World.JS

A JavaScript Wrapper for World Vocoder Powered by Emscripten

About World.JS

World.JS is A JavaScript Wrapper for World Vocoder Powered by Emscripten.

You can use World Vocoder in your Browser,Node or Electron app and many more.

About Emscripten

Emscripten is a toolchain for compiling to asm.js and WebAssembly, built using LLVM, that lets you run C and C++ on the web at near-native speed without plugins.

Fast as near-native speed.
Easy to use.
Powerful

About World

Morise's World Vocoder is a fast and high-quality vocoder. World Vocoder parameterizes speech into three components:

Pitch (fundamental frequency, F0) contour
Harmonic spectral envelope
Aperiodic spectral envelope (relative to the harmonic spectral envelope)

For more information, please visit Morise's World repository and the official website of World Vocoder

Build Status

Ubuntu:

API

World Native:
- [x] WavRead_JS             # .wav input functions.
- [x] DIO_JS                 # F0 estimation based on DIO (Distributed Inline-filter Operation).
- [x] Harvest_JS             # F0 estimation based on Harvest.
- [x] DisplayInformation     # Display Information.
- [x] CheapTrick_JS          # Spectral envelope estimation on the basis of the idea of CheapTrick.
- [x] D4C_JS                 # Band-aperiodicity estimation on the basis of the idea of D4C.
- [X] Synthesis_JS           # Voice synthesis based on f0, spectrogram and aperiodicity.
- [x] WavWrite_JS            # Output Wav File.

Wrapper:
- [ ] Wav2World              # Wav to world.
- [ ] WavInJS                # Load wav file as WavRead do.

Example

Expand View full directory trees

WavRead_JS

.wav input functions

<script src="WorldJS.js"></script> 

<script>
    let x = Module.WavRead_JS("vaiueo2d.wav");
    console.log(x);
    console.log(x.x);
    console.log(x.fs);
    console.log(x.nbit);
    console.log(x.x_length);
</script>

DIO_JS

F0 estimation based on DIO (Distributed Inline-filter Operation).

<script src="WorldJS.js"></script> 

<script>
    let x = Module.WavRead_JS("vaiueo2d.wav");
    let f0 = Module.Dio_JS(x.x,x.fs,2.0);
    console.log(x);
    console.log(f0);
</script>

Harvest_JS

F0 estimation based on Harvest.

<script src="WorldJS.js"></script> 

<script>
    let x = Module.WavRead_JS("vaiueo2d.wav");
    let f0 = Module.Harvest_JS(x.x,x.fs,2.0);
    console.log(x);
    console.log(f0);
</script>

DisplayInformation

Display Information

<script src="WorldJS.js"></script> 

<script>
    let x = Module.WavRead_JS("vaiueo2d.wav");
    Module.DisplayInformation(x.fs, x.nbit, x.x_length);
</script>

CheapTrick_JS

Spectral envelope estimation on the basis of the idea of CheapTrick.

<script src="WorldJS.js"></script> 

<script>
    let x = Module.WavRead_JS("vaiueo2d.wav");
    let f0 = Module.Dio_JS(x.x, x.fs, 2.0);
    Module.CheapTrick_JS(x.x, f0.f0, f0.time_axis, x.fs, 2.0);
</script>

D4C_JS

Band-aperiodicity estimation on the basis of the idea of D4C.

<script src="WorldJS.js"></script> 

<script>
    let x = Module.WavRead_JS("vaiueo2d.wav");
    let f0 = Module.Dio_JS(x.x, x.fs, 2.0);
    let sp = Module.CheapTrick_JS(x.x, f0.f0, f0.time_axis, x.fs, 2.0);
    Module.D4C_JS(x.x, f0.f0, f0.time_axis, sp.fft_size, x.fs, 2.0);
</script>

Synthesis_JS

Voice synthesis based on f0, spectrogram and aperiodicity.

<script src="WorldJS.js"></script> 

<script>
    let x = Module.WavRead_JS("vaiueo2d.wav");
    let f0 = Module.Dio_JS(x.x, x.fs, 2.0);
    let sp = Module.CheapTrick_JS(x.x, f0.f0, f0.time_axis, x.fs, 2.0);
    let ap = Module.D4C_JS(x.x, f0.f0, f0.time_axis, sp.fft_size, x.fs, 2.0);
    Module.Synthesis_JS(f0.f0, sp.spectral, ap.aperiodicity, sp.fft_size, x.fs, 2.0)
</script>

WavWrite_JS

Output Wav File

<script src="WorldJS.js"></script> 
<script>
    let x = Module.WavRead_JS("vaiueo2d.wav");
    Module.WavWrite_JS(x.x, x.fs, x.nbit, "out.wav")
</script>

Build World.JS

Before Build

You must install these software in advance:

CMake
Python3 environment
Emscripten SDK
Node
VSCode
- VSCode extension: CMake Tools
Software development tool chain like make gcc etc.

