Common Voice Spontaneous Speech 3.0 - Betawi

bew — Betawi (bew)

This datasheet is for sps-corpus-3.0-2026-03-09 of the Mozilla Common Voice Spontaneous Speech dataset for Betawi [bew - bew]. The dataset contains 1336 clips representing 10.49 hours of recorded speech (9.78 hours validated) from 21 speakers.

Language

Betawi language originally belongs to Austronesian language with a full name of Melayu-Betawi. This language is considered as one of Malay dialects, but historically it grew together with other major languages, such as Arabic, Hokkien, Sundanese, Javanese, and Malay in Sumatra - a tiny portion with Portuguese and Dutch. The language vitality status is Endangered according to https://www.ethnologue.com/language/bew/. At the moment, Indonesian standard and English in general influence the native speakers, allowing code switching and code mixing happens in a spontaneous speech. The specific variation of this dataset is Betawi Ora or Betawi Pinggiran (Peripheral Betawi), taken from several locations of Bekasi District/City, West Java Province, Indonesia. This variation is unique in terms of geo-politics: language is spoken only in the community, but it is not taught at school. Instead, the community is taught Sundanese language, which is dominated in West Java Province in general.

Data splits for modelling

The dataset clips are categorised by transcription status and training-set assignment. The following tables summarise the distribution.

Audio clips

Training splits

Training split coverage: 1,270 of 1,270 transcribed & validated clips (100.0%)

Transcriptions

The transcription system uses general Latin script, but involves allophone variants of three /e/, these are /é/, /è/, and /e/.

• Prompts: 199

• Duration: 39396960[ms]

• Avg. Transcription Len: 292

• Avg. Duration: 28.28[s]

• Valid Duration: 36776.84[s]

• Total hours: 10.94[h]

• Valid hours: 10.22[h]

Transcription status

Writing system

Historically, this language used Pegon, Arabic script, but now Latin is adapted.The writing system in this dataset uses general Latin script, but involves allophone variants of three /e/, these are /é/, /è/, and /e/.

Symbol table

a b c d é è ȇ e f g h i j k l m n o p q r s t u v w y z 

Samples

Questions

There follows a randomly selected sample of questions used in the corpus.

1. *Kalo Ente bisé keluar negeri, Ente pengen liburan ke mané? Dan kenapé di sono? *

2. Kalo Ente, pernah belajar kesenian apé? Dan, kenapé belajar tuh kesenian?

3. Apé ajé nyang bikin Ente suké buat pake sosial media?

4. Ceritain pengalaman Ente liburan bareng keluargé!

5. Dari segi pendidikan, apé pendidikan terakhir di masyarakat sekitar?

Responses

There follows a randomly selected sample of transcribed responses from the corpus.

1. Pendidikan di daerah gua tuh di masyarakat sekitar ya biasanyé nya si sampé SMA ya untuk pendidikan gitu tapi untuk jenjang kuliah itu tergantung dari ekonomi setiap keluarga tersebut gitu, kaya di daerah gua gitu ga yang semuanya kuliah, paling ya cuma sampe SMA aja sih.

2. Jam bérapé biasanyé enté tidur?

3. Kalo untuk kesenian secara pribadi dan dasar buat saya begitu pentingnya. Karna kesenian itu adalah hasil daripada ééé mengutip sejarah masa lalu juga untuk mengikuti jejak-jejak daripada orang yang terdahulu yang menciptakan kesenian tersebut. Makanya wajib dong perlu. Kenapa kita bina dan kita bangun kesenian dan kita ciptakan dan kita kembangkan? Karna untuk mencontohkan nanti anak-anak kité, masa depannya, setelah kité udah tua renta. Nah maka anak-anak kité harus udah tau untuk cara mengenal tentang seni maupun kesenian. Makanya indahnya orang-orang yang mempunyai seni dan menciptakan kesenian supaya memberikan inspirasi dan eee apa namanya model-model buat anak-anak masa depannya biar supaya berkembang di era globalisasi maupun digital. Gitu menurut ayé sih. Makanya kalo buat ayé sih, seni maupun kesenian itu penting kita galakan untuk kita kembangkan untuk memberikan kepada anak-anak kita maupun mencontohkan supaya lebih mengenal kesenian maupun seni, gitu.

4. *Jadi kalo ayémêh kalo mau liburan itu jauh jauh arih ayé ngerencanain ngadain liburan seperti kalo itu paling juga liburan sekolah anak-anak deh paling juga maennyé itu ke laut berenang *

5. *karna kan kalo sosial media kan lebih cepet lebih canggih gituloh lebih enak cepet untuk nyampein berita apapun *

Recommended post-processing

(1) Observe the non-linguistic aspects, such as filler, (2) Make sure your machine learning does not differ the suprasegmental aspect, like intonation which does not change the word and its meaning.

Fields

Each row of a tsv file represents a single audio clip, and contains the following information:

• client_id - hashed UUID of a given user

• audio_id - numeric id for audio file

• audio_file - audio file name

• duration_ms - duration of audio in milliseconds

• prompt_id - numeric id for prompt

• prompt - question for user

• transcription - transcription of the audio response

• votes - number of people that who approved a given transcript

• age - age of the speaker1

• gender - gender of the speaker1

• language - language name

• split - for data modelling, which subset of the data does this clip pertain to

• char_per_sec - how many characters of transcription per second of audio

• quality_tags - some automated assessment of the transcription--audio pair, separated by |

  • transcription-length - character per second under 3 characters per second

  • speech-rate - characters per second over 30 characters per second

  • short-audio - audio length under 2 seconds

  • long-audio - audio length over 5 minutes

Get involved

Community links

• https://referensi.data.kemendikdasmen.go.id/budayakita/wbtb/objek/AA000491

• https://petabahasa.kemdikbud.go.id/ (Web of peta bahasa does not consider Betawi language is part of Indonesia, particularly in Jakarta and West Jawa Province.

• Common Voice translators on Pontoon

• Common Voice Communities

Discussions

• Common Voice on Matrix

• Common Voice on Discourse

• Common Voice on Discord

• Common Voice on Telegram

Contribute

• Contribute questions

• Validate questions

• Answer questions

• Transcribe recordings

• Validate transcriptions

Acknowledgements

Datasheet authors

• Yacub Fahmilda <yacub.fahmilda@gmail.com>

• Riska Legistari Febri <riskalegistari25@gmail.com>

Funding

This dataset was partially funded by the Open Multilingual Speech Fund managed by Mozilla Common Voice.

Licence

This dataset is released under the Creative Commons Zero (CC-0) licence. By downloading this data you agree to not determine the identity of speakers in the dataset.

Footnotes

1. For a full list of age, gender, and accent options, see the demographics spec. These will only be reported if the speaker opted in to provide that information. ↩ ↩2