somu9/iisc_mono_hindi_female

---
language:
- hi
license: cc-by-4.0
task_categories:
- text-to-speech
tags:
- hindi
- tts
- single-speaker
- female
- iisc
- syspin
- studio-quality
size_categories:
- 10K<n<100K
dataset_info:
features:
- name: audio
dtype: audio
- name: text
dtype: string
- name: domain
dtype: string
- name: speaker_id
dtype: string
- name: language
dtype: string
splits:
- name: train
num_examples: 21662
- name: test
num_examples: 396
---

# IISc Mono Hindi Female

Studio-quality single-speaker Hindi female TTS dataset from the SYSPIN project by Indian Institute of Science (IISc), Bengaluru.

## Dataset Description

| Property | Value |
|---|---|
| **Source** | [IISc SYSPIN Project](https://syspin.iisc.ac.in/) |
| **Speaker** | Single professional female voice artist (42 yrs, 21 yrs experience) |
| **Language** | Hindi (hi) |
| **Total Duration** | 54 hours 54 minutes 44 seconds |
| **Utterances** | 22,058 (train: 21,662 / test: 396 EVAL domain) |
| **Audio** | 48kHz, 24-bit, mono, embedded in parquet |
| **Recording** | Neumann TLM-103 microphone, professional studio, ~40dB SNR |
| **Domains** | Agriculture, Books, Education, Evaluation, Finance, General, Health, Others, Politics, Weather |

## Domain Distribution

| Domain | Hours | Sentences | Description |
|---|---|---|---|
| BOOK | 23:03:20 | 8,358 | Books |
| OTHE | 7:21:25 | 3,081 | Others |
| GENE | 6:00:43 | 2,540 | General |
| EDUC | 5:04:10 | 2,060 | Education |
| WEAT | 4:22:58 | 1,873 | Weather |
| POLI | 2:57:51 | 1,235 | Politics |
| AGRI | 1:48:18 | 848 | Agriculture |
| HEAL | 1:48:31 | 835 | Health |
| FINA | 1:46:45 | 832 | Finance |
| EVAL | 0:40:38 | 396 | Evaluation (test set) |

## Fields

| Column | Type | Description |
|---|---|---|
| `audio` | Audio (48kHz) | Speech waveform |
| `text` | string | Hindi transcription (Devanagari) |
| `domain` | string | Content domain (BOOK, GENE, etc.) |
| `speaker_id` | string | `hindi_female_spk001` |
| `language` | string | `hi` |

## Splits

- **train**: 21,662 utterances (all domains except EVAL)
- **test**: 396 utterances (EVAL domain — recommended by creators for TTS evaluation)

## Usage

```python
from datasets import load_dataset

ds = load_dataset("somu9/iisc_mono_hindi_female", split="train")

# Listen to first sample
print(ds[0]["text"])
audio = ds[0]["audio"] # {"array": np.array, "sampling_rate": 48000}
```

## Speaker Metadata

- **Language:** Hindi
- **Gender:** Female
- **Age:** 42
- **Experience:** 21 Years
- **Languages known:** Hindi, English, Tamil
- **Mother tongue:** Hindi

## Recording Setup

- **Microphone:** Neumann TLM-103
- **Environment:** Professional studio
- **Conditions:** Studio quality at ~40dB SNR

## License

This dataset is released under the [CC-BY-4.0 license](https://creativecommons.org/licenses/by/4.0/legalcode.en).

TTS data created under SYSPIN project by Indian Institute of Science, Bengaluru. The copyright in the TTS data belongs to Indian Institute of Science, Bengaluru.

## Acknowledgments

We extend our heartfelt gratitude to the talented voice artist whose contributions were fundamental to this project's success. We are particularly grateful to the project of German Development Cooperation "FAIR Forward - AI for All" for their financial support in developing this TTS corpus, and Bhashini AI Solutions Private Limited for their financial support for part of the corpus beyond 44 hours for every voice artist in developing this TTS corpus.

## Citation

```bibtex
@misc{SYSPIN_S1.0_Corpus,
Title = {SYSPIN_S1.0 Corpus - A TTS Corpus of 900+ hours in nine Indian Languages},
Authors = {Abhayjeet Et al.},
Year = {2025}
}
```

## Contact

SPIRE Lab, EE Dept., IISc, Bengaluru
Email: [e-mail]