SplitVoice

This app can separate singing voice and instruments.

This is a sound source separation app that separates music or video files (mp3, wav, m4a, mp4) in your device into vocal and instrumental track, which can be freely remixed and played back.

For example, you can

- Create karaoke by removing vocals.

- Listen to a cappella by extracting vocals.

- Listen to a remixed version of the song by freely changing the volume of vocal and instrumental track.

- Listen repeatedly to a specified section of a song.

- Remove only the voice from the audio of a video file.

Please use this app for creating karaoke, practicing singing, making MAD videos, etc.

[Limitations]

This application uses the engine described below to perform the separation process into instrument tracks in the terminal without using a server.

This process uses a technology called deep learning, which requires a lot of memory (RAM) and processing time.

When I checked on my device, Pixel3a (RAM: 3,592MB), it took about 4 minutes to separate a 4-minute song.

You can playback the separated parts during the process, so please be patient while listening.

I have tried various ways to reduce the weight of the app, but so far we have not found a solution.

If you cannot use this application on your device, I am sorry, but please give up and uninstall it.

[Software used]

This application uses Spleeter by deezer as the sound source separation engine.

https://github.com/deezer/spleeter

Spleeter is a software provided under MIT license.

https://github.com/deezer/spleeter/blob/master/LICENSE

[Please note]

The development of this application is done when the author feels like it.

I am sorry, but I do not provide support (e.g., answers to inquiries), so please use this application to the extent that you can understand and use it in its current state.

[Disclaimer]

This application has been tested on the author's own device and used by the author himself, but the author assumes no responsibility for any damages incurred by the user as a result of using this application.

Also, please understand that the author does not provide any support (e.g., response to inquiries) regarding this application.