Skip to content

Spleeter GUI by Chataigne. Deezer separation tool & models.Separate vocal from others.

License

Notifications You must be signed in to change notification settings

zak-45/SpleeterGUI-Chataigne-Module

Repository files navigation

image

SpleeterGUI-Chataigne-Module

Spleeter GUI by Chataigne. Deezer separation tool & models

GUI Front end for Spleeter - AI source separation image

This project will make it easy for users to run Spleeter without needing to use the command line tools to do so. Built-in Chataigne audio feature will provide you all necessary to see/test the result with few clicks. You should be able to hear the song without bass, drums , vocals etc ...or only one of them.

This should work on all OS where Chataigne/Spleeter are supported (Win/macOS/Linux...)

Windows users can have it running in few minutes.

You do not need to install python or spleeter, this is a portable version of python pre-loaded with spleeter.

*** Videos DEMO ***

Spleeter_2stems_demo.mp4

2 stems

Spleeter_5stems_demo.mp4

5 stems

*** Prerequest ***

Need python 3 (3.10.x x64)/ spleeter & tensorflow modules / ffmpeg

pip3 install spleeter==2.4.0

Spleeter : https://github.com/deezer/spleeter/wiki/1.-Installation should install required python modules

see this for ffmpeg:

ffmpeg download : https://ffmpeg.org/download.html

Win user can use winget:

winget install ffmpeg

For Win users, portable version has been setup with all prerequest (except ffmpeg).

Win Installer

image image

Installation :

Manual

Copy this repository to :
<MyDocuments>/Chataigne/modules/Spleeter
nota: name must be 'Spleeter'

Win users can use the portable version with all pre-requests

Win Installer nota: you steel need to have ffmpeg installed

Use it :

Open  Chataigne.

Go to Modules, right click, Custom, Spleeter.

image

On Inspector:
	Spleeter command: command to execute
		This is the full path for the command to launch. Need to be adapted to your needs/ installation.
	Spleeter options : options to pass to the command
		Options that will be passed to the executed command.
	OutputFolder :Folder where to store separate audio tracks
		This one will store the separate audio file output
		If left blank, will be set to audio_output under temp location.
	Force : force command execution
		To optimize the process, SpleeterGUI will check if an already separation file exist and re-use it
		This parameter will force spleeter to execute and overwrite existing files.
	Verbose : provide more informations during command running.

image

On Command Tester, Spleeter: all SpleeterGUI available commands.
	separate : execute the command / options set in Parameters for audio track separate
		* sequence : audioclip file path from an existing sequence (with audio file enveloppe)  . Take in priority if set.
		* file : mp3 audio file name.
		* Model to use : select 2 / 4 or 5 stems

image

 Example after a 5 stems execution:

image

-- INFO --

This is CPU intensive: on Intel(R) Core(TM) i5-8300H CPU @ 2.30GHz 2.30 GHz, need 70 seconds to separate 3'30" audio track.Second run should be faster anymore.

Not all spleeter features has been implemented. Focus has been set to the separate. SpleeterGUI need a mp3 file as input and produce a mp3 file as result.

All others commands can be executed by using spleeter command line if necessary.

Spleeter WIKI : https://github.com/deezer/spleeter/wiki/2.-Getting-started

If you want the look and feel of the first screenshots, just use the spleeter.noisette file and the spleeter_bg.png provided.