摘要线性预测编码作为一种基于全极点模型假定和均方预测误差最小准则下的波形逼近技术，是语音编码中的一种重要技术,运用这种编码方式通过对音频信号的时域和频域分析，线性预测编解码，滤波等处理,能够在尽量保持原音质的基础上合成出高质量的语音.本文通过对波形编码，参量编码，混合编码,中低速率编码的介绍，重点讨论了线性预测(LPC)的原理、Levinson-Durbin 算法及 MATLAB 在处理语音信号时的应用。最后以 MATLAB 为开发工具，编写了线性预测编码的程序,对语音信号进行了线性预测编码处理,并对阶数的选取做了比较分析。 关键词 语音信号 LPC 莱文森-杜宾算法 MATLAB 仿真AbstractLPC all-pole model as based on assumptions and the smallest mean squared prediction error criterion of waveform approximation technique ， is an important speech coding techniques，By using this encoding audio signals in time domain and frequency domain analysis, linear prediction coding and decoding, filtering processing, able to keep the original sound quality on the basis of high-quality voice synthesis。In this paper ， based on the introduction of the waveform coding,parametric coding， hybrid coding， low rate coding, focuses on the linear prediction （LPC) principle,Levinson—Durbin algorithm and MATLAB in the application of speech signal processing. Finally, with MATLAB as the development tool ， write a linear predictive coding procedures ， voice signal processing of linear predictive coding。，and the number of orders to do a comparative analysis。Keywords voice signal LPC Levinson — Durbin algorithm MATLAB simulation目 录第 1 章 绪论...................................................11.1 语音信号 LPC 分析技术的基本概念....................................11。2 语音信号 LPC 分析技术的进展.......................................21.3 语音信号 LPC 分析技术的应用........................................21。4 基本设计要求及设计思路.............................................
