一个语音基频提取新算法  

A New Algorithm for Speech Fundamental Frequency Extraction

作　　者：阳瑞新[1] 苏新菊[1] 姜占才[1] 

机构地区：[1]青海师范大学,西宁810008

出　　处：《电脑开发与应用》2010年第3期32-33,39,共3页Computer Development & Applications

摘　　要：为了准确提取语音基频,设计了一种新的基频提取算法,并对此算法的抗噪性能进行了评估,同一段语音进行加噪和未加噪两类处理,通过实验提取两类语音的基频,以未加噪基频作为准标值,对不同信噪比的语音基频与准标值的偏离情况进行了分析,从而了解算法提取基频的抗噪性。以便全面掌握此算法的性能,达到优化算法,减小估计带来的误差,提高语音基频提取的目的。In order to accurately extract the voice fundamental frequency, the design of a new fundamental frequency extraction algorithm, and the anti-noise performance of this algorithm was evaluated with a voice, the additional noise and without noise added two types of processing, extracting two experiments class voice fundamental frequency in order to not add noise fundamental frequency as the quasi-standard values for different signal to noise ratio of voice fundamental frequency and quasi-standard deviation values were analyzed to understand algorithm for extracting the fundamental frequency of the noise immunity. In order to fully grasp the performance of this algorithm to achieve optimal algorithm, reducing the estimation error caused by increase voice fundamental frequency extraction purposes.

关 键 词：语音信息处理 基频检测 算法优化 算法设计 抗噪性能 

分 类 号：TN912.3[电子电信—通信与信息系统]