voicepaw/so-vits-svc-fork

so-vits-svc fork with realtime support, improved interface and more features.

What it solves

This project provides a system for singing voice conversion, allowing a user to change the voice of a source audio file or live microphone input to sound like a specific target speaker. It improves upon the original so-vits-svc by adding real-time capabilities, a more user-friendly interface, and faster training.

How it works

It uses a VITS-based architecture for voice conversion. The pipeline involves preprocessing audio (resampling, splitting, and diarization), extracting features using models like HuBERT or CREPE for pitch estimation, and training a model on a target speaker's dataset. Once trained, the model can be used for inference via a command-line interface (CLI) or a graphical user interface (GUI) to convert audio in real-time or from files.

Who it’s for

It is designed for users who want to create voice clones for singing or speech, ranging from beginners who can use the one-click installer and GUI to more advanced users who can train their own models using local GPUs or cloud environments like Google Colab and Paperspace.

Highlights

  • Real-time conversion: Supports live voice changing from a microphone.
  • Improved Interface: Includes both a GUI and a unified CLI for easier operation.
  • Faster Training: Training speed is approximately twice as fast as the original repository.
  • Simplified Installation: Available via pip and includes a one-click .bat installer for Windows.
  • Enhanced Accuracy: Uses CREPE for more precise pitch estimation and fixes previous misuse of ContentVec.

Related

  • Project
    yxlllc/DDSP-SVC

    An efficient singing voice conversion project that enables high-quality AI voice changing with significantly lower hardware requirements and faster training times than traditional SVC models.

  • Project
    prophesier/diff-svc

    A singing voice conversion tool that uses diffusion models to transform input singing voices into a target timbre with basic pitch correction support.

  • Project
    abus-aikorea/voice-pro

    An AI-powered multimedia tool that integrates speech recognition, translation, and voice cloning for automated dubbing and multilingual content creation.

  • Project
    RVC-Boss/GPT-SoVITS

    A powerful few-shot voice conversion and text-to-speech system that can clone a voice using as little as 5 seconds of audio.

  • Project
    jianchang512/pyvideotrans

    An open-source video translation tool that automates speech recognition, subtitle translation, and AI dubbing with support for voice cloning and multi-role synthesis.