AISHELL-4 多通道中文会议语音数据库

AISHELL-4 Open Source Mandarin Multi-channel Meeting Speech Corpus

      AISHELL-4是一个通过麦克风阵列实录的八通道中文普通话会议场景语音数据集。该数据集共包含211场会议,每场会议4至8人,数据集共120小时左右。该数据集旨在促进实际应用场景下多说话人处理的研究。AISHELL-4数据包括了实际会议场景下各种重要特性,例如停顿、重叠、说话人轮转、噪声等。同时数据集提供了准确的音字转写文本及时间戳信息,方便研究者进行诸如前端处理、语音识别、说话人分割等单独任务,并可以进行联合优化。

 

The AISHELL-4 is a sizable real-recorded Mandarin speech dataset collected by 8-channel circular microphone array for speech processing in conference scenario. The dataset consists of 211 recorded meeting sessions, each containing 4 to 8 speakers, with a total length of 120 hours. This dataset aims to bride the advanced research on multi-speaker processing and the practical application scenario in three aspects. With real recorded meetings, AISHELL-4 provides realistic acoustics and rich natural speech characteristics in conversation such as short pause, speech overlap, quick speaker turn, noise, etc. Meanwhile, the accurate transcription and speaker voice activity are provided for each meeting in AISHELL-4. This allows the researchers to explore different aspects in meeting processing, ranging from individual tasks such as speech front-end processing, speech recognition and speaker diarization, to multi-modality modeling and joint optimization of relevant tasks. We also release a PyTorch-based training and evaluation framework as baseline system to promote reproducible research in this field.

The setup of the recording environment.

120 小时 丨 120 Hours

Speech front-end processing

Speech Recognition

Speaker Diarization

开源系统 

Open Source

211 场会议 丨 211 Meeting Sessions

10个 会议室 丨 10 Meeting Rooms

60 人 丨 60 Speakers

AISHELL-4 is part of the AISHELL-ASR0055 Corpus

 

The setup of the recording environment.

29 个会议室 丨 29 Meeting Rooms

956 场会议 丨 956 Meeting Sessions

666 小时/单通道 丨 666 Hours/Single Channel

261 人 丨 261 Speakers

Non-Open Source

 

       数据使用申请                Company:bd@aishelldata.com      

 

Service  Application          Academic Institution:aishell.foundation@gmail.com     

数据下载

 

论 文

 

License: Apache License v.2.0

Dataset
arxiv

基线系统

 

相关课程

 

语音识别实战

 

了解课程详情
Recipe

样例下载

 

Sample

数据介绍

 

Readme