The CHiME-7 UDASE task: Unsupervised domain adaptation for conversational speech enhancement - Inria - Institut national de recherche en sciences et technologies du numérique Access content directly
Conference Papers Year : 2023

The CHiME-7 UDASE task: Unsupervised domain adaptation for conversational speech enhancement

Abstract

Supervised speech enhancement models are trained using artificially generated mixtures of clean speech and noise signals, which may not match real-world recording conditions at test time. This mismatch can lead to poor performance if the test domain significantly differs from the synthetic training domain. This paper introduces the unsupervised domain adaptation for conversational speech enhancement (UDASE) task of the 7th CHiME challenge. This task aims to leverage real-world noisy speech recordings from the target domain for unsupervised domain adaptation of speech enhancement models. The target domain corresponds to the multi-speaker reverberant conversational speech recordings of the CHiME-5 dataset, for which the ground-truth clean speech reference is unavailable. Given a CHiME-5 recording, the task is to estimate the clean, potentially multi-speaker, reverberant speech, removing the additive background noise. We discuss the motivation for the CHiME-7 UDASE task and describe the data, the task, and the baseline system.
Fichier principal
Vignette du fichier
main.pdf (163.94 Ko) Télécharger le fichier
Origin Files produced by the author(s)

Dates and versions

hal-04156930 , version 1 (10-07-2023)
hal-04156930 , version 2 (02-10-2023)

Licence

Identifiers

Cite

Simon Leglaive, Léonie Borne, Efthymios Tzinis, Mostafa Sadeghi, Matthieu Fraticelli, et al.. The CHiME-7 UDASE task: Unsupervised domain adaptation for conversational speech enhancement. 7th International Workshop on Speech Processing in Everyday Environments (CHiME), Aug 2023, Dublin, Ireland. ⟨10.21437/CHiME.2023-2⟩. ⟨hal-04156930v2⟩
104 View
71 Download

Altmetric

Share

Gmail Mastodon Facebook X LinkedIn More