AudioFingerprinterBase<T>

Base Classes · AiDotNet.Audio.Fingerprinting

Base class for audio fingerprinting algorithms.

Implements: IAudioFingerprinter<T>

Known derived types: ChromaprintFingerprinter<T>, SpectrogramFingerprinter<T>

For Beginners

Audio fingerprinting is like Shazam - it can identify a song from a short audio clip, even if the audio is noisy or compressed.

How it works:

  1. Audio is converted to a spectrogram
  2. Key features (peaks, patterns) are extracted
  3. Features are hashed into a compact fingerprint
  4. Fingerprints can be matched against a database

This base class provides:

  • Hash computation utilities
  • Hamming distance for comparison
  • Time alignment for matching

How It Works

Audio fingerprinting creates compact identifiers that can recognize audio content even after degradation (compression, noise, speed changes). Unlike neural network approaches, fingerprinting typically uses signal processing techniques.

Properties

Property Summary
double FrameDuration { get; set; } Gets or sets the duration of each fingerprint frame in seconds.
int FingerprintLength { get; set; } Gets the fingerprint length in bits or elements.
int SampleRate { get; set; } Gets the expected sample rate for input audio.
string Name { get; } Gets the name of the fingerprinting algorithm.

Methods

Method Summary
AudioFingerprint<T> Fingerprint(Tensor<T> audio) Generates a fingerprint from audio data.
AudioFingerprint<T> Fingerprint(Vector<T> audio) Generates a fingerprint from audio data.
IReadOnlyList<FingerprintMatch> FindMatches(AudioFingerprint<T> query, AudioFingerprint<T> reference, int minMatchLength) Finds matching segments between two fingerprints.
double ComputeSimilarity(AudioFingerprint<T> fp1, AudioFingerprint<T> fp2) Computes the similarity between two fingerprints.