Loading datasets...
License:
NOODL-1.0
Steward:
Institute of African Digital HumanitiesDataset ID:
cmqf6c04x06gel207o5vjtni7
Task: TTS
Release Date: 6/15/2026
Format: MP3, TSV
Size: 64.35 MB
Share
Ngiemboon-TTS-Dataset is a scripted speech dataset dedicated to the documentation and technological development of Ngiemboon (ISO 639-3: nnh), a Grassfields Bantu language spoken in the Bamboutos Division of the West Region of Cameroon. The dataset was compiled in the framework of the Mozilla Data Collective initiative (2026), as a supplement to the Common Voice Scripted Speech 25.0 – Ngiemboon dataset (https://mozilladatacollective.com/datasets/cmn1qf3al00xzo107byg0pine). The dataset comprises 995 high-quality MP3 audio recordings of Ngiemboon sentences read by a native speaker across 10 recording sessions, together with per-session sentence-to-audio mapping files enabling precise alignment between textual and acoustic data. Sentences were drawn from a scripted speech prompt list and read in a controlled environment. The transcription of all sentences follows the General Alphabet of Cameroon's Languages (AGLC; French acronym: Alphabet Général des Langues Camerounaises), the reference standard for Cameroonian national languages. The Ngiemboon orthography employed in this dataset is distinguished by an extended vowel inventory — including the open-mid front unrounded vowel ɛ, the open-mid back rounded vowel ɔ, the high central rounded vowel ʉ, the high central unrounded vowel ɨ (barred i), and the close front rounded vowel ÿ — as well as a series of labialized consonants written by appending ẅ (w with diaeresis) to the base consonant (e.g., kẅ, gẅ, sẅ, zẅ, tsẅ), a multi-register tone-marking system combining level (acute, grave) and contour (caron, circumflex) diacritics applied to vowels and syllabic nasals, and the modifier letter apostrophe (ʼ) for glottal closure. The parallel availability of AGLC-transcribed text and aligned speech makes the dataset suitable for a wide range of applications, including text-to-speech (TTS) synthesis, automatic speech recognition (ASR), forced alignment, pronunciation modelling, and language learning tools. It also directly supports efforts to standardise and normalise the digital representation of Ngiemboon in language technology contexts.
Licensing
Nwulite Obodo Open Data Licence 1.0 (NOODL-1.0)
https://licensingafricandatasets.com/nwulite-obodo-licenseRestrictions/Special Constraints
By downloading this dataset, you agree: - To use it for research and scientific use only - That you will not re-host or re-share this dataset
Forbidden Usage
You agree not to use the data for: determining the identity of any speaker in the dataset; attempting to clone any voice or train models that imitate any speaker in this dataset; Generative AI; reproduction; duplication; modification; augmentation; copying; distribution; transmission; display; sale; transfer; publication or creation of derivative works without the explicit permission of the legal owner of the dataset.
Intended Use
(a) Speech-related tasks: - Text-to-speech (TTS) synthesis: The dataset provides clean sentence–audio pairs from multiple recording sessions and is directly suited for training, fine-tuning, and evaluating speech synthesis models for Ngiemboon. The availability of AGLC-transcribed sentences with aligned audio enables the development of TTS systems capable of producing natural-sounding Ngiemboon speech. - Automatic speech recognition (ASR): Audio–text alignment enables the training and evaluation of speech recognition models for Ngiemboon. The per-session structure and controlled recording conditions make the dataset suitable for building and evaluating ASR models for this under-resourced language. - Speech–text alignment / forced alignment benchmarking: Fine-grained audio–text pairing provides ground truth for evaluating phoneme- or word-level aligners adapted to Grassfields Bantu languages of the Bamboutos area. - Pronunciation modelling: The AGLC-transcribed sentences, combined with aligned audio, provide a resource for developing grapheme-to-phoneme (G2P) models and pronunciation lexicons for Ngiemboon. (b) Linguistic and lexicographic tasks: - Phonological analysis: The dataset enables systematic study of the phonological and tonal system of Ngiemboon, including its multi-register tone system, extended vowel inventory (ɛ, ɔ, ʉ, ɨ, ÿ), labialized consonant series (kẅ, gẅ, sẅ, tsẅ, etc.), and the behaviour of syllabic nasals (ḿ, ń, ǹ) as tone-bearing units. - Orthographic standardisation and normalisation: The dataset can serve as a reference corpus for evaluating and training text normalisation models aligned with the AGLC standard for Ngiemboon. - Language documentation: The dataset contributes to the digital documentation of Ngiemboon scripted speech in AGLC orthography, supporting efforts to extend the digital presence of this Grassfields Bantu language of the West Region of Cameroon.
Ngiemboon (also written Ngyɛmbɔŋ) is a Grassfields Bantu language belonging to the Niger-Congo phylum, classified within the Mbam-Nkam branch. It is spoken in the Bamboutos Division of the West Region of Cameroon. Despite its sociolinguistic significance within Cameroon, Ngiemboon remains substantially underrepresented in language technology resources.
