Skip to content
This repository has been archived by the owner on Sep 30, 2020. It is now read-only.
/ Vocalnet Public archive

A wavenet and waveglow based singing synthesizing system.

License

Notifications You must be signed in to change notification settings

xushengyuan/Vocalnet

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

7 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Vocalnet(In progress)

基于Wavenet与Waveglow的中文歌声合成系统

Webdemo

http://39.98.177.168/

安装

使用源码

  1. 从Github下载代码
git clone https://github.com/xushengyuan/Vocalnet.git
  1. 安装相关依赖包
pip install -r requirements.txt
  1. 从Github的release中下载预训练模型,保存到项目根目录。大陆用户可使用多线程下载工具(如aria)提高下载速度及稳定性。

生成

程序接收由Vocalloid编辑器保存的.vsqx工程文件,仅读取第一条轨道中的音符。

一键合成

python Vocalnet_synthesis.py in.vsqx

分布合成

  1. 在vsqx中保存工程到根目录,执行vsqx2npy.py.vsqx解析

    python vsqx2npy.py in.vsqx out.npy
    
  2. 合成梅尔谱

    python synthesis.py checkpoint_step000800000_ema.pth ./out --conditional="out.npy"
    
  3. 合成波形

        python waveglow_generate.py --checkpoint="model.ckpt-660000.pt" --local_condition_file="./out/checkpoint_step000800000_ema.npy"
    

波形文件保存在checkpoint_step000800000_ema.wav

训练

Wavenet

演示视频

参考

About

A wavenet and waveglow based singing synthesizing system.

Topics

Resources

License

Stars

Watchers

Forks

Packages

No packages published

Languages