Bi-phoneme Frequency of Korean Based on the Analysis of Spontaneous Speech Data

Shin, Jiyoung

ORIGINAL ARTICLE

Commun Sci Disord. 2008;13(3): 477-502.

Jiyoung Shin^`

신지영(Jiyoung Shin)

ABSTRACT

Background & Objectives
The aim of the present study was to provide some information on frequencies of occurrence for units of Korean bi-phoneme based on the careful analysis of spontaneous speech spoken by adult speakers of Korean. Information on bi-phoneme frequency as well as phoneme and syllable frequencies is one of the most fundamental and basic knowledges on Korean language and is used in the segmentation of speech signal into meaningful utterance units.
Methods
57 adult (28 male and 29 female) speakers of standard Korean were employed as subjects for the present study. Spontaneous speech spoken by the subjects were recorded and phonemically transcribed. Bi-phoneme frequency was calculated from the database. The data size of the present study was as follows: 35,439 utterances, 403,572 syllables, and 858,479 phonemes.
Results
980,668 classified into 8 types (C+V, V+C, C$C, C$V, V$C, V$V, #+P, P+#) according to prosodic boundaries between phonemes, were examined. Logically possible numbers of bi-phoneme types were 1,225 types and 1,009 types of bi-phonemes were found in the data. However, only 5.5% (55 types) of bi-phoneme types covered over 50% of the speech data.
Discussion & Conclusion
Information on bi-phoneme frequency of Korean is one of the most fundamental and basic data for speech pathology, Korean language teaching, and speech engineering as well as Korean linguistics.