ETRW on Speaker Characterization in Speech Technology Edinburgh, Scotland, UK June 26-28, 1990

Speaker-Dependent and Speaker-Independent Parameters in Intonation

Alex I. C. Monaghan (1), D. Robert Ladd (2)

(1) Centre for Speech Technology Research; (2) Department of Linguistics, University of Edinburgh, Scotland, UK

Given the highly structured nature of intonational phonology, we would expect that intonation could be described by a small set of phonetic parameters, and that inter-speaker variation would be constrained to a few specific parameters within this set. In order to characterise the range of inter-speaker phonetic variation in the realisation of a single phonological specification, it is necessary to refer to a model of the phonetics of intonation whose parameters correspond to tyhe dimensions along which speaker characteristics vary. We present a model of pitch contour generation, originally developed for speech synthesis applications, by which speakers' intonational behaviour can be reduced to just such a small set of phonetic parameters, and we illustrate the range of inter-speaker and intra-speaker variation which this model can produce. Problems with the interpretation of these parameters are also discussed.

Full Paper

Bibliographic reference.  Monaghan, Alex I. C. / Ladd, D. Robert (1990): "Speaker-dependent and speaker-independent parameters in intonation", In SCST-1990, 167-174.