Temporal organization of the speech intensity curve, in particular the ascending (positive) and descending (negative) intensity slopes across an utterance, was found to be speaker-specific in Zurich German, and idiosyncratic articulation was believed to be the driving force behind this phenomenon. Nevertheless, several issues are still pending, including (a) whether similar results can be obtained from speakers of other typologically different languages, (b) whether utterance-final lengthening has an effect on the result, and (c) whether utterance length has an effect on the result. The present article addresses these issues by analyzing read utterances from Thai speakers in addition to those from Zurich German speakers. Results suggest that in general, Zurich German and Thai speakers exhibit similar patterns in how between-speaker variability is explained by measures of positive and negative intensity slopes, in that measures of negative intensity slopes are more informative about speaker idiosyncrasy. Further analyses with the same datasets (Zurich German and Thai) with either the final or initial two syllables removed indicate that Thai data are more sensitive to the phenomenon of final lengthening than Zurich German. The Zurich German data were further partitioned into longer and shorter utterances, but the result does not seem to be affected by utterance length. Implications for forensic phonetic practices and semiautomatic speaker recognition as well as further research are discussed.
Zhang et al. (Mon,) studied this question.