🎵 AI Music Genre Classifier

A sophisticated web application that uses machine learning to classify music genres from audio files with 73% accuracy. Features comprehensive audio analysis, real-time visualizations, and multiple music processing tools.

✨ Features

🎯 Core Functionality

• Genre Classification: ML model with 73% accuracy across 10 genres
• Real Audio Analysis: MFCC, spectral, and temporal feature extraction
• Confidence Scoring: Detailed probability breakdowns with uncertainty detection

🎨 Visualization & Analysis

• Real-time Audio Visualizations: Waveform, spectrum, MFCC, and time-domain graphs
• Interactive Audio Player: Vinyl-style player with equalizer
• Feature Extraction: Tempo, key, energy, spectral centroid, and more

⚡ Advanced Tools

• Batch Processing: Upload and analyze multiple files simultaneously
• Audio Mashup Mixer: Blend two tracks with adjustable parameters
• Music Visualizer: Real-time animated visualizations
• Audio Comparison: Side-by-side track analysis with similarity scoring

📊 Export & Sharing

• Multiple Formats: JSON, CSV, PDF, and text reports
• Batch Export: Comprehensive results for multiple files
• Shareable Results: Clipboard copy and social sharing

🎼 Supported Genres

• Blues 🎸
• Classical 🎻
• Country 🤠
• Disco 💃
• Hiphop 🎤
• Jazz 🎷
• Metal 🤘
• Pop 🎤
• Reggae 🌴
• Rock 🎸

🚀 Quick Start

Prerequisites

• Python 3.8+
• Modern web browser with Web Audio API support

Installation

1. Clone the repository

   git clone https://github.com/yourusername/music-genre-classifier.git
   cd music-genre-classifier

2. Install Python dependencies

   pip install -r requirements.txt

3. Train the model (optional - pre-trained model included)

   python train_model_v2.py

4. Start the Flask server

   python app.py

5. Open your browser

   http://localhost:5000
   

🛠️ Technical Stack

Backend

• Flask: Web framework and API
• Scikit-learn: Machine learning (Random Forest + Gradient Boosting ensemble)
• Librosa: Audio feature extraction
• NumPy/SciPy: Scientific computing

Frontend

• Vanilla JavaScript: Modern ES6+ features
• Web Audio API: Real-time audio processing
• Canvas API: Dynamic visualizations
• Tailwind CSS: Responsive styling

Audio Processing

• MFCC Analysis: 13 coefficients with statistical features
• Spectral Features: Centroid, rolloff, bandwidth, contrast
• Temporal Features: Tempo, zero-crossing rate, RMS energy
• Harmonic/Percussive: Source separation analysis

music-genre-classifier/ ├── app.py # Flask backend server ├── train_model.py # Original model training ├── train_model_v2.py # Improved ensemble model ├── debug_model.py # Model debugging and testing ├── models/ # Trained model files ├── static/ │ ├── script.js # Main application logic │ ├── real-audio-visualizer.js # Audio visualization │ ├── batch-processor.js # Batch upload system │ ├── export-manager.js # Export functionality │ └── keyboard-shortcuts.js # UX enhancements ├── templates/ │ └── index.html # Main interface └── requirements.txt # Python dependencies

🎹 Usage

Single File Analysis

• Drag & drop an audio file (MP3/WAV, up to 50MB)
• View real-time analysis with interactive visualizations
• Explore genre prediction with confidence scores
• Export results in multiple formats

Batch Processing

• Upload multiple audio files
• Monitor progress with live updates
• View aggregated statistics and genre distribution
• Export comprehensive batch reports

Audio Tools

• Mashup Mixer: Combine two tracks with volume control and crossfade
• Music Visualizer: Real-time animated spectrum displays
• Audio Comparison: Compare tracks with similarity scoring

🔧 API Endpoints

• POST /api/classify - Analyze single audio file
• GET /api/health - Server status check
• GET /api/genres - List supported genres

📈 Model Performance

• Accuracy - 73% (improved ensemble model)
• Featurea - 43 audio characteristics per sample
• Genres - 10 distinct music categories
• Processing - ~2-10 seconds per file (depending on size)

🌟 Key Features

Smart Confidence Handling

• Automatic uncertainty detection for low-confidence predictions
• Alternative genre suggestions with probabilities
• Visual indicators for prediction reliability

Professional Visualizations

• Real waveform rendering from audio buffer data
• Interactive frequency spectrum analysis
• MFCC coefficient displays
• Beat detection and tempo visualization

Enterprise-Grade Tools

• Concurrent batch processing with progress tracking
• Comprehensive export system (JSON, CSV, PDF, text)
• Responsive design for all devices
• Keyboard shortcuts for power users

🤝 Contributing

We welcome contributions! Please feel free to submit pull requests, report bugs, or suggest new features.

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.

🙏 Acknowledgments

• GTZAN Dataset for genre classification benchmarks
• Librosa library for audio feature extraction
• Web Audio API for browser-based audio processing