【摘要】 用非线性动力学方法对正常语速的汉语语音进行了初步研究，对关税维算法进行了改进以适合语音信号的特点，文章给出了摩擦音和单元音汉语的相空间重构图及关联维曲线，发现由于它们在发声机制上的不同，关联维算法能够区分摩擦音和单元音，同时初步研究表明，关联维算法也能为区分汉语四声提供信息。更多还原

【Abstract】 Nonlinear dynamic method is used in studying the Chinese spoken in normal speed, the improved correlationdimension algorithm is made for the characterization of speech signal, the reconstructed phase space and correlationdimension curves of fricatives and vowels are also given. It is found that because of the different mechanism betweenthem, the correlation dimension algorithm can distinguish fricative from vowel. Also our study shows that it can provideinformation for distinguishing four basic tones in mandarin speech.更多还原