Publicly Available Emotional Speech Dataset (ESD) for Speech Synthesis and Voice Conversion

This is the GitHub page for publicly available emotional speech dataset (ESD) for speech synthesis and voice conversion.

This dataset contains 350 parallel utterances spoken by 10 native Mandarin speakers, and 10 English speakers with 5 emotional states (neutral, happy, angry, sad and surprise). The transcripts are provided.

License

Please complete a License and send it to zhoukun@u.nus.edu. This database can only be used for research purpose.

Download

The whole dataset can be downloaded here: Google Drive

Note: For users from mainland China, google drive cannot be accessed. Please send an email to zhoukun@u.nus.edu. We will share this dataset with you through other platforms.

Paper

If you are using this dataset, please kindly cite this paper:

@inproceedings{zhou2021seen,
  title={Seen and unseen emotional style transfer for voice conversion with a new emotional speech dataset},
  author={Zhou, Kun and Sisman, Berrak and Liu, Rui and Li, Haizhou},
  booktitle={ICASSP 2021-2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)},
  pages={920--924},
  year={2021},
  organization={IEEE}
}
@article{zhou2021emotional,
title = {Emotional voice conversion: Theory, databases and ESD},
journal = {Speech Communication},
volume = {137},
pages = {1-18},
year = {2022},
issn = {0167-6393}
}