Speaker verification in mismatch training and testing conditions

TL;DR

This paper investigates the robustness of various parameterizations in speaker verification across diverse recording conditions using a new database, demonstrating that combining parameterizations enhances performance.

Contribution

It provides an extensive analysis of parameterization robustness in speaker verification with a new database covering multiple recording scenarios.

Findings

Combining parameterizations improves robustness across scenarios.

New database includes telephonic, microphonic, and multilingual recordings.

Covariance matrices are effective in text-independent speaker verification.

Abstract

This paper presents an exhaustive study about the robustness of several parameterizations, with a new database specially acquired for the purpose of a speaker recognition application. This database includes the following variations: different recording sessions (including telephonic and microphonic recordings), recording rooms, and languages (it has been obtained from a bilingual set of speakers). This study has been performed with covariance matrices in a text independent speaker verification application. It reveals that the combination of several parameterizations can improve the robustness in all the scenarios.