The following report introduces an ongoing project to produce a pitch-controllable speech synthesis model for Nigerian Pidgin, a widely-spoken but poorly-resourced language of West Africa. The first dedicated Nigerian Pidgin TTS model, NaijaTTS, is intended to provide a tool for linguists wishing to study the prosody of this language in an experimental context. In this paper, we present the key objectives of our model, the progress made thus far, and the challenges involved in building a TTS model for this low-resource language.
Cite as: Strickland, E., Aubakirova, D., Doncenco, D., Torres, D., Evrard, M. (2023) NaijaTTS: A pitch-controllable TTS model for Nigerian Pidgin. Proc. 12th ISCA Speech Synthesis Workshop (SSW2023), 248-249
@inproceedings{strickland23_ssw,
author={Emmett Strickland and Dana Aubakirova and Dorin Doncenco and Diego Torres and Marc Evrard},
title={{NaijaTTS: A pitch-controllable TTS model for Nigerian Pidgin}},
year=2023,
booktitle={Proc. 12th ISCA Speech Synthesis Workshop (SSW2023)},
pages={248--249}
}