Sixth International Conference on Spoken Language Processing
October 16-20, 2000
The Acoustic Characteristics of Japanese Identical Vowel Sequences in Connected Speech
Tsutomu Sato (1), John A. Maidment (2)
(1) English Department, Meiji Gakuin University, Tokyo, Japan
It has been pointed out that a word which contains an
identical vowel sequence may not be distinguished,
especially in faster speech, from a word which has a
long vowel and phonologically the same mora-count.
However, this indication has been made without any
objective evidence and it is argued that a quantitative
study is needed to investigate the acoustic
characteristics of identical vowel sequence in
comparison with those of a long vowel. This paper
reports the results of production experiments for this
purpose done by 10 native speakers of Japanese.
The results will be discussed from the viewpoints of
glottal stop insertion between an identical vowel
sequence and durational ratio comparison. More
specifically, the actual number of glottal stop insertions
on citation and sentence levels will be clearly shown
and the acoustic features of the glottal stop such as its
duration will be investigated. In addition, the ratios
of the whole token, consonant + vowel(s) and the
vowel part of identical vowel sequence examples will
be made clear and compared with those of long vowel
examples in both citation and sentence reading forms.
(2) Department of Phonetics and Linguistics, University College London, UK
Sato, Tsutomu / Maidment, John A. (2000):
"The acoustic characteristics of Japanese identical vowel sequences in connected speech",
In ICSLP-2000, vol.2, 632-635.