In order to improve the problem that it is difficult to select the key parameters of VMD and the lack of acoustic features in single-channel speech enhancement,a single-channel speech enhancement model based on BDO and VMD-EAM algorithm is proposed.Firstly,the DBO is used to process the input noisy speech to obtain the opti-mal combination of the number of decomposition modes and penalty factor,so that the adaptive optimization of the key parameters of VMD is realized.Then,the classification of IMF component is completed according to the similari-ty evaluation results of EAM,and the high-frequency information of noise component is filtered by wavelet threshold method.Finally,the signal was reconstructed based on the reconstruction principle of IVMD-EAM model.Simulation and real experiments show that the proposed speech enhancement model can effectively suppress noise and improve the enhancement effect of English speech,and all performance evaluation indicators are significantly better than the other two single-channel speech enhancement models.
维普
万方数据