So-vits-svc

So-vits-svc (SoftVC VITS Singing Voice Conversion) is an open-source neural network for converting singing vocals, letting you change the performer’s voice in an audio track while keeping the melody and phrasing.

What it does

• Replace a singer’s timbre in an audio file
• Preserve intonation and pitch
• Keep key vocal characteristics for more natural results
• Adjust conversion parameters such as speed, volume, and expressiveness
• Create a new voice based on selected settings
• Fine-tune/adapt a model for specific voices

How to run and download

So-vits-svc can be installed as a local program, and it’s also supported in Google Colab so you can run it in the cloud without installing extra software.

• Download files and documentation from the official GitHub repository
• Install via command line with: `pip install -U so-vits-svc-fork`

Typical workflow

• Prepare the source audio (the song/vocal you want to convert)
• Upload the audio and choose conversion settings
• Run conversion to generate covers/remixes with a different voice

So-vits-svc is free and available in English. Some forks add extra features, such as improved UI and real-time conversion (e.g., so vits svc fork voicepaw).