ECE 535 DIGITAL SIGNAL PROCESSING


Problem Set 9

Spring 2001

Issued: Tuesday, April 10, 2001            Due: Tuesday, April 17, 2001



Reading in Oppenheim and Schafer

    
Week of 4/10/01 --- Chapter 7
    Section 5.7
    Appendix B



Although this problem set may seem long, note that problems 9.2, 9.3, and 9.4 require VERY short answers.

Problem 9.1
The decimation-in-time FFT algorithm was developed in Section 9.3 for radix 2, i.e., N=2n. A similar approach leads to a radix-3 algorithm when N=3n. Draw a flow graph for a 9-point decimation-in-time algorithm using a 3x3 decomposition of the DFT.

Problem 9.2
Oppenheim and Schafer, problem 7.3, part a only


Problem 9.3 Suppose we design a discrete-time filter using the impulse invariance technique with a CT lowpass filter as a prototype. The prototype filter has a cutoff frequency of Wc=2p(1000) rad/sec and the impulse invariance transformation uses T=0.2 ms. Assume that aliasing effects are negligible. What is the cutoff frequency wc of the resulting discrete-time filter?

Problem 9.4
We wish to design a discrete-time lowpass filter using the bilinear transformation on a continuous-time ideal lowpass filter. Assume that the continuous-time prototype has a cutoff frequency Wc=2p(2000) rad/sec and we choose the bilinear transformation parameter T=0.4 ms. What is the cutoff frequency wc of the resulting discrete-time filter?

Problem 9.5
Design a single-pole lowpass discrete-time filter with a 3-dB bandwidth of 0.2p using the bilinear transformation applied to the analog filter
Hc(s)=
Wc
s+Wc
where Wc is the 3-dB bandwidth of the analog filter.

(a)
Design the filter, i.e. determine H(z).

(b)
Sketch a structure for realizing this filter.
Problem 9.6
Suppose that you have a causal continuous-time filter defined by the system function:
Hc(s)=
s+0.1
(s+0.1)2+9
.
(a)
Use impulse invariance to determine the system function, H(z) of a discrete-time filter.
(b)
Sketch the pole-zero plot and the frequency response magnitude of the resulting discrete-time filter.
Problem 9.7
Consider designing a discrete-time filter with the system function H(z) from a continuous-time filter with rational system function Hc(s) by the transformation
H(z)=Hc(s) ½
½
 
s=(1-z-2)/(1+z-2)
.             (9.1-1)

(i)
What contour does the jW axis in the s-plane map to in the z-plane? Justify your answer.
(ii)
Determine the mapping between W and w defined by this transformation. Sketch W as a function of w.
(iii)
Suppose that the continuous-time filter is a stable lowpass filter with passband frequency response such that
1-d1£ |Hc(jW)|£ 1+d1     for    |W| £ 1.
If the discrete-time system is obtained by the transformation in Equation 9.1-1 above, determine the values of w in the interval |w|£p for which
1-d1£ |H(e
jw
 
)|£ 1+d1 .
Does this transformation provide a reasonable way to design DT lowpass filters from analog lowpass filter prototypes?

This document was translated from LATEX by HEVEA.