acoustid/chromaprint

What it solves

Chromaprint is designed to identify near-identical audio files and streams. It focuses on creating compact fingerprints that allow for high-performance searching, making it ideal for identifying full audio files, detecting duplicates, and monitoring long audio streams.

How it works

The library generates audio fingerprints based on research into music identification and efficient index-based matching. It can be built with various Fast Fourier Transform (FFT) libraries—such as FFmpeg, FFTW3, KissFFT, or vDSP—to handle the signal processing required to generate these fingerprints.

Who it’s for

It is intended for developers building audio identification systems, such as MusicBrainz Picard, or those needing to detect duplicate audio content across large datasets.

Highlights

• Compact Fingerprints: Optimized for search performance over absolute precision.
• Flexible FFT Support: Compatible with multiple FFT libraries to suit different licensing or performance needs.
• ** SBOM/Integrations**: Integrated into tools like FFmpeg and GStreamer, with bindings available in numerous languages including Python, Rust, and Go.