---
license: cc-by-nc-4.0
task_categories:
- audio-classification
- automatic-speech-recognition
language:
- zh
- en
tags:
- emotion
- speech
- voice
size_categories:
- 10K<n<100K
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
dataset_info:
  features:
  - name: audio
    dtype: audio
  - name: transcript
    dtype: string
  - name: emotion
    dtype: string
  - name: speaker_id
    dtype: string
  - name: gender
    dtype: string
  - name: language
    dtype: string
  splits:
  - name: train
    num_bytes: 3353221499.0
    num_examples: 35000
  download_size: 3145534453
  dataset_size: 3353221499.0
---

# Emotional Speech Dataset (ESD)

The Emotional Speech Dataset (ESD) is a multilingual emotional speech corpus containing parallel recordings in English and Chinese across 5 emotions.

## Dataset Details

- **Total samples**: 35,000
- **Speakers**: 20 (10 Chinese, 10 English)
- **Emotions**: anger, happiness, neutral, sadness, surprise (7,000 each)
- **Languages**: Chinese (zh), English (en) - 17,500 each
- **Gender**: 10 male, 10 female speakers

## Dataset Structure

| Column | Description |
|--------|-------------|
| `audio` | Audio waveform (WAV) |
| `transcript` | Text transcription |
| `emotion` | anger, happiness, neutral, sadness, surprise |
| `speaker_id` | Speaker identifier (0001-0020) |
| `gender` | male / female |
| `language` | zh (Chinese) / en (English) |

## Usage

```python
from datasets import load_dataset

dataset = load_dataset("jspaulsen/esd")
```

## Citation

```bibtex
@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}
}
```