Clone Files

git clone https://github.com/YuzukiTsuru/World.JS.git
git submodule init
git submodule update

Set emsdk

Edit it in .vscode/settings.json and change the path into your own.

"cmake.configureArgs": ["-DCMAKE_TOOLCHAIN_FILE=~/sdk/emsdk/upstream/emscripten/cmake/Modules/Platform/Emscripten.cmake"]

Run build script

Before run it, make sure vaiueo2d.wav is existence in the build folder or change the CMakeLists.txt and delete the preload-file, this file is only for test.

And make with vscode.

Troubleshooting

Uncaught could not load memory initializer

Access to XMLHttpRequest at 'file:///World.JS/test/WorldJS.data' from origin 'null' has been blocked by CORS policy: Cross origin requests are only supported for protocol schemes: http, data, chrome, chrome-extension, https.
WorldJS.js:15 Uncaught could not load memory initializer WorldJS.js.mem

Unfortunately several browsers (including Chrome, Safari, and Internet Explorer) do not support file:// XHRrequests, and can’t load extra files needed by the HTML (like a .mem file, or packaged file data as mentioned lower down). For these browsers you’ll need to serve the files using a webserver. The easiest way to do this is to use the python HTTPServer (in the current directory do python -m SimpleHTTPServer (Python2) python -m http.server (python3) and then open http://localhost:8000).

File System Develop

The File System is based on Emscripten,visit Emscripten home for more detials:

https://emscripten.org/docs/porting/files/index.html

References

[1] M. Morise, F. Yokomori, and K. Ozawa: WORLD: a vocoder-based high-quality speech synthesis system for real-time applications, IEICE transactions on information and systems, vol. E99-D, no. 7, pp. 1877-1884, 2016. [2] M. Morise: D4C, a band-aperiodicity estimator for high-quality speech synthesis, Speech Communication, vol. 84, pp. 57-65, Nov. 2016. http://www.sciencedirect.com/science/article/pii/S0167639316300413

In CheapTrick, you can refer the following references.

[3] M. Morise: CheapTrick, a spectral envelope estimator for high-quality speech synthesis, Speech Communication, vol. 67, pp. 1-7, March 2015. http://www.sciencedirect.com/science/article/pii/S0167639314000697 [4] M. Morise: Error evaluation of an F0-adaptive spectral envelope estimator in robustness against the additive noise and F0 error, IEICE transactions on information and systems, vol. E98-D, no. 7, pp. 1405-1408, July 2015.

In DIO, you can refer the following reference.

[5] M. Morise, H. Kawahara and H. Katayose: Fast and reliable F0 estimation method based on the period extraction of vocal fold vibration of singing voice and speech, AES 35th International Conference, CD-ROM Proceeding, Feb. 2009.

In Harvest, you can refer the following reference.

[6] M. Morise: Harvest: A high-performance fundamental frequency estimator from speech signals, in Proc. INTERSPEECH 2017, pp. 2321–2325, 2017. http://www.isca-speech.org/archive/Interspeech_2017/abstracts/0068.html

In the codec of spectral envelope, you can refer the following reference.

[7] M. Morise, G. Miyashita and K. Ozawa: Low-dimensional representation of spectral envelope without deterioration for full-band speech analysis/synthesis system, in Proc. INTERSPEECH 2017, pp. 409-413, 2017. http://www.isca-speech.org/archive/Interspeech_2017/abstracts/0067.html

A paper was published to demonstrate that the current version of WORLD was superior to the similar vocoders in the sound quality of re-synthesized speech. This paper also includes the detailed information in the D4C LoveTrain used in the latest version.

[8] M. Morise and Y. Watanabe: Sound quality comparison among high-quality vocoders by using re-synthesized speech, Acoust. Sci. & Tech., vol. 39, no. 3, pp. 263-265, May 2018. https://www.jstage.jst.go.jp/article/ast/39/3/39_E1779/_article/-char/en

File Tree

.
├─build
├─example
│  └─html
├─lib
│  └─WorldCmake
│      └─World
├─src
│  └─WorldJS
├─test
└─tools
    └─third_party
        └─emscripten