AISHELL-1

LDC2018S14

Introduction

AISHELL-1, Linguistic Data Consortium (LDC) Catalog Number LDC2018S14 and ISBN 1-58563-866-8, was developed by Beijing Shell Shell Technology Co., Ltd. It contains approximately 520 hours of Chinese Mandarin speech from 400 speakers recorded simultaneously on three different devices with associated transcripts.

The goal of the collection was to support speech recognition system development in 11 domains, including smart homes, autonomous driving, entertainment, finance and science and technology. Participants read 500 sentences covering the domains; sentences were chosen for their speech and phonetic characteristics.

Speakers were recruited from different accent areas across China, including North, South and Yue-Gui-Min regions. There were 214 female speakers and 186 male speakers, constituting 53% and 47% of the database, respectively. Additional demographic information about the participants is included in this release.

Data

Speech was recorded in a quiet indoor environment on a high fidelity microphone and two mobile phones (Android and iOS). All speech is presented as 16-bit flac compressed wav files; the microphone speech sample rate is 44.1kHz and the phone speech sample rate is 16kHz. Each speech file ranges from approximately 1 second to 14 seconds in length.

Transcripts are stored as UTF-8 encoded plain text files and are not time-aligned.

Directory Structure

Please see file.tbl for a complete file list as well as checksums for this publication.

Updates

Additional information, updates, bug fixes may be available in the LDC catalog entry for this corpus at LDC2018S14.

Content Copyright

Portions © 2018 Beijing Shell Shell Technology Co., Ltd., © 2018 Trustees of the University of Pennsylvania