Loading datasets...
Release Date: 7/14/2026
Format: WAV, CSV
Size: 2.50 GB
Share
This dataset, curated by Digital Divide Data (DDD), provides approximately 12 hours of high-quality Singaporean Tamil speech recordings paired with verified text transcripts. The collection consists of unique read-speech utterances designed to support the development of Text-to-Speech (TTS) systems and other low-resource language technologies. All transcripts have undergone manual verification to ensure high linguistic accuracy. The recordings represent a balanced mix of genders and various age groups to promote diversity and help reduce bias in downstream AI models. The dataset is intended primarily for Text-to-Speech (TTS) model development and evaluation, and may also be used for speech synthesis research, pronunciation modelling, and linguistic research on the Singaporean Tamil language.
Licensing
Creative Commons Attribution 4.0 International (CC-BY-4.0)
https://spdx.org/licenses/CC-BY-4.0.htmlRestrictions/Special Constraints
For research and scientific use only Redistribution or re-hosting of this dataset is not permitted without written permission from Digital Divide Data.
Forbidden Usage
You agree not to attempt to determine the identity of any speaker in the dataset. It is forbidden to use this dataset for voice cloning, biometric identification, surveillance, or the creation of synthetic voices mimicking speakers.
Ethical Review
All participants gave informed consent prior to recording. Speakers were briefed about the use of their audio for research and could withdraw at any time. No personal names or identifying information are included in the dataset.
Intended Use
This dataset is intended for developing and evaluating Automatic Speech Recognition (ASR) systems, studying cross-dialect variation, training self-supervised audio models, and supporting linguistic research on Singaporean Tamil.
TECHNICAL DATASHEET SINGAPOREAN TAMIL TEXT-TO-SPEECH (TTS) DATASET (12 HOURS)
Digital Divide Data
CONSIDERATIONS 1.1 Forbidden Usage Do not attempt to identify or re-identify any speaker. Do not link voices to real people or external datasets. Do not use the dataset for surveillance or biometric identification. Do not re-host, re-share, or redistribute the dataset without written permission. 1.2 Privacy and Ethics Demographic metadata is intended for aggregate analysis only. Individual-level inference is strictly discouraged.
DATASET OVERVIEW 2.1 Dataset Summary
A curated collection of high-quality read speech recordings from native Singaporean Tamil speakers. The recordings were collected through controlled voice collection sessions and are paired with verified text transcripts. The dataset is intended to support the development of Text-to-Speech (TTS) systems and other low-resource language technologies.
The dataset contains approximately 12 hours of recorded speech.
2.2 Primary Intended Task Text-to-Speech (TTS) model training and evaluation. 2.3 Secondary Possible Uses Speech synthesis research Pronunciation modelling Low-resource language technology development Educational and linguistic research 3. LANGUAGE
Language: Singaporean Tamil
DATA CREATION Native Singaporean Tamil speakers were recruited. Sentence prompts were prepared by linguistically competent contributors. Speech recordings were manually reviewed and paired with verified transcripts.
COLLECTION ENVIRONMENT
Speech was collected during controlled recording sessions to promote consistency in recording quality.
DATA SOURCES Generated specifically for this project. Read speech recordings paired with verified text transcripts. Original recordings are stored in WAV (.wav) format.
TECHNICAL SPECIFICATIONS Property Specification Dataset Type Text-to-Speech (TTS) Dataset Duration Approximately 12 hours Audio Format WAV (.wav) Speech Type Read speech Language Singaporean Tamil Recording Environment Controlled recording sessions Transcript Format CSV metadata with verified text transcripts Dataset Organization Audio files with accompanying metadata Metadata Format Comma-Separated Values (CSV)
DATASET STRUCTURE
The dataset contains:
An audio/ directory containing WAV recordings. A metadata.csv file containing transcripts and metadata for each recording. 9. METADATA DESCRIPTION Field Description file_name Audio filename language Recording language age Speaker age gender Speaker gender theme Recording theme domain Recording domain sub_theme Topic category sentence_ID Unique sentence identifier sentence Verified transcript 10. INTENDED USE
This dataset is intended for:
Text-to-Speech (TTS) model development and evaluation Speech synthesis research Academic and educational research Low-resource language technology development Pronunciation and linguistic analysis
This dataset is not intended for:
Speaker identification Voice biometric systems Surveillance applications Individual profiling or re-identification 11. CITATION
Digital Divide Data. (2026). Singaporean Tamil Text-to-Speech (TTS) Dataset (12 Hours).
