An Investigation of Distribution Alignment in Multi-Genre Speaker Recognition

TL;DR

This paper studies the effectiveness of distribution alignment methods in multi-genre speaker recognition, revealing that current approaches have limited and inconsistent improvements across different genre distributions.

Contribution

It provides a comprehensive analysis of mainstream distribution alignment techniques on multi-genre data, highlighting their limitations and the need for more effective solutions.

Findings

Within-between distribution alignment (WBDA) performs relatively better

None of the methods consistently improve performance across all cases

Distribution alignment alone may not fully solve multi-genre recognition challenges

Abstract

Multi-genre speaker recognition is becoming increasingly popular due to its ability to better represent the complexities of real-world applications. However, a major challenge is the significant shift in the distribution of speaker vectors across different genres. While distribution alignment is a common approach to address this challenge, previous studies have mainly focused on aligning a source domain with a target domain, and the performance of multi-genre data is unknown. This paper presents a comprehensive study of mainstream distribution alignment methods on multi-genre data, where multiple distributions need to be aligned. We analyze various methods both qualitatively and quantitatively. Our experiments on the CN-Celeb dataset show that within-between distribution alignment (WBDA) performs relatively better. However, we also found that none of the investigated methods consistently improved performance in all test cases. This suggests that solely aligning the distributions of speaker vectors may not fully address the challenges posed by multi-genre speaker recognition. Further investigation is necessary to develop a more comprehensive solution.