Vocalforge

Professional AI Vocal Extraction Studio

A cutting-edge web application that uses advanced AI to separate vocals from musical instruments in audio and video files. Built with Flask and powered by the state-of-the-art Demucs neural network, VocalForge delivers studio-quality vocal extraction with an intuitive, professional interface.

Perfect for creating halal content, acapella versions, vocal practice tracks, or any project requiring isolated vocals.

---

✨ Features

🎨 Professional Enterprise-Grade Interface

• Modern Layout - Split-screen design with input on left, output on right
• Professional Color Scheme - Deep blue & gold accent palette for trust and premium feel
• Proper File Structure - Separated CSS and JavaScript in static folder
• Gradient Header - Eye-catching branding with live statistics
• Smooth Animations - GPU-accelerated transitions and micro-interactions
• Fully Responsive - Adapts beautifully to desktop, tablet, and mobile
• Real-time Feedback - Instant visual response to all user actions
• Auto-Cleanup - Automatic file cleanup prevents disk space buildup

🚀 Powerful AI Processing

• 6-Stem Model - Uses htdemucs_6s for best jazz/classical/orchestral separation
• Studio Quality - Professional-grade vocal extraction at 320kbps MP3
• Advanced Processing - Overlap and segmentation for cleaner results
• Jazz/Classical Optimized - Better handling of trumpet, piano, strings
• GPU Accelerated - Automatic CUDA detection for 10-20x faster processing
• Smart Progress - Live progress tracking with device info and time estimates
• Batch Ready - Process files up to 500MB

📁 Comprehensive File Support

• Drag & Drop - Intuitive file upload with visual feedback
• Audio Formats - MP3, WAV, FLAC, M4A, OGG
• Video Formats - MP4, AVI, MKV, MOV (extracts audio automatically)
• In-Browser Playback - Preview results instantly
• One-Click Download - Get your vocals immediately

---

📖 How It Works

VocalForge uses a sophisticated AI pipeline:

1. Upload - Drag or select your audio/video file (up to 500MB)
2. AI Processing - Demucs neural network analyzes and separates audio sources
3. Extraction - Vocals are isolated from instruments with studio-quality precision
4. Output - Download or preview your 320kbps MP3 vocals

---

💻 Installation

Prerequisites

• Python 3.8+ - Download Python
• FFmpeg - Required for audio processing
• NVIDIA GPU (Optional) - For CUDA acceleration

Install FFmpeg

Windows:

# Download from https://ffmpeg.org/download.html
# Or use Chocolatey:
choco install ffmpeg

Linux (Ubuntu/Debian):

sudo apt-get update
sudo apt-get install ffmpeg

macOS:

brew install ffmpeg

Setup VocalForge

1. Clone the repository

git clone <repository-url>
cd VocalForge

1. Create virtual environment

python -m venv venv

1. Activate virtual environment

Windows:

venv\Scripts\activate

macOS/Linux:

source venv/bin/activate

1. Install dependencies

pip install -r requirements.txt

---

🎯 Usage

1. Activate virtual environment

# Windows
venv\Scripts\activate

# macOS/Linux
source venv/bin/activate

1. Launch VocalForge

python app.py

1. Open in browser

http://localhost:5000

1. Process your files
   • Upload audio/video file via drag & drop or file browser
   • Monitor real-time progress with device info
   • Preview results in the browser
   • Download your vocals

---

⚙️ Technical Details

AI Model

• Model: Demucs v4 (htdemucs_ft)
• Developer: Meta AI Research
• Architecture: Hybrid transformer with temporal convolutions
• Quality: State-of-the-art music source separation
• Output: 320kbps MP3 (studio quality)

Performance

• GPU (CUDA): ~40-80 seconds for 3-min song ⚡
• CPU: ~4-8 minutes for 3-min song
• First run: Downloads AI model (~300MB)
• Quality: 6-stem model (htdemucs_6s) optimized for jazz/classical

Supported Formats

• Audio: MP3, WAV, FLAC, M4A, OGG
• Video: MP4, AVI, MKV, MOV (audio extracted)
• Max size: 500MB

---

📁 Project Structure

VocalForge/
├── app.py                 # Flask application & API
├── requirements.txt       # Python dependencies
├── README.md             # Documentation
├── templates/
│   └── index.html        # Professional dual-column UI
├── uploads/              # Temporary upload storage
├── outputs/              # Processed vocal files
└── venv/                 # Virtual environment

---

🔧 Troubleshooting

Error: "demucs command not found"

• Activate virtual environment: venv\Scripts\activate
• Reinstall: pip install demucs

Error: "FFmpeg not found"

• Install FFmpeg (see Installation section)
• Ensure FFmpeg is in system PATH

Processing is slow

• Install CUDA-enabled PyTorch for GPU acceleration
• Close other applications to free resources
• Process shorter audio segments

Upload fails

• Check file size (max 500MB)
• Verify file format is supported
• Ensure sufficient disk space

---

🚀 Performance Tips

1. Use GPU - 10-20x faster processing
2. Close Apps - Free system resources
3. SSD Storage - Faster file operations
4. Shorter Files - Process 3-5 min segments

---

📄 License

This project is open source and available under the MIT License. The Demucs model is subject to its own license from Meta AI.

---

🙏 Credits

• Demucs - Meta AI's music source separation
• Flask - Web framework
• FFmpeg - Multimedia processing
• PyTorch - Deep learning platform

---

VocalForge - Professional AI Vocal Extraction Studio Powered by Demucs AI | Built with Flask