5th International Conference on Spoken Language Processing

Sydney, Australia
November 30 - December 4, 1998

Suprasegmental Cues for the Segmentation of Identical Vowel Sequences in Japanese

Kazuhiko Kakehi (1), Yuki Hirose (2)

(1) Nagoya University, Japan
(2) The Graduate School and University, The City University of New York, USA

This paper investigates how hearers cope with a sequence of more than two identical vowels --a common occurrence in Japanese speech. In the segmentation of identical vowels there are no spectral cues and very small power envelope changes in usual utterances containing identical vowels. We consider the effects of suprasegmental information such as duration, pitch pattern and rhythm of speech as important cues, and examine how, and to what extent a hearer can successfully make use of such information to segment each mora in a consecutive vowel series with and without the preceding sentential context.

Full Paper

Bibliographic reference.  Kakehi, Kazuhiko / Hirose, Yuki (1998): "Suprasegmental cues for the segmentation of identical vowel sequences in Japanese", In ICSLP-1998, paper 1011.