摘要语音的特征提取技术是一门语音学与信号处理技术相结合的交叉学科。它的 发展与当今在仪器领域掀起浪潮的微电子技术，计算机技术等有着密切的联系。

通过收集、整理前人对于特征提取研究的资料。综合成本低，便于实现等各 方面因素。选择了设计一个 LabVIEW 与 MATLAB 相结合的方案对语音的特征提取 进行研究。69252

在学习了语音特征提取的相关知识后，着手设计了一个基于 LabVIEW 的虚 拟式语音特征提取程序。该程序主要对语音信号进行采集，回放与处理分析。主 要分析了语音信号的时域特性及频域特性。利用 LabVIEW 与 MATLAB 相结合在对 采集到的语音进行了预处后提取了部分语音特征进行研究。这些语音特征主要包 括短时能量、短时过零率、短时傅里叶变换、LPCC（线性预测倒谱系数）、MFCC

（Mel 倒谱系数）等。

该毕业论文有图 31 个，有表 1 个，参考文献 26 篇。 关键词：语音信号 采集 预处理 LabVIEW

Speech Feature Extraction Based on LabVIEW

Abstract

Voice feature extraction technique is a phonetics and signal  processing technology combined with the cross-disciplinary. Its development t set off a wave of microelectronic technology, computer technology and other closely related in the instrumen today.

By collecting and sorting predecessors for feature extraction studies. Integrated low cost, easy to implement and so on various factors. Choose a design based on LabVIEW and MATLAB programs to study voice feature extraction technique.

After learning the speech feature extraction of relevant knowledge. I set out to design a virtual voice feature extraction procedures based on LabVIEW. The program mainly speech signal capture, playback and processing and analysis. It analyzes the time domain and frequency domain characteristics of the speech signal. Using LabVIEW and MATLAB combined in the collected speech after a pre-speech feature extraction part of the study. These voice features include short-term energy, zero-crossing rate of short-term, short-time Fourier transform, LPCC (Linear Predictive Cepstral Codings), MFCC (Mel Frequency Cepsturm Coefficients) and the like.

Key Words: voice signal acquisition pretreatment LabVIEW

目录

摘要 I

Abstract II

目录 II

图清单 IV

1 绪论 1

1.1 语音特征提取的发展 1

1.2 语音特征提取的应用 2

1.3 论文的主要工作与安排 2

2 语音信号理论与采集 4

2.1 语音信号概述 4

2.2 语音信号的采集与系统实现 5

3 语音信号分析 8

3.1 语音信号分析理论 8

3.2 语音信号的时域分析及系统实现 8

3.3 语音信号的频域分析 15

3.4 倒谱分析及系统实现 18

3.5 语音特征提取 19

4 程序设计及仿真原理 23

4.1 声卡概述