Ewondo-ASR-Dataset

Language

Ewondo is a Narrow Bantu language belonging to the Beti-Fang group of the Benue-Congo branch. It is indigenous to a population located primarily in the Centre Region of Cameroon, with significant speech communities in the South and East Regions. Ewondo also functions as a vehicular language in those regions and has given rise to a creolised variety known as Mongo Ewondo. Ethnologue estimates the number of speakers at approximately 900,000, including first- and second-language users. Despite its relatively large speaker base, Ewondo remains significantly underrepresented in language technology resources.

Variants

The glossonym 'Ewondo' designates a set of closely related linguistic varieties whose speakers may or may not identify with the label, depending on geographical, social and pragmatic factors. In the framework of the Atlas Linguistique du Cameroun (ALCAM) project, Ewondo is listed as one of the major micro-languages of the Beti-Fang macro-language, alongside Fang, Bulu, Ntumu and Eton. Varieties such as Yezoum (Haut-Nyong Division), Yanda and Moog-Ebanda are considered sub-varieties of Ewondo in standard classifications, though this classification is not always accepted by their speakers.

The present dataset represents speakers of the Ewondo variety as spoken in the Yaoundé area (Centre Region), recruited at the École Normale Supérieure de Yaoundé.

Writing System

The writing system used for the transcription of Ewondo in this dataset is the General Alphabet of Cameroon's Languages (AGLC), as adopted by the Ministry of Basic Education of Cameroon and regularly updated by the Direction de la Promotion des Langues Nationales. The AGLC provides a phonologically motivated orthographic standard for Cameroonian national languages and serves as the reference framework for Ewondo literacy materials, including those produced by the Catholic and Protestant missionary traditions that have subsequently aligned with this standard.

1. Vowels

The vowel system attested in the dataset includes the following oral vowels:

a, e, ə, i, o, u, ɔ

Long vowels are represented by vowel doubling (e.g. aa, ee, oo).

2. Consonants

The consonant inventory reflected in the dataset includes simple, prenasalized and digraph consonants:

b, d, dz, f, g, h, k, l, m, mb, mv, n, nd, ng, nk, nz, ny, ŋ, p, s, t, ts, v, w, y, z

Special symbols: ə (mid central vowel), ŋ (velar nasal)

3. Tone system

Ewondo is a tonal language with lexical and grammatical contrastive tones. The dataset employs systematic tone marking on vowels in accordance with the AGLC convention:

• High tone (H): á, é, ə́, í, ó, ɔ́, ú

• Low tone (L): à, è, ə̀, ì, ò, ɔ̀, ù

• Falling tone (HL): â, ê, ə̂, î, ô, ɔ̂, û

• Rising tone (LH): ǎ, ě, ə̌, ǐ, ǒ, ɔ̌, ǔ

Unmarked vowels represent tonally neutral or contextually determined syllables. This explicit tone notation distinguishes the present dataset from the Common Voice Scripted Speech 25.0 – Ewondo resource, in which tone diacritics are systematically absent.

Source

The dataset was compiled from scripted speech prompt lists read by native speakers of Ewondo in recording sessions held at the École Normale Supérieure de Yaoundé in 2026, in the framework of the Mozilla Data Collective project. Sentences were selected to provide broad phonological coverage of Ewondo and were transcribed in accordance with the AGLC orthographic standard, with full tone marking.

Domain

The dataset represents scripted speech in Ewondo, covering a broad range of everyday sentence types drawn from a general-purpose ASR/TTS prompt list. All utterances are scripted rather than spontaneous.

Size

Total audio duration: 11,457 seconds (03:10:57), distributed across 1,781 MP3 audio clips in 19 recording sessions contributed by 16 native speakers of Ewondo. Total uncompressed dataset size: approximately [X] MB.

Structure

The dataset comprises:

1. 1,781 MP3 audio clips read by 16 native speakers of Ewondo, with a total duration of 11,457 seconds (03:10:57), distributed across 19 recording sessions:

   • Session ewo_01: 97 clips (12m 03s)

   • Session ewo_02: 99 clips (14m 41s)

   • Session ewo_03: 99 clips (07m 52s)

   • Session ewo_04: 10 clips (01m 03s)

   • Session ewo_07: 99 clips (08m 51s)

   • Session ewo_08: 98 clips (08m 03s)

   • Session ewo_09: 96 clips (07m 09s)

   • Session ewo_13: 96 clips (10m 23s)

   • Session ewo_13-1: 97 clips (17m 29s)

   • Session ewo_14: 99 clips (12m 23s)

   • Session ewo_14-1: 99 clips (19m 09s)

   • Session ewo_15: 99 clips (08m 46s)

   • Session ewo_18: 100 clips (12m 04s)

   • Session ewo_18-1: 100 clips (10m 52s)

   • Session ewo_19: 96 clips (10m 21s)

   • Session ewo_22: 99 clips (05m 49s)

   • Session ewo_30: 99 clips (06m 55s)

   • Session ewo_31: 99 clips (08m 46s)

   • Session ewo_32: 100 clips (08m 08s)

2. Nineteen per-session sentence-to-audio mapping files (mapping.tsv), each with 4 columns.

Description of columns (mapping.tsv)

• #audio_filename: filename of the audio clip (MP3)

• #key: unique hash identifier of the recording

• #sentence: sentence text as read by the speaker, transcribed in AGLC orthography with tone marking

• #attempts: number of recording attempts before acceptance

Sample