The dataset is currently empty. Upload or create new data files. Then, you will be able to explore them in the Dataset Viewer.

Description

Mandarin Chinese Seperated Track Spontaneous Dialogue Paralanguage Annotated Speech Synthesis Corpus, with a free dialogue style. Given a topic, the speaker can express themselves, and in each conversation, each person's audio is stored in their own separate WAV file. Professional linguists have annotated 16 types of paralanguage annotations, text annotations, timestamps, and other information to accurately match the research and development needs of speech synthesis.

For more details, please refer to the link: https://www.nexdata.ai/datasets/tts/1620?source=huggingface

Specifications

Format

48kHz, 24 bit, wav, mono channel

Recording condition

Recording studio

Content category

Spontaneous dialogue in given topics

Speaker

294 people (Non-Professional Voice Actors) in total, gender balanced (144 females and 150 males), 18~60 years old

Features of annotation

16 kinds of paralanguage annotation; text transcription; speaker ID, special symbol

Recording device

Microphone

Language

Mandarin Chinese

Country

China(CHN)

Language(Region) Code

zh-CN

Accuracy

Character Accuracy Rate 99%

Licensing Information

Commercial License

Downloads last month
7