| Acoustic feedback is a common phenomenon in sound amplifying system, Sound will be badly destroyed and electronic components even will be burned when it comes serious due to certain frequency bands of the signal are repeated processing gain in a closed loop. There are a lot of methods for acoustic feedback suppression, including frequency shift method, phase shifting method and so on. The algorithms are not widely accepted because of damages to sound quality, need for professional operation and other shortcomings, although them has respective advantages.The author of this paper attends in his tutor’s research subject “Research on DSP intelligent audio mixer and industrialization” and chooses to research the technology of sound feedback suppression. This paper analyses the howling signal decision process, the notch-filter parameter design, adaptive algorithm and other aspects of notch-filter-based howling suppression(NHS) algorithm and presents the NHS algorithm using multi band technology. The paper also presents the double adaptive notch filters algorithm based on normalized least mean square algorithm with Pre-emphasis method and simple linear regression analysis. The algorithms was studied experimentally on DSP platform order to retain more original signal information, get less damages to sound quality and suppress howling signals at the same time.Although automatically howling suppressor is widely accepted, in the market of professional howling suppressor mostly are the foreign products with high price. Analyzed the performance of efficiency, function parameters based on the demand of the conference system, author decide use ADI SHARC DSP 21489 and CCES and sigma platform to develop the conference system in the cooperation with Foshan HMAUDIO Ltd. and produce the M1000 intelligent mixer. |
PDF Full Text Request