Song Classifier: A Desktop App for Automatic Music Genre Recognition Using Machine Learning

Song Classifier is a desktop app released by monk10 on GitHub. Designed with a QML interface, it uses machine learning to analyze audio features (such as MFCC), supporting recognition of 10 music genres, multi-format audio processing, and batch operations. It helps users efficiently manage music collections and generate automatic tags, suitable for scenarios like personal music management and DJ tools.

Original Author and Source

• Original Author/Maintainer: monk10
• Source Platform: GitHub
• Project Title: song-classifier
• Original Link: https://github.com/monk10/song-classifier
• Release Date: 2026-05-25

Project Overview

Song Classifier is a user-friendly desktop app that aims to automatically classify songs into 10 music genres via audio analysis technology, simplifying the music collection management process, helping users understand the composition of their music library, and suitable for local music organization or streaming tag generation.

Audio Feature Extraction

• Time-domain features: Zero-crossing rate (reflects brightness)
• Frequency-domain features: FFT analysis of frequency energy distribution
• MFCC: A common feature simulating human auditory perception
• Rhythm features: Beat detection, BPM estimation
• Timbre features: Spectral centroid, rolloff, etc.

Machine Learning Model

By training on large-scale music datasets, it learns feature patterns of different genres, establishes a mapping from features to genre labels, and achieves high-accuracy classification.

1. 10 Genre Support: Covers mainstream genres for precise music collection organization
2. Modern QML Interface: Intuitive interaction, responsive layout, and adaptation to different screens
3. Multi-format Support: Compatible with MP3, WAV, FLAC, AAC, etc., no conversion needed
4. Drag-and-Drop & Batch Processing: Drag to import files, process multiple at once to improve efficiency
5. Result Export: Classification results can be exported as CSV reports for easy integration with other tools

• Personal Music Management: Automatically add genre tags to quickly build a structured music library
• DJ/Music Producers: Filter tracks by style to improve work efficiency
• Music Education & Research: Analyze music style evolution and process large datasets
• Streaming Reference: Similar technology can be used for automatic tagging, recommendations, and preference analysis

• QML & Qt Framework: QML builds a beautiful UI, integrated with C++ backend to balance performance and experience
• Cross-Platform: Supports Windows 10+ and macOS 10.12+ for a consistent experience
• Open Source Community: Open-source model with transparent and customizable code; community contributions and feedback are welcome

System Requirements

• Operating System: Windows 10+ / macOS 10.12+
• Hardware: 4GB+ RAM, dual-core 2GHz+ processor, 500MB+ available space

Installation Steps

1. Download the installation package for your system from the GitHub Releases page
2. Windows: Extract and run the exe file
3. macOS: Drag the app into the Applications folder to use

Song Classifier is a practical application of machine learning in music management. It helps users organize music collections through automated analysis, with advantages like a modern interface and wide format support. Although there may be audio processing issues, its core functions demonstrate technical potential. It is recommended for music enthusiasts and users needing management tools to try and follow this project.