倒谱域提取基音信号.docVIP

PAGE 摘 要 人类在讲话时发出的一种别具一格的声音，被称之为语音。一系列的音汇聚成为语音，一系列的规则又控制着各个音的排列顺序。被称之为搭载特定模拟信号信息的的语音，是人们日常社会生产和生活中得到以及传播资讯的重要方式。为什么要处理语音信号的原因其实就是为了既能提取有效的信息还能在复杂的环境中做到这一点。 基音周期是语言信号的重要参数，因此也是语音激励源的特色所在，对描述其特征有重要作用。现阶段，该参数主要应用于指导听觉障碍人群、诊断发音系统方面的疾病、编码低码率语音以及语音识别等多个方面。其中语音识别与语音分析更是在许多行业得到应用。倒谱法被认为是一种比较成功的基音检测方法, 倒谱法作为信号处理的重要的方法之一，能够得到比较好的识别性能。 本毕业设计拟采用在倒谱域进行语音信号的基音提取和估算，从而了解语音信号的生成机理、掌握基音生成的概念和产生原理、掌握倒谱域提取基音信号的方法和过程。通过MATLAB仿真进行倒谱法基音检测以及简单的后处理，对仿真结果加以分析。最后能顺利检测提取到基音。 关键词：语音信号；基音周期；倒谱域 ABSTRACT A distinctive sound that humans make when they speak is called phonology. A series of sounds converge into speech, and a series of rules control the order in which each sound is arranged. Speech, which is called as carrying specific analog signal information, is an important way for people to obtain and disseminate information in daily social production and life. And the reason why you want to deal with voice signals is because you want to be able to extract valid information and do that in complex environments. To describe the important characteristics of speech excitation source, pitch period as one of the important parameters of speech signal must be necessary. Pitch cycle information is widely used in speech recognition, speaker recognition, speech analysis and synthesis, low bitrate speech coding, speech system disease diagnosis, language guidance for hearing impaired people and other fields. Cepstrum method is considered as a successful pitch detection method. As one of the important methods of signal processing, cepstrum method can get better recognition performance. This graduation project intends to adopt pitch extraction and estimation of speech signal in cepstrum domain, so as to understand the generation mechanism of speech signal, master the concept and generation principle of pitch generation, and master the method and process of pitch extraction in cepstrum domain. MATLAB simulation is carried out to detect the fundamental sound of cep