DSP Midterm

October 26, 2023 Close Book (maximum score 195) 09:10pm – 12:00pm
Note: If no indications are given, x[n], y[n], and h[n] as well as their associated z-transforms
denote the input, the output, and the impulse response of the system, respectively.
1. Multiple Answers and Multiple Choices (25%) (could have no or many solutions)
(1) ( ) Which systems are linear?
n n
(a) y[n] = 2x[n] + y[n−1]; (b) h[n] = (0.2) u[n−1] + 2 u[−n−2].; (c) y[n] = 3x[n−2] +2nx[n−3];
(d) y[n] =3x[n−1]+2x[n−3] + 5; (e) y[n] = 3x[n−1] + x[n] with y[−1] = 3;
 x[n / 3] if n = 3k
(2) ( ) Which systems are stable? (a) y[n] =  ; (b) y[n] = 4x[n] + x[n−1]+1;
 0 if n ≠ 3k
(c) y[n] = x2[n]; (d) y[n] = min{y[n−1], x[n], x[n+1]]}; (e) y[n] = 2x[n−1] + x[n] with y[−1] = 0.
(3) ( ) Which systems are linear and time-invariant?
(a) y[n] = 2x[n] + 3δ[n−3]; (b) y[n] = x[n/2]+x[n] for n = even and y[n] = x[n] for n =odd; (
c) y[n] = x[n]y[n]; (d) y[n] = 2x[n−1] + x[n] with y[−1] = 0.; (e) y[n] = 2x[n−1]+ 0.5nx[n].
z + 2 z −1
(4) ( ) Which systems are stable? (a) y[n] = 0.8y[n−1] − 4x[n−2] + 2; (b) H(z) = forz > 0.5
2z − 1
z2 + 5
n n
(c) y[n] = x[2n]; (d) h[n] = (0.5) δ[n−1] −(0.5) u[n−4]; (e) H(z)= forz< 5.
z +5
(5) ( ) Which systems are causal? (a) h[n] = (0.3)nu[n−1] + 2nu[−n−2]; (b) y[n] − 2y[n−1] = 0.5x[n] with
z 2z + 1
y[−1] = 0; (c) h[n]= 3δ[n+1]+2δ[n−1]; (d) H(z)= forz> 2; (e) H(z)= forz> 0.8
z−2 z − 0.8
2. For the LTI systems, which are characterized by the following z-transforms, please
(a) Identify them if the system is stable and causal or not? (12%)
(b) Identify them if they consist of an initial or final value? If yes, find it position and the value (12%)
2 z −2 − 1 2 z −3 + 2 z −1 + 3
(a) H1(z) = for z < 2 ; (b) H2(z) = for z < 2 ;
( 2 z −1 − 1)( z −1 − 3) 4 z −2 − 1
−1 −1
(c) H3(z) = ( z + 1)( 2 z − 1) for z > 1 / 3 ; (d) H4(z) = 3 z −1 + 1 for z > 1 / 2
z −1 − 3 ( z −1 − 2)( z −1 + 3)
3z 2 + 2 4 z −2 + 2 z −1 + 1
(e) H5(z) = for z < 1 / 2 ; (f) H6(z) = for z < 1 / 3 .
(1 − 3z )(1 − 2 z ) ( z −1 + 3)2
3. In the following formulations, each equation could be a sequence or a value. Compute the result if it is a
value or draw a diagram with a specific time index (in time index k or n) if it is a sequence: (24%)
5 ∞ 4
8
(1) ∑ 2n δ[k − n] ; (2) ∑ (2k + 4k )δ [k ] ;
k = −3
(3) ∑ 2k δ [k + 3] ;
k =0
(4) ∑ 2 δ [n − k ] ;
k =2
k

n=0
5
(5) 7δ[k ] + 3δ[k − 1] + δ[k − 2] − 6δ[k − 3] ; (6) ∑ cos(kπ / 4)δ [n − k ] ;
k =0
9
(7) ∑{7δ [k ] + 3δ [k − 1] + δ [k − 2] − 6δ [k − 3]} ;
k =5
(8) 4δ [k − 3] + 2δ [k − 2] + 7δ [k + 1] + 6δ [k + 3] .
4. For linear and time-invariant systems characterized by the following impulse responses, please identify
them whether the systems are stable, causal, memory or not? (12%)
6

