Lecture 11: FIR Filter Designs XILIANG LUO 2014/11 1 Windowing Desired frequency response: Fourier series for a periodic function with period 2pi Convergence of the Fourier series 2 Windowing 3 Windowing 4 Windowing Rectangular window: 5 Common Windows 6 Common Windows 7 Common Windows M=50 Rectangular Window 8 Common Windows M=50 Hamming Window 9 Common Windows M=50 Blackman Window 10 Comparisons 11 Kaiser Window 12 Kaiser Window 13 Kaiser Window 14 Kaiser Window 15 Kaiser Window 16 Optimal FIR Filter Design Type-1 FIR filter: 17 Optimal FIR Filter 18 Optimal FIR Filter Parks-McClellan algorithm is based on the reformulating the filter design problem as a problem in polynomial approximation. 19 Optimal FIR Filter Approx. Error: only defined in interested subintervals of [0, pi] 20 Optimal FIR Filter Parks-McClellan, MinMax criterion: 21 Optimal FIR Filter 22 Parks-McClellan Alternation theorem gives necessary and sufficient conditions on the error for optimality in the Chebyshev or minimax sense! Optimal FIR should satisfy: 23 Parks-McClellan 2(L+2) unknowns 𝜔𝑝 , 𝜔𝑠 are two alternation frequencies 24 Parks-McClellan Given set of the extremal frequencies, we can have: 25 Parks-McClellan Given set of the extremal frequencies, we can have: Evaluate on other frequencies 26 Parks-McClellan 27 Flow Chart of Parks-McClellen 28 29 30 1.4 1.2 1 0.8 0.6 0.4 0.2 0 0 0.5 1 1.5 2 2.5 3 3.5 31