The Research Of Speech Enhancement Method Based On Microphone Array

In the actual environment,the speech signal is inevitably affected by the noise from different directions,which not only affect the intelligibility and intelligibility of speech,but also easily lead to human auditory fatigue.Therefore,it is necessary to carry on the noise reduction processing to the noisy speech signal.Compared with the traditional single-microphone speech enhancement method,the microphone array speech enhancement method adds spatial processing on the basis of time domain and frequency domain processing,while the received speech signal and noise overlap in time and spectrum.It is possible to process the signals from different directions of the space by air-time-frequency combining and has a high spatial resolution,which can get a better noise reduction effect,make up for a single microphone on the lack of noise suppression.This article describes the microphone array speech enhancement of background knowledge,and then introduces the generalized sidelobe canceller(GSC)traditional speech enhancement algorithm.According to the traditional GSC algorithm requires accurate delay estimation,a generalized sidelobe canceller based on the transfer function(TF)is proposed.The algorithm uses the transfer function coefficient ratio to build the blocking matrix.Therefore,it can improve the adaptability to the environment.The principle of noise suppression is based on the premise that the noise is coherent noise,it can effectively suppress the coherent noise in the signal,but the ability to suppress the incoherent noise is not strong.The post-filtering speech enhancement algorithm can effectively suppress the incoherent noise,so combine two algorithms above to form a multi-channel speech enhancement algorithm based on TF-GSC.The key of the algorithm is to obtain accurate priori signal to noise ratio(SNR)estimation and noise estimation.First,we introduce several algorithms used to estimate the probability of speech commonly,then improve one of the algorithms,and use MATLAB simulation software to simulate.The simulation results show that the improved algorithm can improve the signal-to-noise ratio and reduce the speech distortion compared with the traditional multi-channel post-filter algorithm.The hardware test platform based on Labview is built and the usability of the system is tested.The test results show that the system is suitable for microphone array speech enhancement research.