The interaction of long-term voice quality with the realization of focus

Voice quality shifts have been shown to be associated with the realisation of accent, focus and deaccentuation. Mostly, accented and focally accented syllables are reported to exhibit a tenser mode of phonation than the unaccented, but accentuation with a laxer/breathier quality is also reported. Possibly, the long-term voice quality of the speaker (or of the utterance) influences the voice quality of accented/unaccented syllables. This paper examines the hypothesis that speakers who have a baseline tense phonatory setting, or who are using a tenser mode of phonation in a particular utterance, will signal accentuation more through breathy phonation than through increased tenseness. To test this hypothesis we analysed utterances produced by a single informant with three voice qualities; modal, breathy and tense. These utterances were produced with variations in the location of a focal accent. The utterances were manually inverse filtered and the voice source parameters F0, EE, UP, RG, CQ and RD were obtained by fitting the LF (Liljencrants-Fant) model to the glottal flow signal. The change in parameter settings in focally accented syllables was examined relative to the adjacent unaccented syllables. The hypothesis was not supported. Overall, the focally accented syllables were associated with tenser phonation, except in the utterance-initial position where they were associated with breathier voice irrespective of the sentence’s intended phonation type. In these data, the voice source modulations associated with focal accentuation appear to depend on the position of the focal syllable in the phrase.