The Design Of Adaptive Acoustic Feedback Suppressor Based On SHARC DSP

The ability of an adaptive filter to track time variations of input statistics and automatically adjust itself according to the input signal makes the adaptive filter widely used in radar, communications, biomedical engineering and image processing. Acoustic feedback is a serious problem in sound reinforcement system. Acoustic feedback is both annoying and reduces the maximum usable gain of sound reinforcement system. Considering the deficiencies of the traditional acoustic feedback suppression algorithm, the adaptive filter is used to suppress acoustic feedback phenomenon in this dissertation.Acoustic feedback reduction (AFR) algorithm is proposed based on the adaptive system identification theory. The system uses an adaptive filter to approximate the feedback path, and then removes the acoustic feedback signal in the system inputs as more as possible. To achieve this result, the mathematical model of acoustic feedback reduction with adaptive filter is given. Based on this model, the dissertation mainly focuses on the following aspects:Firstly, an acoustic feedback reduction algorithm is proposed on the basis of in-depth study of the adaptive filter theory. The structure of the adaptive filter and the principle of adaptive system identification are analyzed. A mathematical derivation is made for the Least Mean Square (LMS) algorithm and Normalized Least Mean Square (NLMS) algorithm. In order to improve the performance of the system, a preliminary research on noise suppression algorithm is made. The noise suppression algorithm and delay module are merged into the model, and then the improved version of AFR model is proposed.Secondly, the AFR algorithm is emulated on MATLAB. On the basis of the algorithm preliminary meets the system requirements, several comparative analyses are designed to see the effects of various parameters on the performance of the algorithm. Noise influence is also taken into consideration. Eventually it is confirmed that the improved version of AFR model has better acoustic feedback suppression.Lastly, the overall design of the system is described. Based on the introduction of the powerful SHARC DSP series, the description of the hardware implementation of the overall system design elaborated. The algorithm is replanted on DSP experimental kit with C programming language and optimized by instructions. Eventually it is shown that the algorithm achieves good results in the actual test.