摘要
为了改善语音去噪的效果,提出了一种基于小波分析的新阈值去噪方法。该方法克服了硬阈值函数的不连续性,软阈值函数中存在恒定偏差的问题,构造了一种任意阶可导的新阈值函数,便于进行各种数学处理。用MATLAB软件对硬阈值、软阈值和新阈值方法进行仿真分析比较,实验结果表明,新阈值方法不但在视觉上,而且在信噪比和最小均方误差等方面都优于传统的硬、软阈值方法,比传统方法具有更大的优越性和有效性。
In order to improve speech de-noising effect,a new threshold de-noising method based on wavelet analysis is proposed.This method overcomes the hard-threshold function is not continuous,the soft threshold function that exists the constant deviation,a kind of arbitrary order derivable the new threshold function is constructed,which is easy to carry out a variety of mathematical treatment.At last,hard threshold,soft threshold and the new threshold methods are analyzed and compared by MATLAB software.Experimental results show that the new threshold method is not only visually,but also is better than the traditional hard and soft-threshold method in signal to noise ratio（SNR）,minimum mean square error（MSE） and so on,which has more advantages and effectiveness than the traditional methods.
出处
《计算机工程与设计》
CSCD
北大核心
2011年第3期1099-1102,共4页
Computer Engineering and Design
基金
国家自然科学基金项目(60672001)
陕西省教育厅专项基金项目(05JC03)
