A comparison of conversational and clear speech versions of three Green Mong tones

Abstract

Green Mong has a phonological tone system that incorporates both fundamental frequency (F0) and phonation-type distinctions. Previous work using discriminant analyses examined three tones which have similar F0 contours, but which are characterized by the distinctive use of breathy, creaky and modal phonation [Andruski and Ratliff, JIPA (in press)]. Results indicated that the relative amplitude of lower and higher harmonics (measured as the H1–H2 amplitude difference) most reliably discriminates productions of these three tones. The contributions of jitter, duration, F0 and shimmer were also significant. This study compares the same three tones in one male speaker’s clear speech and conversational speech productions. Clear speech is typically more intelligible than conversational speech. The increase in intelligibility appears to result from exaggeration of acoustic cues that contribute to phonological distinctions in the language. The question explored here is the extent to which cues relating to phonation type are exaggerated in these three tones, as compared to F0 and F0 contour cues. Results of acoustic analyses of clear and conversational speech tokens will be presented, as well as results of a discriminant analysis using the discriminant functions generated in the previous study from the productions of six other speakers.