Loading datasets...
Release Date: 9/17/2026
Format: MP3
Size: 109.74 MB
A collection of spontaneous responses to questions in Basaa (Basaa).
Restrictions/Special Constraints
None provided.
Forbidden Usage
It is forbidden to attempt to determine the identity of speakers in the Common Voice datasets. It is forbidden to re-host or re-share this dataset.
Intended Use
This dataset is intended to be used for training and evaluating automatic speech recognition (ASR) models. It may also be used for applications relating to computer-aided language learning (CALL) and language or heritage revitalisation.
bas)This datasheet is for sps-corpus-5.0-2026-09-11 of the Mozilla Common Voice Spontaneous Speech dataset for Basaa [Basaa - bas]. The dataset contains 773 clips representing 5.37 hours of recorded speech (5.06 hours validated) from 11 speakers.
Basaa is a narrow Bantu language spoken across a geographical area spanning three administrative regions in Cameroon: the Centre, Littoral and South regions. It is estimated that there are currently around 600,000–700,000 speakers. This figure includes different varieties, as well as diasporic populations who identify as Basaa speakers.
The vitality of the Basaa language is stable (Ethnologue online). However, intergenerational transmission of Basaa is increasingly threatened among parents aged 50 and under, particularly in urban areas.
Although Basaa is taught in schools, this does not significantly impact the vitality of the language, mainly due to the current pedagogical approach, which relies on rule-based and descriptivist teaching methods.
The glossonym 'Basaa' is a generic term that encompasses a range of varieties, the speakers of which may identify with the 'Basaa' label to varying degrees, depending on a complex set of geographical, social, political, situational and pragmatic factors. Whether a language variant is considered Basaa depends greatly on the perspective of the person 'telling the story'. Some of the most commonly acknowledged varieties of Basaa include:
Other varieties, such as Ndonga, Mbaa (also known as Mbay-Bati) and Hijuk, may also be classified as Basaa. However, as previously mentioned, not everyone agrees on this classification.
The dataset clips are categorised by transcription status and training-set assignment. The following tables summarise the distribution.
| Bucket | Clips | % |
|---|---|---|
| Transcribed & Validated | 734 | 95.0% |
| Transcribed & Pending | 38 | 4.9% |
| Not transcribed | 1 | 0.1% |
| Bucket | Clips | % |
|---|---|---|
| Train | 203 | 26.3% |
| Dev | 279 | 36.1% |
| Test | 252 | 32.6% |
| Unassigned | 39 | 5.0% |
Training split coverage: 734 of 734 transcribed & validated clips (100.0%)
| Bucket | Clips | % |
|---|---|---|
| Validated | 734 | 95.1% |
| Pending | 38 | 4.9% |
| Edited | 500 | 64.8% |
Basaa has several competing writing norms. The most widely used are the Catholic missionary orthography, the Prostestant missionary orthography, and a version of the General Alphabet of Cameroonian Languages that was adapted to Basaa.
This dataset is mostly based on the Protestant missionary's orthography, with minor alterations concerning, for example, the signaling of b as implosive [ɓ]. For example, m'bôñ "cassava" vs mbôñ "poison". Other alterations includes the signaling of the n- prefix followed by the y symbol, to distinguish it from the complex symbol ny. For example, nyo "mouth" vs a n'yo "he stole palm wine from the palm trunk".
There follows a randomly selected sample of questions used in the corpus.
There follows a randomly selected sample of transcribed responses from the corpus.
Each row of a tsv file represents a single audio clip, and contains the following information:
client_id - hashed UUID of a given useraudio_id - numeric id for audio fileaudio_file - audio file nameduration_ms - duration of audio in millisecondsprompt_id - numeric id for promptprompt - question for usertranscription - transcription of the audio responsevotes - number of people that who approved a given transcriptage - age of the speaker1gender - gender of the speaker1language - language namesplit - for data modelling, which subset of the data does this clip pertain tochar_per_sec - how many characters of transcription per second of audioquality_tags - some automated assessment of the transcription--audio pair, separated by |
transcription-length - character per second under 3 characters per secondspeech-rate - characters per second over 30 characters per secondshort-audio - audio length under 2 secondslong-audio - audio length over 5 minutesnon-allowed-script - transcription contains characters from a writing system not associated with the languagemixed-script-words - a single word contains characters from multiple writing systemsmixed-script-transcription - transcription spans multiple writing systems, but each word consistently uses only oneThe compilation of this dataset was made possible thanks to grant awarded by the Mozilla Foundation
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.
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