Split

API that hosts the Spleeter neural network for source music separation and uploads and downloads files via REST.

Python Version used in the project was:

V3.6

The Last Python Version that works is here:

V3.7

Install PyEnv to manage multiple Versions:

https://realpython.com/intro-to-pyenv/

Create a virtual Python3.8 environment on Linux, Win, or Mac:

python -m venv venv3-8

source venv3-8/bin/activate

install requirements and libraries

pip install -r requirements.txt

Run the application in the Rest.py file, for now.

Command inside project folder:

python app.py

Routes:

127.0.0.1:5000/ - GET - Returns a test string in Json format

127.0.0.1:5000/upload - POST - File upload by parameter in Body - "audio" : <nameFile> - and neural network processing:

curl --location '127.0.0.1:5000/upload' \
--form 'audio=@"/home/filipe-borato/Documents/It dont mean i think_ master.mp3"'

127.0.0.1:5000/download - GET - Download the last track processed by the neural network

Split connects to another project of mine called CompressorAndSplit

The Split project is coupled to a plugin I created with the name Boratio, in C++ with Juce Framework, and it's a github project called CompressorAndSplit

THIS IS MY SERVER AND HE HOSTED THE SPLEETER NEURAL NETWORK

this project is a neural network that removes or separates the voice between the instruments

Prerequisite: FFmpeg

Spleeter uses FFmpeg to decode audio (mp3, wav, etc). Make sure ffmpeg is installed and available on your PATH before using /upload.

• Ubuntu/Debian (WSL/Linux):

  sudo apt-get update && sudo apt-get install -y ffmpeg
  

• macOS (Homebrew):

  brew install ffmpeg
  

• Windows:
  • Using winget: winget install Gyan.FFmpeg
  • Or Chocolatey: choco install ffmpeg
  • Ensure the bin folder (containing ffmpeg.exe) is added to your PATH.

Verify installation:

ffmpeg -version

Restart the Flask server after installation.