README.md

---
dataset_info:
- config_name: Human_3
  features:
  - name: original_wav
    dtype: audio
  - name: normalized_wav
    dtype: audio
  - name: speaker_id
    dtype: string
  - name: transcription
    dtype: string
  splits:
  - name: train
    num_bytes: 554661238.0
    num_examples: 907
  - name: test
    num_bytes: 65929372.0
    num_examples: 100
  download_size: 601830149
  dataset_size: 620590610.0
- config_name: Synthetic
  features:
  - name: original_wav
    dtype: audio
  - name: normalized_wav
    dtype: audio
  - name: speaker_id
    dtype: string
  - name: transcription
    dtype: string
  splits:
  - name: train
    num_bytes: 16298373533.296
    num_examples: 20056
  - name: test
    num_bytes: 1735872207.904
    num_examples: 2048
  download_size: 15976481639
  dataset_size: 18034245741.2
- config_name: default
  features:
  - name: original_wav
    dtype: audio
  - name: normalized_wav
    dtype: audio
  - name: speaker_id
    dtype: string
  - name: transcription
    dtype: string
  splits:
  - name: train
    num_bytes: 17035340160.108
    num_examples: 20963
  - name: test
    num_bytes: 1820617200.704
    num_examples: 2148
  download_size: 16584190993
  dataset_size: 18855957360.812
configs:
- config_name: Human_3
  data_files:
  - split: train
    path: Human_3/train-*
  - split: test
    path: Human_3/test-*
- config_name: Synthetic
  data_files:
  - split: train
    path: Synthetic/train-*
  - split: test
    path: Synthetic/test-*
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: test
    path: data/test-*
license: cc-by-4.0
task_categories:
- text-to-speech
- text-to-audio
- automatic-speech-recognition
language:
- ar
pretty_name: arvoice
size_categories:
- 10K<n<100K
---

<h2 align="center">
  <b>ArVoice: A Multi-Speaker Dataset for Arabic Speech Synthesis</b>
</h2>

<div style="font-size: 16px; text-align: justify;">
    <p>ArVoice is a multi-speaker Modern Standard Arabic (MSA) speech corpus with fully diacritized transcriptions, intended  for multi-speaker speech synthesis, and can be useful for other tasks such as speech-based diacritic restoration, voice conversion, and deepfake detection. <br> 
      ArVoice comprises: (1) professionally recorded audio by 2 male and 2 female voice artists from diacritized transcripts, (2) professionally recorded audio by 1 male and 1 female voice artists from undiacritized transcripts, (3) a modified subset of the
Arabic Speech Corpus, and (4) synthesized speech using commercial TTS systems. The complete corpus consists of a total of 83.52 hours of speech across 11 voices; around 10 hours consist of human voices from 7 speakers. <br> <br>
      <strong> This repo consists of only Parts (3), ASC subset, and (4) synthetic subset </strong>; to access the main subset, part (1,2), which consists of six professional speakers, <a href="/"> please sign this agreement</a> and email it to us. 
      <br><br> If you use the dataset or transcriptions provided in Huggingface, <u>place cite the paper</u>. 
</p>
</div>

Usage Example

```python
df = load_dataset("MBZUAI/ArVoice", "Human_3")  #data_dir options: Human_3, Synthetic,
print(df)

DatasetDict({
    train: Dataset({
        features: ['original_wav', 'normalized_wav', 'speaker_id', 'transcription'],
        num_rows: 907
    })
    test: Dataset({
        features: ['original_wav', 'normalized_wav', 'speaker_id', 'transcription'],
        num_rows: 100
    })
})

```


Data Statistics
| Type      | Part            | Gender     | Speaker Origin | Duration (hrs) | Text Source                  |
|-----------|-----------------|------------|----------------|----------------|------------------------------|
| Human     | ArVoice Part 1  | M          | Egypt          | 1.17           | Tashkeela                    |
|           |                 | F          | Jordan         | 1.45           |                              |
|           |                 | M          | Egypt          | 1.58           |                              |
|           |                 | F          | Morocco        | 1.23           |                              |
|           | ArVoice Part 2  | M          | Palestine      | 0.93           | Khaleej                      |
|           |                 | F          | Egypt          | 0.93           |                              |
|           | ArVoice Part 3  | M          | Syria          | 2.69           | ASC                          |
| Synthetic | ArVoice Part 4  | 2×M, 2×F   | -              | 73.5           | Tashkeela, Khaleej, ASC      |


License: [https://creativecommons.org/licenses/by/4.0/](https://creativecommons.org/licenses/by/4.0/)

### Citation

```
@misc{toyin2025arvoicemultispeakerdatasetarabic,
      title={ArVoice: A Multi-Speaker Dataset for Arabic Speech Synthesis}, 
      author={Hawau Olamide Toyin and Rufael Marew and Humaid Alblooshi and Samar M. Magdy and Hanan Aldarmaki},
      year={2025},
      eprint={2505.20506},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2505.20506}, 
}
```