Loading datasets...
License:
CC-BY-NC-SA-4.0
Steward:
CommunityDataset ID:
cmtimj5tm0045mj07qrjo74gc
Release Date: 9/1/2026
Format: WEBM, TSV
Size: 129.43 MB
Share
This dataset is a collection of natural Javanese speech texts designed for text-to-speech (TTS) research and development, with a focus on Javanese Ngoko and variations across everyday spoken contexts in Java Island, Indonesia. The texts consist primarily of personal narratives, memories, daily experiences, hobbies, family stories, travel experiences, education, work, food, entertainment, art, culture, and other familiar topics. Many samples use natural conversational Javanese with occasional Indonesian words and expressions, reflecting the way Javanese is commonly used in informal communication and code-switching contexts. The dataset also includes vocabulary and expressions related to different aspects of everyday life, allowing TTS systems to learn varied sentence structures, word combinations, pacing, and pronunciation patterns. The narratives are intentionally written in an accessible, informal style rather than formal literary Javanese, making them suitable for modeling natural-sounding speech and evaluating how well TTS systems handle longer, continuous utterances in Javanese. The collection is particularly intended to support research involving Javanese language generation, pronunciation, speech synthesis, and the representation of dialectal and informal language characteristics.
Licensing
Creative Commons Attribution Non Commercial Share Alike 4.0 International (CC-BY-NC-SA-4.0)
https://spdx.org/licenses/CC-BY-NC-SA-4.0.htmlRestrictions/Special Constraints
This dataset is not for commercial use. This dataset is intended for research and education purposes only, and proper citation is a must. Any use of this dataset requires written permission by submitting an access request with a clear statement of your intended use.
Forbidden Usage
Using this dataset to clone or imitate specific speakers/individuals and train chatbots or large language models is strictly forbidden. Re-uploading or redistributing this dataset is prohibited.
Ethical Review
This dataset was created by developing texts in informal Javanese Ngoko representing multiple regional dialects, with occasional code-mixing in Indonesian and English. The texts were read and recorded by speakers through the hosting platform (https://sabre-2.onrender.com/), and the resulting audio recordings were compiled into a comprehensive dataset for Javanese speech and TTS research.
Intended Use
This dataset is intended to support the preservation, development, and dissemination of the Javanese language, particularly informal Javanese Ngoko and its regional dialectal variations, for text-to-speech (TTS) development, linguistic research, language technology, and educational purposes.
This dataset contains Javanese speech representing a mix of Central Javanese and East Javanese dialects, Indonesia. These two geographical dialects reflect language used in everyday activities and conversations. The dataset also includes natural code-mixing between Javanese, Indonesian, and English.
Created by the owner of the dataset, considered as (linguists) and native speakers.
The dataset covers a wide range of everyday and personal domains, including family and relationships, education and school experiences, work and career, travel and leisure, food and dining, hobbies and entertainment, art and culture, music, pets and animals, technology and social media, health and lifestyle, sports and exercise, personal memories, and everyday observations. The topics are presented through informal personal narratives and conversational-style texts, providing varied vocabulary and contexts for natural Javanese speech.
Around 2 hours 15 minutes
Public open access is permitted with proper attribution and citation of the dataset source.
Surti Syafiurahmi. (2026). TTS Javanese - Multidialect [dataset]. Mozilla Data Collective. URL [link dataset].
Audio file name, text
“Yen aku liburan, aku wajib banget sing jenenge kulineran.”
“Aku nduwe impian sederhana, yaiku pengen nduwe omah ning pinggiran kutha gedhe”
“Areke asline rodok atos dan teges, makane dee diwedeni lan diajeni banget.”
“Trainer e yo nerangake jelas lan suasana kelas dadi luwih santai. “
“Dadi nek ditakoni selera musikku opo, aku kadang bingung jawabnya. “
Latin alphabet (A–Z), Arabic numerals (0–9)