Voice Cloning - Speak with any voice

Feb 28, 2024#AI33

AI Translation

This post is translated from Chinese into English through AI.View Original

AI-generated summary

Clone-voice is a voice cloning tool with a web interface that allows users to record audio using their own voice or any other voice. After using it, the result may be around 50% similar. The necessary files can be downloaded from Github, including the main precompiled file, models for sound-to-sound and text-to-sound conversion, and an optional tool for separating background and vocals. Users need to copy and extract the model files into the tts directory of the main file, then double-click app.exe to start the tool.

Introduction#

clone-voice is a voice cloning tool with a web interface that allows you to record audio using your voice or any sound.

After using it, I feel like it might be 50% similar, somewhere between similar and dissimilar, haha.

Download#

The following steps 1 and 2 can be accelerated using Github File Acceleration | GitMirror.

Required Pre-compiled main file (1.7G): github.com/jianchang512/clone-voice/releases
Required Models (3G) model-sts.7z and model-tts.7z: github.com/jianchang512/clone-voice/releases/tag/v0.0.1
1. sts: Sound to sound
2. tts: Text to speech
Optional Background and vocal separation tool: github.com/jianchang512/vocal-separate/releases

Configuration#

Copy the two model files to the tts directory of the pre-compiled main file, and extract by choosing "Extract to current location", the final effect is as follows:

Start#

Double-click on the app.exe in the pre-compiled main file and wait for it to start up.