According to the Administrative Atlas of Cameroon's Languages (Breton & Bikia Fohtung 1991), Ngiemboon comprises the following dialects:
Balatchi
Bamoungong
Bangang
Batcham
The writing system used for the transcription of Ngiemboon in this dataset is the General Alphabet of Cameroon's Languages (AGLC). The AGLC provides a phonologically motivated orthographic standard for Cameroonian national languages and serves as the reference framework for Ngiemboon literacy materials.
The vowel system attested in the dataset includes the following oral vowels:
a, e, i, o, u, ɛ, ɔ, ʉ, ɨ, ÿ
Where:
ɛ (epsilon): open-mid front unrounded vowel
ɔ (open-o): open-mid back rounded vowel
ʉ (barred u): high central rounded vowel
ɨ (barred i): high central unrounded vowel
ÿ (y-diaeresis): close front rounded vowel
Long vowels are represented by vowel doubling (e.g., aa, ɛɛ, ɔɔ, uu, ii).
The consonant inventory reflected in the dataset includes simple, digraph, and labialized consonants:
b, c, d, f, g, h, j, k, l, m, n, p, s, sh, t, ts, v, w, y, z, ŋ
Labialized consonants are formed by appending ẅ (w with diaeresis) to the base consonant or cluster:
kẅ: labialized velar stop
gẅ: labialized voiced velar stop
sẅ: labialized alveolar fricative
zẅ: labialized voiced alveolar fricative
tsẅ: labialized alveolar affricate
nzẅ, nkẅ: labialized nasal-consonant clusters
Special symbols:
ŋ (eng): velar nasal consonant
ẅ (w with diaeresis): labialization marker, appended to consonants
ʼ (modifier letter apostrophe): glottal stop / glottal closure marker
Ngiemboon attests syllabic nasal consonants that function as tone-bearing units. The following tone-marked syllabic nasals are represented in the dataset:
ḿ (m with acute): syllabic bilabial nasal, high tone
ń (n with acute): syllabic alveolar nasal, high tone
ǹ (n with grave): syllabic alveolar nasal, low tone
Ngiemboon is a tonal language with multiple contrastive pitch levels and contour tones. The dataset employs systematic tone marking on vowels and syllabic nasals in accordance with the AGLC convention. The following diacritics are attested in the dataset:
Level tones:
High tone (H): acute accent — á, é, í, ó, ú, ɛ́, ɔ́, ʉ́, ɨ́, ÿ́
Low tone (L): grave accent — à, è, ì, ò, ù, ɛ̀, ɔ̀, ʉ̀, ɨ̀, ÿ̀
Contour tones:
Falling tone (HL): circumflex — â, ê, î, ô, û, ɛ̂, ɔ̂, ʉ̂
Rising tone (LH): caron — ǎ, ě, ǐ, ǒ, ǔ, ɛ̌, ɔ̌, ÿ̌
Mid tone is generally left unmarked in the Ngiemboon AGLC orthography.
The dataset was compiled from scripted speech prompt lists read by a native speaker. Sentences were selected to provide broad phonological coverage of Ngiemboon and were transcribed in accordance with the AGLC orthographic standard.
The dataset represents scripted speech in Ngiemboon, covering a broad range of everyday sentence types drawn from a general-purpose TTS/ASR prompt list. All utterances are scripted rather than spontaneous.
Total audio duration: 7,280 seconds (02h 01m 20s), distributed across 995 MP3 audio clips in 10 recording sessions.
The dataset is organised into 10 recording sessions:
Session tts_dataset_nnh_01: 100 clips (14m 46s)
Session tts_dataset_nnh_02: 100 clips (16m 34s)
Session tts_dataset_nnh_03: 100 clips (10m 28s)
Session tts_dataset_nnh_04: 100 clips (11m 06s)
Session tts_dataset_nnh_05: 100 clips (14m 07s)
Session tts_dataset_nnh_06: 100 clips (13m 24s)
Session tts_dataset_nnh_07: 100 clips (10m 59s)
Session tts_dataset_nnh_08: 100 clips (09m 32s)
Session tts_dataset_nnh_09: 100 clips (07m 56s)
Session tts_dataset_nnh_10: 95 clips (12m 24s)
Each session folder contains:
MP3 audio clips
One per-session sentence-to-audio mapping file (mapping.tsv), with 4 columns
#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
#attempts: number of recording attempts before acceptance
| audio file | sentence (Ngiemboon, AGLC) |
|---|---|
| d2d7174fe45cf83b8b89c49daad332aa.mp3 | Menkẅɛ̌ ndá lezyéen fʉʼ ntsèm |
| ca2914db43399709f83e39928006a0b3.mp3 | Lepǔ ḿbyág mentí fÿàg ntsèm jʉ̀' ntsèm, à fʉ̀' ntsém. |
| 8ef3a54eaebd558ae8752a565abc1e2b.mp3 | Atèmte ncwò lɔg ńdiŋ nkʉ̀a lezíŋ lɔgɔ ńgwɔ́ tsɔ̌ tàʼ na meliŋé menkʉ̀a mezíŋ métá |
| a5ed566a13cb0cc0b039b6d1d4ae1c4a.mp3 | ḿbiŋ ńgʉa na menkàŋ, ńtsɔ́ʼ ntɔɔn saŋtí mítà 30 |
| 5c71748b491175f5477d39d686822ed9.mp3 | Lezíŋ tá pa Lôŋtsyě ée le wɔ̌? |