On the Use of Self-Supervised Representation Learning for Speaker Diarization and Separation

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Baroudi, Séverin, Bredin, Hervé, Razik, Joseph, Marxer, Ricard
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912770796027904
author Baroudi, Séverin
Bredin, Hervé
Razik, Joseph
Marxer, Ricard
author_facet Baroudi, Séverin
Bredin, Hervé
Razik, Joseph
Marxer, Ricard
contents Self-supervised speech models such as wav2vec2.0 and WavLM have been shown to significantly improve the performance of many downstream speech tasks, especially in low-resource settings, over the past few years. Despite this, evaluations on tasks such as Speaker Diarization and Speech Separation remain limited. This paper investigates the quality of recent self-supervised speech representations on these two speaker identity-related tasks, highlighting gaps in the current literature that stem from limitations in the existing benchmarks, particularly the lack of diversity in evaluation datasets and variety in downstream systems associated to both diarization and separation.
format Preprint
id arxiv_https___arxiv_org_abs_2512_15224
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle On the Use of Self-Supervised Representation Learning for Speaker Diarization and Separation
Baroudi, Séverin
Bredin, Hervé
Razik, Joseph
Marxer, Ricard
Audio and Speech Processing
Self-supervised speech models such as wav2vec2.0 and WavLM have been shown to significantly improve the performance of many downstream speech tasks, especially in low-resource settings, over the past few years. Despite this, evaluations on tasks such as Speaker Diarization and Speech Separation remain limited. This paper investigates the quality of recent self-supervised speech representations on these two speaker identity-related tasks, highlighting gaps in the current literature that stem from limitations in the existing benchmarks, particularly the lack of diversity in evaluation datasets and variety in downstream systems associated to both diarization and separation.
title On the Use of Self-Supervised Representation Learning for Speaker Diarization and Separation
topic Audio and Speech Processing
url https://arxiv.org/abs/2512.15224