摘要
研究了语音信号的混沌动力学机制，并用相空间重建图、语音信号极坐标表示、语音信号时域波形等方式研究了语音信号的分维数，改进了ＢｏｘＣｏｕｎｔｉｎｇ的分维数计算方法，改进后的计算方法准确可靠、处理速度快．用该方法对汉语语音的分维数进行了统计分析，得到汉语语音的基本分形特征．
Based on chaos and fractal cheories the dynamic modeling of speech signal on chaos and fractal theories is presented,and the methods of phase reconstruction,polar repesentation,power spectrum and wave form representation are proposed to analysis speech signal.The paper compares several methods for fractal dimension estimation,and a more accurate and efficient boxcounting method based on a novel algorithm(NBCA)is developed.With the help of NBCA,fractal characterization of Chinese speech signal is studied.
出处
《四川大学学报（自然科学版）》
CAS
CSCD
北大核心
1998年第3期414-419,共6页
Journal of Sichuan University(Natural Science Edition)
关键词
混沌
分形
语音信号处理
chaos, fractal,speech signal processing
