worldveil/dejavu
Dejavu
Dejavu is a Python audio fingerprinting library and recognition engine. It functions as a digital audio signature tool used to analyze sound waves and create unique identifiers for the purposes of audio search and retrieval.
The project enables automatic music identification by matching live audio feeds or recorded clips against a database of fingerprints. It covers audio content matching and digital audio archiving to identify original source recordings from a stored collection.
The system incorporates capabilities for generating audio fingerprints, identifying audio tracks, and recognizing audio segments from both files and microphone streams.
Features
- Audio Recognition Tools - Provides a complete engine for identifying audio tracks from digital fingerprints against a stored collection.
- Fingerprint Indexing - Provides a keyed database indexing system to allow rapid lookup of matching audio fingerprints.
- Track Verification - Identifies specific audio tracks by comparing live feeds or files against a database of fingerprints.
- Audio Fingerprinting - Generates unique digital signatures from audio files to enable future identification and retrieval.
- Audio Source Matching - Enables the comparison of two different audio samples to verify if they share the same source recording.
- Audio Signature Tools - Provides a utility for analyzing sound waves to create unique identifiers for audio search and retrieval.
- Audio Temporal Alignment - Validates candidate matches by ensuring the temporal distance between fingerprints is consistent across the recording.
- Fingerprint Generation - Analyzes spectrograms to identify peak energy points, creating unique digital signatures for audio tracks.
- Automatic Track Identification - Automatically identifies original music tracks by matching live audio feeds or clips against a fingerprint library.
- Audio Fingerprinting Libraries - Implements a specialized Python library for creating digital signatures from audio files for track recognition.
- Fourier Transforms - Uses Fast Fourier Transforms to convert raw audio waveforms into the frequency domain for spectral analysis.
- Audio Fingerprint Matching - Implements a constellation-map matching system that pairs frequency peaks into hashes for robust audio identification.
- Audio Collection Archiving - Supports digital audio archiving by generating fingerprints to organize and search large audio collections.
- Real-Time Audio Streaming Buffers - Implements real-time audio streaming buffers to capture microphone input for continuous recognition.
- Audio Processing - Audio fingerprinting and recognition.
