摘要
线性预测法是语音信号处理中的核心技术。在语音信号的处理中,常常需要将线性预测的LPC系数与LSF参数相互转换。本文根据Chebyshev多项式求根法,研究了几种由LPC求解LSF的算法,分析了它们各自的特点及相互关系。分析并推导了由LSF求解LPC的算法。
Linear prediction is a kernel technology of speech processing. In speech processing, transformation between LPC and LSF is usually needed. In this research, according to the method of Chebyshev polynomial root finding, several algorithms of converting from LPC to LSF are studied, their features and the relation between each other are analyzed. The method of converting from LSF to LPC is analyzed and presented.
出处
《信号处理》
CSCD
2004年第5期521-524,共4页
Journal of Signal Processing
