License:
NOODL-1.0
Steward:
Institute of African Digital HumanitiesDataset ID:
cmqj75zev01i8nq07v1qzjctb
Task: OTH
Release Date: 6/18/2026
Format: WAV, XML
Size: 366.79 MB
Share
EMAC-Dataset-CAR is a digitised ethnomusicological audio collection dedicated to the documentation and preservation of traditional vocal music from the Central African Republic (CAR), formerly known as the Central African Empire (Empire centrafricain). The dataset was constituted in the framework of the Ethnomusicologie d'Afrique Centrale (EMAC) project, initiated by CERDOTOLA (Centre de Recherche et de Documentation sur les Traditions Orales et pour le Développement des Langues Africaines) following the founding meeting held in Sangmélima, Cameroon, from 17 to 22 September 1979. The dataset comprises 5 high-quality WAV audio recordings representing two distinct oral music genres: the chante-fable and the musique épique (epic music / mélopée). Two chante-fable sessions are included, of which one (CHF_002) is distributed across three consecutive audio files. The recordings were originally captured on magnetic tape and digitised in 2015 at CERDOTOLA in Yaoundé, Cameroon, under the coordination of Dr. Emmanuel Ngue Um, using Adobe Audition 3.0. At the time of the Sangmélima meeting (1979), the CAR (then Empire centrafricain) was represented by Jean-Marie Toninga, who described a rich and varied musical culture insufficiently studied at national level. The Service des Musées held approximately 1,500 tape recordings of traditional music collected during provincial tours, though conservation challenges were significant. EMAC-Dataset-CAR provides a digitally preserved subset of that heritage. As an ethnomusicological rather than a linguistic dataset, it does not provide transcriptions or sentence-level alignments.
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, education, and cultural heritage purposes only - That you will not re-host or re-share this dataset without the explicit permission of the legal owner
Forbidden Usage
You agree not to use the data for: determining the identity of any performer in the dataset; attempting to clone any voice or train models that imitate any performer 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) Ethnomusicological and cultural heritage tasks: - Ethnomusicological research: The dataset provides authentic recordings of CAR chante-fable and epic vocal traditions and is suited for the study of vocal style, narrative structure, melodic contour, and genre conventions in Central African oral music. - Cultural heritage archiving: The dataset contributes to the digital preservation of intangible cultural heritage from the Central African Republic at a time when national archival infrastructure faced significant challenges, making this collection particularly valuable. - Cross-cultural comparative analysis: Used alongside the other six EMAC country datasets, these recordings enable comparative studies of shared and divergent oral music traditions across Central Africa. (b) Computational and technological tasks: - Computational ethnomusicology: Audio signal analysis, genre classification, melodic transcription, and acoustic modelling using traditional Central African vocal music as data. Ethical Review Process All recordings were collected with the knowledge and participation of the performers and their affiliated institutions. The dataset is made available for research and educational use in the spirit of the EMAC project's mission of documenting and promoting Central African musical heritage.
The Central African Republic harbours a rich mosaic of oral music traditions associated with its diverse ethnic communities. CAR traditional music is characterised by complex polyrhythmic structures, call-and-response vocal patterns, and the intertwining of music with oral narrative in genres such as the chante-fable. The musical traditions represented in this dataset belong to the sphere of community oral literature and ceremonial performance, where song functions both as an aesthetic expression and as a vehicle for the transmission of cultural knowledge. The EMAC corpus for CAR was constituted during a period when the national ethnomusicological infrastructure was still nascent, making these recordings among the earliest systematic collections of CAR traditional vocal music.
Chante-fable (CHF): A hybrid narrative genre combining sung passages with spoken prose. Two sessions are represented: CHF_001 (1 audio file) and CHF_002 (3 audio files, parts 1–3 of 3).
Musique épique / Mélopée (MEP): A genre of extended melodic vocal performance with an epic or lyrical character. One session is represented (MEP_001, 1 audio file).
The musical traditions represented in this dataset are transmitted exclusively through oral and performative channels. No written notation system is employed for these genres in their traditional transmission context. The chante-fable is communicated through performance events where singers and storytellers alternate or combine melodic and spoken registers. The mélopée / musique épique is a melodically sustained form requiring specialised vocal technique, learned through extended community exposure. The EMAC recordings constitute elicited performances in a controlled environment and serve as primary archival documentation of these traditions.
The recordings in this dataset were collected by researchers and technicians of the EMAC project in the Central African Republic, approximately between 1979 and 1982. The project engaged the Service des Musées of the CAR as its principal institutional partner. Jean-Marie Toninga, maître d'éducation musicale and cultural affairs officer, represented the CAR at the Sangmélima founding meeting. The original recordings were stored on magnetic tape at CERDOTOLA and digitised in 2015 under the coordination of Dr. Emmanuel Ngue Um.
The recordings represent two genres of traditional vocal music from the Central African Republic: chante-fables (combining narrative and song) and an extended epic melodic performance (musique épique / mélopée). All performances were elicited in a controlled recording environment at the request of the EMAC project.
Total audio duration: approximately 57 minutes (3,408 seconds), distributed across 5 WAV audio files. IMDI 3.03 XML metadata files: 1 (corpus-level only), in IMDI/ subfolder.
The dataset comprises 5 WAV audio files organised as follows:
EMAC_CF_CHF_001-1-1.wav — Chante-fable 001 (1 part of 1)
EMAC_CF_CHF_002-1-3.wav — Chante-fable 002 (part 1 of 3)
EMAC_CF_CHF_002-2-3.wav — Chante-fable 002 (part 2 of 3)
EMAC_CF_CHF_002-3-3.wav — Chante-fable 002 (part 3 of 3)
EMAC_CF_MEP_001-1-1.wav — Musique épique 001 (1 part of 1)
File naming convention: EMAC_{CC}{GENRE}{NNN}-{part}-{total}.wav
CC = ISO country code (CF = Central African Republic)
GENRE = genre code (CHF, MEP)
NNN = zero-padded session number
part / total = part index and total number of parts for multi-file sessions
IMDI metadata files (IMDI/ subfolder):
EMAC_Centrafrique.imdi — Corpus-level (Ethnomusicologie de la République centrafricaine)
| Audio file | Genre | Session | Parts |
|---|---|---|---|
| EMAC_CF_CHF_001-1-1.wav | Chante-fable | 001 | 1/1 |
| EMAC_CF_CHF_002-1-3.wav | Chante-fable | 002 | 1/3 |
| EMAC_CF_CHF_002-2-3.wav | Chante-fable | 002 | 2/3 |
| EMAC_CF_CHF_002-3-3.wav | Chante-fable | 002 | 3/3 |
| EMAC_CF_MEP_001-1-1.wav | Musique épique / Mélopée | 001 | 1/1 |