Abstract
The purpose of this paper is to investigate difference of vowels
between Chinese and Japanese based on chaos analysis. At first, we study characteristic features of Chinese vowels data's by means of chaos analysis. Thus, we investigate embedding dimension of
data's by making use of correlation dimension analysis and false neighbor method, and then we reconstruct an attractor of them by means of Takens'embedding method. According to the attractor, we perform Lyapunov spectrum analysis, Lyapunov dimension analysis and information dimension analysis. In addition, we investigate the meaning of the results by comparing with the difference in direct listening.