摘要
小波变换在语音处理中有广泛的应用,但是传统的阈值函数存在不连续且在临界阈值处不能平滑过渡等缺陷。针对这些缺陷,提出了一种改进的小波阈值函数,同时在获取阈值时采用贝叶斯阈值方法。最后通过仿真表明,改进的小波阈值函数与贝叶斯阈值方法结合能够实现临界阈值处的平滑过渡和解决不同层数阈值恒定的问题,而且提出的算法可以获得更高的信噪比和更小的均方误差,在降低信号失真和抑制噪声方面均有所提高。
Wavelet transform has a wide range of applications in speech processing,but the traditional threshold functions have some defects,such as discontinuity,no smooth transition at the critical threshold,etc.Aiming at these defects,an improved wavelet threshold function is proposed.At the same time,the Bayesian threshold method is used to obtain the threshold value.Finally,the simulation results show that the improved wavelet threshold function combined with Bayesian threshold method can realize the smooth transition at the critical threshold and solve the problem of constant threshold of different layers,and the proposed algorithm can obtain higher signal-to-noise ratio and smaller mean square error,and improve the noise reduction and suppression.
作者
卢勇
Lu Yong(Air Traffic Management Center,Civil Aviation Flight University of China,Guanghan 618307, China)
出处
《信息技术与网络安全》
2019年第8期38-41,共4页
Information Technology and Network Security
基金
青年基金项目(Q2019-072)
关键词
语音处理
小波变换
阈值函数
speech processing
wavelet transform
threshold function
