Saved in:
Bibliographic Details
Main Authors: Luo, Longjie, Li, Lin, Hong, Qingyang
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2505.24450
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915354461077504
author Luo, Longjie
Li, Lin
Hong, Qingyang
author_facet Luo, Longjie
Li, Lin
Hong, Qingyang
contents Due to the lack of target speech annotations in real-recorded far-field conversational datasets, speech enhancement (SE) models are typically trained on simulated data. However, the trained models often perform poorly in real-world conditions, hindering their application in far-field speech recognition. To address the issue, we (a) propose direct sound estimation (DSE) to estimate the oracle direct sound of real-recorded data for SE; and (b) present a novel pseudo-supervised learning method, SuPseudo, which leverages DSE-estimates as pseudo-labels and enables SE models to directly learn from and adapt to real-recorded data, thereby improving their generalization capability. Furthermore, an SE model called FARNET is designed to fully utilize SuPseudo. Experiments on the MISP2023 corpus demonstrate the effectiveness of SuPseudo, and our system significantly outperforms the previous state-of-the-art. A demo of our method can be found at https://EeLLJ.github.io/SuPseudo/.
format Preprint
id arxiv_https___arxiv_org_abs_2505_24450
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle SuPseudo: A Pseudo-supervised Learning Method for Neural Speech Enhancement in Far-field Speech Recognition
Luo, Longjie
Li, Lin
Hong, Qingyang
Sound
Audio and Speech Processing
Due to the lack of target speech annotations in real-recorded far-field conversational datasets, speech enhancement (SE) models are typically trained on simulated data. However, the trained models often perform poorly in real-world conditions, hindering their application in far-field speech recognition. To address the issue, we (a) propose direct sound estimation (DSE) to estimate the oracle direct sound of real-recorded data for SE; and (b) present a novel pseudo-supervised learning method, SuPseudo, which leverages DSE-estimates as pseudo-labels and enables SE models to directly learn from and adapt to real-recorded data, thereby improving their generalization capability. Furthermore, an SE model called FARNET is designed to fully utilize SuPseudo. Experiments on the MISP2023 corpus demonstrate the effectiveness of SuPseudo, and our system significantly outperforms the previous state-of-the-art. A demo of our method can be found at https://EeLLJ.github.io/SuPseudo/.
title SuPseudo: A Pseudo-supervised Learning Method for Neural Speech Enhancement in Far-field Speech Recognition
topic Sound
Audio and Speech Processing
url https://arxiv.org/abs/2505.24450