Research Of Speech Enhancement Based On Microphone Array

Under the real environment, speech is usually inevitably be interfered by noise andreverberation, those interferences will reduce intelligibility and clarity of speech. Sospeech enhancement is introduced, the aim is pick up speech information form speechwith noise and obtain high quality speech. In recent years, the speech enhancementmethods based on microphone array have become research hot pot. On this background,this paper studied it, and the contributions are as follows:Firstly, analyzed the topological structure of microphone array, studied the signalmodels of microphone array. Analysis the several common speech enhancementalgorithms, then presents the advantages and disadvantages of these algorithms.Secondly, studied speech enhancement based on the structure of generalized side lobecancellation. Focus on structure and principle of this method. Studied the different kindsof time delay estimation techniques in it, those methods are studied on anti-reverberationand noise resistance in different acoustic environment, then presents the advantages anddisadvantages of these methods. In the different kinds of noise and signal to noise ratio,compared the de-noising performance of the generalized side lobe algorithm.Thirdly, studied speech enhancement based on wavelet transform. Analyzed the basicprinciple of de-noising; discussed in wavelet base, level of decomposition, thresholdfunction and threshold of the method, and improved those. In the different kinds of noiseand signal to noise ratio, compared the de-noising performance of this method.Finally, studied speech enhancement based on GSC and rear wavelet transform.Designed a rear enhancement method combining the advantages of the two methods, jointhe voiced sound judgment in wavelet threshold de-noising part, In the different kinds ofnoise and signal to noise ratio, compared the de-noising performance of those methods. Inreverberation environment, compared the de-noising performance of those methods.