A Robust Audio Fingerprinter Based on Pitch Class Histograms Applications for Ethnic Music Archives

Abstract

In this paper we present a new acoustic fingerprinting system, based on pitch class histograms. The aim of acoustic fingerprinting is to generate a small representation of an audio signal that can be used to identify identical, or recognize similar, audio snippets in a large audio set. A robust fingerprinting system generates similar fingerprints for perceptually similar audio signals. A piece of music with a noise added should generate an almost identical fingerprint as the original. The new system, presented here, has some interesting features which makes it a valuable tool to manage ethnic music archives: the fingerprints are rather robust against pitch shift, tempo changes, several synthetic audio effects, and reversal of the audio. When only part of the audio is used to generate a fingerprint, the system keeps working but retrieval performance degrades.