Datasets:
Formats:
parquet
Size:
1M - 10M
Tags:
indian speech
indian languages
synthetic speech
deepfake
audio deepfake detection
indian deepfake detection
License:
Update README.md
Browse files
README.md
CHANGED
|
@@ -62,7 +62,6 @@ dataset_info:
|
|
| 62 |
|
| 63 |
# IndicSynth
|
| 64 |
*A Large-Scale Multilingual Synthetic Speech Dataset for Low-Resource Indian Languages*
|
| 65 |
-
|
| 66 |
**π Outstanding Paper Award, ACL 2025**
|
| 67 |
|
| 68 |
---
|
|
@@ -71,9 +70,9 @@ dataset_info:
|
|
| 71 |
|
| 72 |
**IndicSynth** is a novel multilingual synthetic speech dataset designed to advance multilingual **audio deepfake detection (ADD)** and **anti-spoofing** research. It covers **12 low-resource Indian languages** and provides both **mimicry** and **diversity** subsets.
|
| 73 |
|
| 74 |
-
|
| 75 |
-
|
| 76 |
-
|
| 77 |
---
|
| 78 |
## π Dataset Structure
|
| 79 |
|
|
@@ -108,8 +107,8 @@ Synthetic data was generated using:
|
|
| 108 |
| vits | TTS | Yes | No |
|
| 109 |
| freevc24 | VC | No | No |
|
| 110 |
|
| 111 |
-
|
| 112 |
-
|
| 113 |
For more details, please see the Table 1 and Section 3 of our paper: https://aclanthology.org/2025.acl-long.1070.pdf
|
| 114 |
---
|
| 115 |
|
|
@@ -163,7 +162,6 @@ doi = {10.1609/aaai.v37i11.26521}
|
|
| 163 |
For questions or feedback, please feel free to reach out at divyas@iiitd.ac.in.
|
| 164 |
|
| 165 |
## π Acknowledgments
|
| 166 |
-
|
| 167 |
-
|
| 168 |
-
> π€ HuggingFace for dataset hosting support
|
| 169 |
|
|
|
|
| 62 |
|
| 63 |
# IndicSynth
|
| 64 |
*A Large-Scale Multilingual Synthetic Speech Dataset for Low-Resource Indian Languages*
|
|
|
|
| 65 |
**π Outstanding Paper Award, ACL 2025**
|
| 66 |
|
| 67 |
---
|
|
|
|
| 70 |
|
| 71 |
**IndicSynth** is a novel multilingual synthetic speech dataset designed to advance multilingual **audio deepfake detection (ADD)** and **anti-spoofing** research. It covers **12 low-resource Indian languages** and provides both **mimicry** and **diversity** subsets.
|
| 72 |
|
| 73 |
+
- 4,000+ hours of synthetic audio
|
| 74 |
+
- 12 languages: Bengali, Gujarati, Hindi, Kannada, Malayalam, Marathi, Odia, Punjabi, Sanskrit, Tamil, Telugu, Urdu
|
| 75 |
+
- Useful for ADD, speaker verification (SV), and bias studies
|
| 76 |
---
|
| 77 |
## π Dataset Structure
|
| 78 |
|
|
|
|
| 107 |
| vits | TTS | Yes | No |
|
| 108 |
| freevc24 | VC | No | No |
|
| 109 |
|
| 110 |
+
- **Mimicry subset**: For anti-spoofing research
|
| 111 |
+
- **Diversity subset**: Contains diverse set of realistic synthetic voices for multilingual audio deepfake detection research
|
| 112 |
For more details, please see the Table 1 and Section 3 of our paper: https://aclanthology.org/2025.acl-long.1070.pdf
|
| 113 |
---
|
| 114 |
|
|
|
|
| 162 |
For questions or feedback, please feel free to reach out at divyas@iiitd.ac.in.
|
| 163 |
|
| 164 |
## π Acknowledgments
|
| 165 |
+
- π ACL Diversity & Inclusion Subsidy for enabling in-person presentation at ACL 2025
|
| 166 |
+
- π€ HuggingFace for dataset hosting support
|
|
|
|
| 167 |
|