This paper presents the performance analysis of energy efficient High Speed adaptive FIR filter for noise cancellation using modified LMS architecture. Synthesis results of adaptive noise cancellation for different CMOS families are concluded in this work. Here, Artix-7, Spartan-6 and Vertex-4 are taken for the most popular adaptive LMS filter for different order as N=8, 16, 32. Presented work is done by using MATLAB (2013b) and Xilinx (14.2). From the synthesis results we can find that CMOS (28nm) achieves lowest power and critical path delay as compared with others and thus proves its efficiency in terms of energy. In this paper MATLAB results are also shown for noise cancellation for different order with different step sizes. These results for different step sizes shows that when a very small step size is chosen then the delay increases in the cancellation of noise and vice versa. In this paper, it considered different parameters along with their optimized results.

Keywords: Adaptive filters, least mean square (LMS) algorithm, finite impulse response (FIR), Complementary Metal Oxide Semiconductor(CMOS)