Speech Signal Denoising Algorithm and Simulation Based on Wavelet Threshold
Gang Yang, Yonggang Song, Jia Du · 2022 4th International Conference on Natural Language Processing (ICNLP) · 2022
In the era of the gradual development of artificial intelligence, people use sound to control the machine to complete the work. When the machine recognizes the human voice signal, the voice must be a signal with less noise interference, so the voice signal denoising technology is particularly important. Common speech signals are non-stationary random signals, mixed with various noises, and it is difficult to filter out common denoising methods. Aiming at this kind of non-stationary random signal denoising processing, this paper uses the characteristics of wavelet transform multi-resolution analysis to improve the wavelet threshold denoising algorithm and uses the improved wavelet threshold function to remove noise and get the original speech signal. The simulation results show that the improved wavelet threshold denoising algorithm has better performance. When analyzing non-stationary random signals, the denoising effect under different signal-to-noise ratios is very significant.