Performance Evaluation Of A Low Order LMS Based Adaptive Feedback Canceller For Public Address System

In an environment wherein the public address system is used to address the general public, acoustic feedback occurs unpredictably which significantly degrades the acoustic quality of the information signal. An adaptive filter is then employed, with the objective that is to reduce the effect of acoustic feedback of the public address system that might occur at any point within the area. However, it is been debated what would be the filter length and adaption constant for its optimum performance. Furthermore, few studies suggest the filter length and adaptation constant for such application. In this paper, the performance of the adaptive filter implemented in FPGA with the least mean square (LMS) adaptive algorithm was evaluated and examine the effects of the filter length adaptation constant. A speaker, microphone, and audio amplifier were positioned to which acoustic feedback will occur. The amplifier's gain was regularly varied at different adaptation constant and filter length and the acoustic activity of the environment was then observed by a spectrum analyzer and spectrograph. Results show that with a large adaptation constant μ, the adaptive filter was able to quickly reduce the howling signal before the system exhibit unstable behavior. However, results also show that the magnitude of the output information signal of the public address system with large μ is lesser compared to a smaller μ. This shows that the adaptation constant μ does not only eliminate the acoustic feedback but also reduces the amplification capability of the public address system with an acoustic feedback controller. Furthermore, there is no significant difference observed on the effect of the length of the adaptive filter.