论文部分内容阅读
目前,人们对噪声背底中增强语言信号的课题日益重视,提出了各种处理方法。在许多方法中,都需要先提取语音的基频参数。本文简述了几种提取基频方法的原理,如平均幅度差函数法(AMDF),自相关法(ACF),极大似然法(ML),在不同信噪比之下,比较测定了这三种方法提取基频的效果。最后,本文还对极大似然法提出一点改进,测试表明,改进后的极大似然法(MLP),提取基频的效果较好。
At present, people pay more and more attention to the problem of enhancing the language signal in the noise background, and put forward various processing methods. In many ways, you need to extract the fundamental parameters of your speech. In this paper, we briefly introduce the principle of extracting the fundamental frequency, such as the average amplitude difference function method (AMDF), the autocorrelation method (ACF) and the maximum likelihood method (ML). Under different SNRs, These three methods extract the effect of the fundamental frequency. Finally, the paper also makes some improvements to the maximum likelihood method. The test shows that the improved maximum likelihood method (MLP) can effectively extract the fundamental frequency.