ExtractVocal

A web application that separates vocals from music using AI-powered audio source separation. Upload audio files or provide YouTube URLs to extract clean vocal and accompaniment tracks.

Features

• File Upload: Upload MP3, WAV, or other audio files directly
• YouTube Integration: Extract audio from YouTube videos by URL
• AI-Powered Separation: Uses Spleeter for high-quality vocal/accompaniment separation
• Dual Output: Download both vocal and accompaniment tracks separately
• Modern UI: React-based frontend with internationalization support

Tech Stack

Backend:

• Python 3
• Flask (REST API)
• Spleeter (audio source separation)
• youtube-dl (YouTube audio extraction)

Frontend:

• React
• Redux (state management)
• Axios (HTTP client)
• i18next (internationalization)

Prerequisites

• Python 3.6+
• Node.js 12+
• FFmpeg (required by Spleeter)

Installation

Backend Setup

1. Navigate to the backend directory:

cd back-end

2. Install Python dependencies:

pip install -r requirements.txt

3. Install FFmpeg if not already installed:

# macOS
brew install ffmpeg

# Ubuntu/Debian
sudo apt-get install ffmpeg

# Windows
# Download from https://ffmpeg.org/download.html

Frontend Setup

1. Navigate to the frontend directory:

cd front-end

2. Install Node.js dependencies:

npm install

Usage

Starting the Backend

cd back-end
python test-spleeter.py

The Flask server will start on http://0.0.0.0:5000

Starting the Frontend

cd front-end
npm start

The React app will start on http://localhost:3000

API Endpoints

• POST /upload - Upload an audio file for vocal separation
• POST /youtubeToMp3 - Provide a YouTube URL to extract and separate vocals
• GET /accompaniment/<timestamp> - Download the accompaniment track
• GET /title - Get the title of the processed audio

How It Works

1. User uploads an audio file or provides a YouTube URL
2. Backend downloads/receives the audio file
3. Spleeter processes the audio using a pre-trained 2-stem model
4. The audio is separated into two tracks: vocals and accompaniment
5. Both tracks are made available for download

Project Structure

ExtractVocal/
├── back-end/
│   ├── pretrained_models/    # Spleeter models
│   ├── test-spleeter.py      # Flask API server
│   └── requirements.txt      # Python dependencies
├── front-end/
│   ├── public/               # Static assets
│   ├── src/                  # React components
│   └── package.json          # Node.js dependencies
└── README.md

License

This project uses Spleeter, which is licensed under the MIT License.

Credits

• Spleeter by Deezer for audio source separation
• youtube-dl for YouTube audio extraction