Combination of Linear Prediction and Phase Decomposition for Glottal Source Analysis on Voiced Speech

TL;DR

This paper introduces a novel glottal source analysis method combining linear prediction and phase decomposition, improving estimation accuracy without requiring glottal model fitting, suitable for real speech processing.

Contribution

It presents a new approach that integrates all-pole odd-order linear prediction with phase decomposition for enhanced glottal source estimation.

Findings

Improved separation of source and filter in glottal flow estimation

No need for glottal model fitting, increasing flexibility

Validated on real speech data showing performance gains

Abstract

Some glottal analysis approaches based upon linear prediction or complex cepstrum approaches have been proved to be effective to estimate glottal source from real speech utterances. We propose a new approach employing both an all-pole odd-order linear prediction to provide a coarse estimation and phase decomposition based causality/anti-causality separation to generate further refinements. The obtained measures show that this method improved performance in terms of reducing source-filter separation in estimation of glottal flow pulses (GFP). No glottal model fitting is required by this method, thus it has wide and flexible adaptation to retain fidelity of speakers's vocal features with computationally affordable resource. The method is evaluated on real speech utterances to validate it.