The Research And Realization Of Acoustic Feedback Suppression Algorithm Based On Space-time Domain

Acoustic feedback is a common problem in the application of modern indoor sound reinforcement systems.When the microphone and speaker are in the same sound field,acoustic coupling will occur,and “howling” will appear as the gain of system sound reinforcement increases,resulting in the power overload or even burn out for the sound reinforcement equipment.Traditional methods of acoustic feedback suppression include frequency shift phase modulation method,pressure limit method,trapping method,etc.However,these methods have limited enhancement of the sound reinforcement system gain.Although the acoustic feedback cancellation method based on the adaptive filter improves the stable gain of the system,the strong correlation between the loudspeaker signal and the effective signal will cause the adaptive algorithm to generate a biased estimate of the acoustic feedback path,making the actual acoustic feedback suppression ability deteriorates.To solve the above problem,this thesis proposes an MPNLMS acoustic feedback suppression algorithm based on two-microphone model,which effectively reduces the biased estimation of adaptive filtering.Based on this,an acoustic feedback system combined with space-time domain and spatial filtering is proposed.Suppress algorithms to improve the ability of sound reinforcement systems to collect effective signals and enhance system performance.The main work of this thesis is organized as follows:(1)The acoustic feedback model,the traditional acoustic feedback suppression method,and the principle of adaptive filter for acoustic feedback suppression were introduced.The simulations of LMS,NLMS,PNLMS,and MPNLMS were compared.The results showed that under the same conditions,the MPNLMS algorithm has the fastest convergence speed and the largest system gain improvement.(2)In order to solve the problem of the eccentricity estimation of adaptive algorithm,an MPNLMS acoustic feedback suppression algorithm based on two-microphone model was proposed,which effectively reduced the influence of "eccentricity estimation" on acoustic feedback suppression.The simulation results showed that the algorithm is no longer subject to the correlation between the speaker output signal and the effective signal,and its convergence speed,error,and maximum gain were better than the MPNLMS algorithm in a single microphone acoustic feedback suppression system.(3)A spatial domain adaptive beamforming algorithm was added on the basis of the improved time domain algorithm.In the classroom scene,adaptive beamforming of the microphone array was used to improve the signal-to-noise ratio of the effective signal collected by the sound reinforcement system.As the result,the probability of howling increased,and the ability of the system to suppress acoustic feedback was enhanced.(4)Combining with airspace adaptive beamforming algorithm and time domain improved algorithm as the software framework of this system,an experimental prototype was built withTMS320C6748 DSP core,and the system was tested in the lab and classroom respectively.The results showed that the system can bring about 8.2dB ~ 8.8dB enhancement of the sound reinforcement.