, range [0, 1]

Requirements

See requirements.txt for the full environment dependencies.