Let us analyze the quantization effects in your designed filters. Quantize the filter coefficients in your original designs to B bits on the interval (-1, 1) using the uencode/udecode commands (see MATLAB help). You need to scale both numerators and denominators by the absolute value of the largest coefficient, and then quantize. The scaling factors up front of the transfer function can be disregarded. (a) Plot on the same graph the magnitude responses (in dB scale) of: (i) your original design without coefficient quantization; (ii) your original filter with [a, b] coefficients quantized to B = 8 bits and B = 4 bits. Comment on these plots. (b) From your results, identify a filter which is robust to quantization errors