ISCA Archive Eurospeech 1993
ISCA Archive Eurospeech 1993

A method of classification among Japanese dialects

Shuichi Itahashi, Kimihito Tanaka

This paper describes a method of dialect classification of spoken Japanese, based on fundamental frequency (F0) contours of speech. Speech data of the famous Japanese folk tale "Peach Boy" were taken from a CD- ROM Japanese dialect speech corpus. The 14 dialect speech data were chosen from the major dialect districts. First, a jFo contour was approximated by a set of polygonal lines so that the mean square error between the lines and jFo values was minimized; the optimum boundaries of the lines were determined using a dynamic programming procedure. The starting frequency, slope and duration of each line were calculated. Secondly, parameters derived from Fq pattern were analyzed by using principal component analysis. Nineteen parameters including mean values and standard deviations of the above parameters were used for the analysis. Results show that the 14 dialects can be classified, based on these parameters.

Keywords: Spoken Japanese dialect, Fundamental frequency, Slope of F0 contour, Principal Component analysis.