ISCA Workshop on Multilingual Speech and Language Processing (MULTILING 2006) Center for Language and Speech Technology, Stellenbosch University, Stellenbosch, South Africa April 9-11, 2006

Computational Models of Prosody in the Nguni Languages

N. Govender, C. Kuun, V. Zimu, Etienne Barnard, Marelie Davel

Human Language Technologies Research Group, Meraka Institute / University of Pretoria, South Africa

We investigate two related issues in the computational modeling of Nguni prosody, based on annotated databases of isiZulu and isiXhosa speech. Firstly, we show that a simple template can be used to describe the tonal characteristics of vowels and adjectives spoken in isolation, and that contextual effects have only a mild impact on this template. This analysis was based on a simple mapping between pitch and tone; in the second part of the paper, we show that pitch and amplitude actually play comparable roles in producing tonal percepts.

Full Paper

Bibliographic reference.  Govender, N. / Kuun, C. / Zimu, V. / Barnard, Etienne / Davel, Marelie (2006): "Computational models of prosody in the Nguni languages", In MULTILING-2006, paper 015.