RESEARCH ON ENHANCEMENT AND DENOISING OF ENGLISH SPEECH IN COMPLEX NOISE ENVIRONMENTS
International Journal of Mechatronics and Applied Mechanics · 2025
Enhancement of speech in complex noise environments is beneficial to communication and interaction.This paper presents a speech enhancement algorithm.The wavelet transform reduces speech noise initially.Then, a convolutional neural network (CNN) is trained using a loss function with a time-frequency masking function for enhancement and denoising.Moreover, the gain coefficient and phase deviation are introduced into the time-frequency masking function to further improve algorithm performance.Simulation experiments were conducted eventually.The results demonstrated that the enhancement and denoising algorithm's effectiveness improved as the signal-to-noise ratio (SNR) of the noise increased.However, the effect became less significant after the SNR exceeded a certain threshold.The improved CNN algorithm not only converges faster with minor errors but also performs exceptionally well in enhancing noise reduction of English speech in complex noise environments.