LPC系数的一种压缩方法

A Method of LPC Coefficients Compression

本文指出了在声码器、语音识别、说话人识别等方面在利用广泛应用的线性预测技术时,压缩LPC系数的重要意义,又在最小均方误差准则下,探讨了将p个LPC系数压缩到q个(1<q<p)系数的原理和方法。并给出了该方法的计算流程和一个具体例子。

This paper points out the importantance of the compression LPC coefficients at the line prediction technique such as voice coder,speech verification,speaker verification,the principle and method of p LPC coefficients compressing into q(1< q<p) coefficients is discussed at the standard of all square error margin,the calculation processing of that method with a example is given.