From ChatGPT

Audio fingerprints, also known as audio fingerprints or audio hashing, are compact and unique representations of audio signals that can be used for various purposes, such as audio identification, content recognition, and audio retrieval. Audio fingerprints are typically generated from the audio content of a given audio file or stream, and they are designed to be robust to common audio transformations such as noise, compression, and format conversions.

Audio fingerprints are usually generated by extracting specific features from an audio signal, such as the amplitude, frequency, or spectral characteristics of the audio, and then converting these features into a compact representation, often in the form of a fixed-length hash or a numerical code. This fingerprint can then be used to compare against a database of pre-existing fingerprints to identify or recognize audio content.

One common application of audio fingerprints is in audio identification, where audio fingerprints of a reference set of audio files are generated and stored in a database, and incoming audio can be compared against these fingerprints to identify the content. Audio fingerprints are also used in audio retrieval systems, where users can search for audio content based on a query fingerprint, and in content recognition systems, where audio fingerprints are used to detect copyrighted or unauthorized audio content.

Audio fingerprints have become an important technology in various industries, including music streaming, broadcast monitoring, copyright enforcement, and audio surveillance, among others. They are also used in various audio-based applications, such as music recommendation, audio search, and audio synchronization.