Filtering in the Time and Frequency Domains

We choose the magnitude function associated with H( s) in (3.3-1) to have the form
| (3.4-1) | |
where ? is a positive, real number not greater than unity, and ? n( ?) is an nth-order polynomial containing only even or only odd powers of ?. If the value of ? n( ?) is small in the passband and large in the stopband, H( j ?) is a good approximation to the rectangular magnitude response in Fig. 3-5 a. Four classes of filter responses obtained in this manner are now discussed.
Setting ? - 1 and ? n( ?) = ? n in (3.4-1) yields the Butterworth response
| (3.4-2) | |
Reference to Section 3.2.1 shows that H( j ?) 2 satisfies the maximally flat criterion at ? = 0, since the required coefficients of powers of ? are zero. As we demonstrated in Example 3-4, a Taylor approximation by H( j ?) 2 is also a Taylor approximation by H( j ?). Since H( j ?) is unity at ? =0 and 1/
at ? = 1 independent of the approximation order n, it is normalized.
The transfer function, obtained by the technique in Example 2-27, is listed in (3.4-3) for n = 2, 3, and...