AISHELL6-whisper: A Chinese Mandarin Audio-visual Whisper Speech Dataset with Speech Recognition Baselines

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Li, Cancan, Su, Fei, Liu, Juan, Bu, Hui, Wan, Yulong, Suo, Hongbin, Li, Ming
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908563445645312
author Li, Cancan
Su, Fei
Liu, Juan
Bu, Hui
Wan, Yulong
Suo, Hongbin
Li, Ming
author_facet Li, Cancan
Su, Fei
Liu, Juan
Bu, Hui
Wan, Yulong
Suo, Hongbin
Li, Ming
contents Whisper speech recognition is crucial not only for ensuring privacy in sensitive communications but also for providing a critical communication bridge for patients under vocal restraint and enabling discrete interaction in noise-sensitive environments. The development of Chinese mandarin audio-visual whisper speech recognition is hindered by the lack of large-scale datasets. We present AISHELL6-Whisper, a large-scale open-source audio-visual whisper speech dataset, featuring 30 hours each of whisper speech and parallel normal speech, with synchronized frontal facial videos. Moreover, we propose an audio-visual speech recognition (AVSR) baseline based on the Whisper-Flamingo framework, which integrates a parallel training strategy to align embeddings across speech types, and employs a projection layer to adapt to whisper speech's spectral properties. The model achieves a Character Error Rate (CER) of 4.13% for whisper speech and 1.11% for normal speech in the test set of our dataset, and establishes new state-of-the-art results on the wTIMIT benchmark. The dataset and the AVSR baseline codes are open-sourced at https://zutm.github.io/AISHELL6-Whisper.
format Preprint
id arxiv_https___arxiv_org_abs_2509_23833
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle AISHELL6-whisper: A Chinese Mandarin Audio-visual Whisper Speech Dataset with Speech Recognition Baselines
Li, Cancan
Su, Fei
Liu, Juan
Bu, Hui
Wan, Yulong
Suo, Hongbin
Li, Ming
Audio and Speech Processing
Computer Vision and Pattern Recognition
Multimedia
Sound
Whisper speech recognition is crucial not only for ensuring privacy in sensitive communications but also for providing a critical communication bridge for patients under vocal restraint and enabling discrete interaction in noise-sensitive environments. The development of Chinese mandarin audio-visual whisper speech recognition is hindered by the lack of large-scale datasets. We present AISHELL6-Whisper, a large-scale open-source audio-visual whisper speech dataset, featuring 30 hours each of whisper speech and parallel normal speech, with synchronized frontal facial videos. Moreover, we propose an audio-visual speech recognition (AVSR) baseline based on the Whisper-Flamingo framework, which integrates a parallel training strategy to align embeddings across speech types, and employs a projection layer to adapt to whisper speech's spectral properties. The model achieves a Character Error Rate (CER) of 4.13% for whisper speech and 1.11% for normal speech in the test set of our dataset, and establishes new state-of-the-art results on the wTIMIT benchmark. The dataset and the AVSR baseline codes are open-sourced at https://zutm.github.io/AISHELL6-Whisper.
title AISHELL6-whisper: A Chinese Mandarin Audio-visual Whisper Speech Dataset with Speech Recognition Baselines
topic Audio and Speech Processing
Computer Vision and Pattern Recognition
Multimedia
Sound
url https://arxiv.org/abs/2509.23833