Simulation and Performance Analysis of Various Least Mean Square Algorithms for Speech Enhancement

In recent years speech technology applications have been rapidly increased because of huge demand in speech applications, but their performance in terms of quality and intelligibility has been affected due to the presence of background noise. The problem of acoustic noise reduction in speech communication devices has attracted a considerable amount of research attention. Adaptive noise cancellers are most suitable because of their ability to control the background noise. In this paper, proposed an optimal adaptive filter out of the variations of the Least Mean Square (LMS) algorithms. The system has been tested under various noise conditions. The performance has been tested by objective and subjective methods to confirm quality and intelligibility of the estimated signals.