Research on intelligent acoustic feedback exploration and suppression system

Abstract

Acoustic feedback is a common phenomenon in public addressing systems. Howling, which results from acoustic feedback, restrains the gain of amplifier and affects the transmissive sound's definition. When feedback goes serious, the public addressing systems even can't work. So making acoustic feedback under control is an key gist to the designer of public address system. By researching the process of howling, we have proposed an effectual scheme: use FFT and Chirp-z transform algorithm to detect acoustic feedback, and create a wave, which has the same amplitude and frequency with the howling signal but have reverse phase, to add on the input signal so that it can remove acoustic feedback from the useful sound. Experimental results show that the system performs a very well in acoustic feedback suppression and it's accuracy is 1Hz.