摘要当今，语音已经成为了人们最常用，最便捷，也是最重要的信息交换手段。因此，对语音信号技术的研究，显得尤为重要。

本文介绍了语音信号技术的研究现状和应用前景，接着，对语音方面的基础知识进行了论述，介绍了语音的定义和基本特性。然后，对本文采用的分析方法进行了介绍。该方法就是从语音的时域入手进行分析，检测出语音信号的端点，从而区分语音信号中的语音段与非语音段。因此我们需要提取语音信号的短时能量和短时过零率两大参数，并绘制参数曲线进行分析。为此，我们用VS2012编写代码程序。该程序要能绘制语音波形图并可以打开wav格式的语音文件，同时还要能分析语音波形图，并通过参数分析计算出短时能量和短时过零率，再绘制出它们的参数曲线。我们运用双门限法对参数曲线进行分析进而检测输入的语音信号的端点。通过对具体实验结果进行分析，我们得出结论：元音段短时能量高而短时过零率低，辅音段短时能量低而短时过零率高。双门限法可以较为准确的检测出语音信号的起始点和结束点。75483

毕业论文关键词  端点检测 双门限法 短时过零率 短时能量 

毕业设计说明书（论文）外文摘要

Title          speech endpoint detection   

Abstract Today, voice has become the most commonly used, the most convenient, but also the most important means of information exchange。 Therefore, it is very important to study the technology of speech signal。

This paper introduces the research status and application prospects of speech signal technology, and then, the basic knowledge of speech is discussed, the definition and basic characteristics of speech are introduced。 Then, this paper introduces the analysis method used in this paper。 According to the analysis of the time domain of speech, the endpoint of speech signal is detected, so as to distinguish the speech segment from the speech signal and the non speech segment。 Therefore, we need to extract the short-time energy and short-time zero crossing rate of the speech signal to the two parameters, and draw the curve of the parameters。 To this end, we use VS2012 to write code program。 The program to draw the speech waveform and can open the wav format audio files, but also the speech waveform analysis, and through the parameter analysis calculate short-time energy and short-time zero crossing rate and then mapped their parametric curves。 We use the basic principle of speech signal endpoint double threshold method to analyze the parameters of the input curve and detection。 Through analysis of the specific experimental results, we draw the conclusion: the speech in a short time high energy and short-time zero crossing rate was low, the speech in a short time of low energy and short-time energy high, double threshold method according to the characteristics of can accurately detect speech signal the start point and end point

Keywords  Edpoint detection, double threshold method，short-time zero crossing rate,short-time energy

目 录

1 绪论。。2

1。1 语音信号处理研究现状 …。。2

1。2本文研究内容2

1。3本文组织结构3

2 语音信号处理的基础知识…。。4

2。1 语音信号的波形特性。。。4

2。2 语音的分类。。。5

3 语音信号的时域特征分析…。。7

3。1 语音信号的预处理 …。。7

3。1。1 语音信号的预加重处理…7

3。1。2 语音信号的加窗处理7

3。2 短时平均能量分析…。。。9

3。3 短时平均过零率分析。10