Research And Implementation Of Speech Enhancement Algorithm Based On Microphone Array

In voice communication,environmental noise will inevitably be introduced,which will affect communication quality.Therefore,people hope to obtain clear voice processing equipment and corresponding algorithms.The traditional single microphone is widely used because of its simple algorithm and certain effect.However,the communication environment is usually more complicated,and the single-microphone noise reduction algorithm cannot obtain satisfactory results in the process of positioning and noise reduction.So people propose microphone array-based speech enhancement algorithm to suppress environmental noise.The microphone array can obtain the spatial information of the sound source and realize the spatial filtering,and the algorithm can use this information to improve the speech enhancement effect.Moreover,the sound source can be tracked and located through the beam,which can be flexibly used in various occasions.In this paper,the direction of arrival estimation algorithm,speech enhancement algorithm and echo cancellation algorithm based on microphone array are studied.The estimation method of signal arrival direction in typical algorithms such as controllable beam location algorithm,subspace sound source location algorithm and sound source location algorithm based on TDOA is studied.Based on this,a beamforming and multi-source selection algorithm is proposed to estimate the direction of arrival of signal.The performance of the algorithm in the real environment is analyzed by the experimental results on the experimental platform.This paper studies fixed beamforming,adaptive beamforming,and post-filtered beamforming.The fixed beamforming algorithm is simple and easy to implement,but it is not applicable to coherent noise.The adaptive filtering beamforming algorithm can remove coherent noise,but it requires more microphones.The post-filtering beamforming method can effectively eliminate the incoherent noise in the environment of large signal-to-noise ratio of the received signal,and the algorithm is relatively simple,but the effect of incoherent noise in the environment of small signal-to-noise ratio is poor.Because these algorithms have their own advantages and disadvantages,a speech enhancement algorithm based on post-filtering and multi-source selection is proposed,and then the principle and framework of the algorithm are analyzed.Finally,the performance of the algorithm is analyzed by the experimental results.It can make the waveform contour of the signal more clear.This paper studies the least mean square algorithm,normalized least mean square algorithm and recursive least square method and other adaptive echo cancellation algorithms.An adaptive echo cancellation algorithm based on LMS is studied,and its principle block diagram is given.Experimental verification of the algorithm is carried out on OMAP-L137 platform,including a series of complete noise reduction processes such as direction of arrival estimation,speech enhancement and echo cancellation,and the audio waveform is clearer after processing.