Chromaprint1.6.2

Generate compact audio fingerprints from recordings with Chromaprint, giving AcoustID-connected software data it can use to identify near-identical audio. This C library is designed for full-file identification and duplicate-audio-file detection, rather than as a general-purpose fingerprinting system.

Developers can build the shared library alongside fpcalc, a command-line utility used by MusicBrainz Picard; this route requires the FFmpeg libraries for fpcalc’s audio decoding and resampling. Chromaprint supports FFT processing through FFmpeg, FFTW3, KissFFT, or macOS’s vDSP, with FFmpeg preferred except on macOS, where vDSP is recommended. FFTW3 carries the GPL, making a Chromaprint binary that uses it GPL-licensed, while bundled KissFFT is slower but permissively licensed and serves as a fallback when other libraries are unavailable. The FFT_LIB option selects the FFT implementation, and FFMPEG_ROOT identifies a nonstandard FFmpeg location; libav may work, but FFmpeg is preferred.

This Windows x64 release, published by AcoustID, is aimed at developers and technical users integrating fingerprint generation into identification workflows. Doxygen can produce HTML API documentation, while building the test suite requires Google Test sources.

Version 1.6.2 fixes a heap-buffer overflow when decoding fingerprints containing data after their encoded values, making the update important for applications processing untrusted fingerprints. It also preserves partial audio frames across buffer calls instead of dropping them and restores builds with FFmpeg 4.x after its channel-layout API changed in 5.1.