Improved multiple birdsong tracking with distribution derivative method and Markov renewal process clustering

TL;DR

This paper introduces an enhanced spectrogram technique based on the distribution derivative method, combined with a Markov renewal process model, to improve the segregation and tracking of multiple birdsongs with rapid pitch modulations.

Contribution

It presents a novel spectrogram representation and an integrated tracking approach that together improve the accuracy of birdsong separation in complex audio mixtures.

Findings

Enhanced segregation performance with the new spectrogram representation

Effective tracking of vocalization patterns using Markov renewal process

Improved recognition of rapid pitch modulations in birdsongs

Abstract

Segregating an audio mixture containing multiple simultaneous bird sounds is a challenging task. However, birdsong often contains rapid pitch modulations, and these modulations carry information which may be of use in automatic recognition. In this paper we demonstrate that an improved spectrogram representation, based on the distribution derivative method, leads to improved performance of a segregation algorithm which uses a Markov renewal process model to track vocalisation patterns consisting of singing and silences.