BEng (Hons) in ELECTRICAL & ELECTRONIC ENGINEERING DIGITAL SIGNAL PROCESSING 2002/2003 Question 1: Given a finite continuous-time signal x(t ) sin( 2t ) 0 t 1 (a) Determine the sampled result x(n) (n starts from 0) with the sampling period TS=0.25, and plot x(n). [5 marks] (b) Decompose x(n) as a sum of scaled and shifted unit samples using x ( n) x(k ) (n k ) . k [5 marks] (c) Determine type of following system, i.e. homogeneous, additive, and/or linear. y (n) 8 x(n 2) 4 x(n) 2 [5 marks] (d) Consider another system with finite unit-sample response h(n) {h(0)=-1, h(1)=1}, choose two different methods to find the response y1 (n) to the input x(n)(i.e. calculate the convolution y1 (n) x(n) h(n) ). [10 marks] Page 1 of 7 BEng (Hons) in ELECTRICAL & ELECTRONIC ENGINEERING DIGITAL SIGNAL PROCESSING 2002/2003 Question 2: (a) Using the definition of z-transform, determine the z-transform X(z) of the following sequence x(n) n u(n 1) , and shade its ROC in the z-plane. [9 marks] (b) Using the result of (a) and the property of z-transform, determine the ztransform Y(z) of the following sequence 1 y (n) (2) n u ( n 1) n [7 marks] (c) Determine the partial fraction of the system function H(z) of the system in figure1, use it and the result of (a) to evaluate the inverse z-transform h(n) Z 1 H ( z ) [9 marks] x(n) H ( z) 6 6 13 z 1 6 z 2 2 z 3 y(n) Figure 1 Page 2 of 7 BEng (Hons) in ELECTRICAL & ELECTRONIC ENGINEERING DIGITAL SIGNAL PROCESSING 2002/2003 Question 3: Sampling a continuous-time signal xa (t ) for 1s generates a sequence of 4096 samples. (a) What is the highest frequency in xa (t ) if it was sampled without aliasing? [5 marks] (b) If a 4096-point DFT of the sampled signal is computed, what is the frequency spacing in hertz between the DFT coefficients? [5 marks] (c) Suppose that we are only interested in the DFT samples that correspond to frequencies in the range 200 f 300 Hz. How many complex multiplications are required to evaluate these values computing the DFT directly, and how many are required if a decimation-in-time FFT is used? [8 marks] (d) How many frequency samples would be needed in order for the FFT algorithm to be more efficient than evaluating the DFT directly? [7 marks] Page 3 of 7 BEng (Hons) in ELECTRICAL & ELECTRONIC ENGINEERING DIGITAL SIGNAL PROCESSING 2002/2003 Question 4: Consider the sequence x(n) (n) 2 (n 2) (n 3) (a) Find the four-point DFT, X (k ) , of x(n) . [5 marks] (b) Find the finite-length sequence q(n) that has a four-point DFT Q(k ) W42 k X (k ) [6 marks] (c) If y (n) is the four-point circular convolution of x(n) with itself, find y (n) and the four-point DFT Y (k ) . [7 marks] (d) With h(n) (n) (n 1) 2 (n 3) , find the four-point circular convolution of x(n) with h(n) . [7 marks] Page 4 of 7 BEng (Hons) in ELECTRICAL & ELECTRONIC ENGINEERING DIGITAL SIGNAL PROCESSING 2002/2003 Question 5: Consider the high-pass filter as shown in figure 2, H (e j ) 1 34 0 3 4 Figure 2 (a) Find the unit sample response, h(n). [8 marks] (b) Sketch the other three ideal frequency selective filters (i.e. low-pass, bandpass, and bandstop filter) using the frequency response magnitude figure. Assume for low-pass filter, the cutoff frequency is 0< C <, and for bandpass and bandstop filter, the cutoff frequencies are 1 and 2 (0< 1 < 2 <). [9 marks] (c) A new system is defined so that its unit sample response is h1 (n) h(2n) . Sketch the magnitude of the frequency response, H 1 (e j ) , of this system. [8 marks] Page 5 of 7 BEng (Hons) in ELECTRICAL & ELECTRONIC ENGINEERING DIGITAL SIGNAL PROCESSING 2002/2003 Question 6: Consider the interconnection of LSI systems shown in figure 3. h2(n) x(n) + h1(n) y(n) + h3(n) h4(n) Figure 3 (a) Express the frequency response H (e j ) of the overall system in terms of H 1 (e j ) , H 2 (e j ) , H 3 (e j ) and H 4 (e j ) . [10 marks] (b) Find the frequency response H (e j ) , if h1(n) (n) 2 (n 2) (n 4) h2 (n) h3 (n) (0.2) n u (n) h4 (n) (n 2) [10 marks] (c) For H 4 (e j ) , determine its magnitude H 4 (e j ) , phase 4 h ( ) , and 4 h ( ) . [5 marks] Page 6 of 7 BEng (Hons) in ELECTRICAL & ELECTRONIC ENGINEERING DIGITAL SIGNAL PROCESSING 2002/2003 Appendix: You might find the following formulas useful during the calculation: 1 aN 1 a n 0 1 an 1 a n 0 N 1 an a 1 ( N 1)a N 1 Na N a (1 a) 2 n 0 a na n a 1 (1 a) 2 n 0 N 1 1 n N ( N 1) 2 n 0 N 1 1 n 2 N ( N 1)( 2 N 1) 6 n 0 N 1 na n Page 7 of 7