基于VMD分解和小波阈值的语音信号去噪

Speech Signal Denoising Based on VMD Decomposition and Wavelet Threshold

为了对非平稳、非连续的语音信号进行降噪,提出一种基于VMD分解和小波阈值的语音降噪方法。通过仿真信号对比分析了VMD、EMD和EEMD算法对信号分解中存在的伪分量、模态混叠问题。先用VMD对语音信号进行分解,再利用小波阈值降噪。实验结果表明,该降噪方法明显优于小波阈值的语音信号降噪、基于EMD和小波阈值的语音信号降噪以及基于EEMD和小波阈值的语音信号降噪。

In order to reduce the noise of non-stationary and discontinuous speech signals, a speech de-noising method based on VMD decomposition and wavelet threshold is proposed in this paper. The simulation results show that the VMD, EMD and EE- MD algorithms have the problem of pseudo component and modal aliasing in the decomposition of the signal. First, the speech signal is decomposed by VMD, and then the wavelet threshold is used to reduce the noise. The experimental results show that the noise reduction method proposed in this paper is superior to the wavelet threshold noise reduction, voice noise de-noising based on EMD and wavelet threshold, and noise reduction based on EEMD and wavelet threshold.