A comparative study of several parameterizations for speaker recognition

TL;DR

This paper systematically compares various parameterizations for speaker recognition, demonstrating that combining multiple parameterizations enhances robustness across different recording conditions, languages, and mismatch scenarios.

Contribution

It provides an exhaustive analysis of parameterizations and evaluates two methods, showing how their combination improves speaker recognition robustness.

Findings

Combining parameterizations improves robustness in speaker verification and identification.

Vector quantization and covariance matrices with sphericity measure are effective methods.

Robustness is maintained across different recording sessions, microphones, and languages.

Abstract

This paper presents an exhaustive study about the robustness of several parameterizations, in speaker verification and identification tasks. We have studied several mismatch conditions: different recording sessions, microphones, and different languages (it has been obtained from a bilingual set of speakers). This study reveals that the combination of several parameterizations can improve the robustness in all the scenarios for both tasks, identification and verification. In addition, two different methods have been evaluated: vector quantization, and covariance matrices with an arithmetic-harmonic sphericity measure.