Summary: A Chinese Mandarin telephone speech corpus published by Beijing DataTang Technology Co., Ltd.
License: Attribution-NonCommercial-NoDerivatives 4.0 International (CC BY-NC-ND 4.0)
About this resource:
The corpus is 200 hours long, which is recorded by Android-system mobile phones (16kHz, 16 bit) and iOS-system mobile phones (16kHz, 16 bit). 600 speakers from different accent areas in China are invited to participate in the recording, which is conducted in a quiet indoor environment or an environment contain background noise that does not affect speech recognition. Gender and age of the participants are evenly distributed. The language materials of the corpus are designed phoneme balanced oral sentences. The manual transcription accuracy is larger than 98% for each sentence.
The contents and the corresponding descriptions of the corpus include:
- audio files: wav format speech data
- transcriptions: manual annotations
- metadata: data label
The corpus aims to support researchers in speech recognition, machine translation, voiceprint recognition, and other speech-related fields. Therefore, the corpus is totally free for academic use.
Please cite the corpus as “AIDataTang_200zh, Free Chinese Mandarin Telephone Corpus”.
The corpus is a subset of a much bigger data set which was recorded in the same environment as this open source data. Please visit our website www.datatang.com for more details.
- Email: firstname.lastname@example.org
- Tel: 400-650-6137
External URL: https://www.datatang.com/webfront/datatang_laboratory_03.html DataTangAIlab