本文对采用减谱法和相关法的语音增强系统进行了研究,系统的噪声为白噪声和普通机房噪声。
In this paper, we have studied several speech enhancement systems that use spectrum subtraction method and Allen's reverberate noise suppression technique.
提出了基于高阶谱的减噪方法,能有效提高回波信号的信噪比,减弱检测噪声对于缺陷散射特征的影响,从而突出缺陷特征信息。
A method for reducing noise based on high-order spectra was presented by which the signal-to-noise ratio (SNR) could be improved significantly and the flaw characteristic information could stand out.
实验表明,算法对于有色噪声干扰下的语音信号有较好的增强效果,并且性能优于改进减谱法。
Simulation results show that this algorithm demonstrates better performance than modified spectrum subtraction algorithm both in thee nvironment of white or colored noise.
应用推荐