∑ 2 δ [n − k ] ;
∞
k
=
(a) h[n] n n
(b) h[n] =(−2) u[n − 1] − (1/ 3) u[n] ; =
(c) h[n] ∑ (0.8) δ [n − k ] ;
k

k =0 k = −3
∞ 5 4
(d)
= h[n] ∑ 3 δ [n − k ] ;
k
2δ [n] + ∑ 3k δ [n − k ] ;
(e) h[n] = (f)
= h[n] ∑ 2 δ [n − k ] .
k

k =0 k =0 k = −1

5. Determine the z-transform of: (a) x[n] = ( − 14 ) n −1 u[n − 3] + 3( 12 ) − n u[n + 1] . (5%)

(b) x[n] = ( − 23 ) n u[n − 2] − 4( 13 ) n n u[n − 1] . (15%)
6. Determine the inverse z-transform of
2 − 4 z −2 (3 z + 2)
(a) X1(z) = for 3> z > 2 (10%); (b) X2(z) = for z > 0.5 (10%)
(1 + 2 z −1 )(1 + 3 z −1 ) ( z − 0.5) 2
7. Let h[n] be the impulse response of a causal system H(z) with four poles at z = 1/3, z = −2, z = 1/3 − j1/4
and z = 1/3 + j1/4 and two zeros at z = -0.5, z = 3, find new poles and zeros of the new systems if we
change their impulse responses to:
n
(a) h1[n] = 2(−0.5) h[n] ; (5%) (b) h2[n] = 2(−1) h[n − 1] ; (5%) (c) h3[n] = 3(−0.5) h[− n − 2] ; (10%)
n n+2

8. If x[n] is shown in Fig. 8, If the z-transforms of y[n] 2 2

and w[n] respectively denote as Y(z) and W(z), please Figure 8 1
(a) find and plot y(n) if Y(z) = z−1X(z) − 2z*X(z*); (5%) −1 2 n
(b) find and plot w(n) if W(z) = z+2X(z) − 3z−1X(1/z); (5%) 0 1 3 4
(‘*’denotes the conjugate of the argument)
−1 −1

9. Fourier transform of x[n] = (−1+3j)δ[n+3] + (2−4j)δ[n+1] + (3−2j)δ[n−1] +(−2+3j)δ[n−3] is denoted as

X(ejω), you could obain the variations of its Fourier transforms with symmetrical properties.
∞
(a) If Y1(ejω) = 3e−jωX*(ejω), please compute ∑ y1[n] =? (5%)
n = −∞
(b) If Y2(ejω) = 2e−2jωXe(e−jω), please find its inverse Fourier transform, y2(n)=? (5%)
(c) If Y3(ejω) =3e−3jωXo*(e−jω), please find its inverse Fourier transform, y3(n)=? (5%)
jω −3jω −jω ∞ 2
(d) If Y4(e ) = 2e X*(e ), please find ∫ e− jωY4 ( e jω ) dω =? (5%)
-∞
10. For difference equation y[n] + 3y[n−1] = x[n] with initial condition y[−1] = 1, if the input is given as
x[n]=(1/2)nu(n), please determine response y[n] by using unilateral transform. (10%)
1, ω < π / L
11. Consider a multirate system, which is shown in Fig. 11(a), where H (e jω ) =  , please
0, π / L < ω ≤ π .
sketch Yc(j Ω) if Xc(j Ω) is shown in Fig. 11(b). (10%)
Xc(j Ω)
xc(t) x[n] v[n] w[n] y [n] yc(t)
C/D L H(ejω) D/C
(−1)n
T T’=T/L −π/Τ π/Τ

Figure 11(a) Figure 11(b)

11. Joke extra: You can have an extra score, if you email a powerpoint file with step-by-step playing
mode in a week to: [email protected] and [email protected] (5%). The score will
be given 3 - 5 based on novelty and funny level.

Useful Formulations:

∞ ∞
1. z-transform: X ( z ) = ∑ x[n ]z −n . 2. Fourier transform: X ( e jω ) = ∑ x[n ]e − jnω
n = −∞ n = −∞
∞ ∞
3. Convolution: y (n) = x[n] * h[n] = ∑ x[n − k ]h[k ] = ∑ h[n − k ]x[k ]
k = −∞ k = −∞
4. Inverse z-transform: h[n] = 1 ∫ H ( z ) z ( n −1) dz or h[n] = 1 ∫ " H ( p −1) p − n −1 dp
2πj C 2πj C
1 π jω j ωn
5. Inverse Fourier transform: x[ n] = ∫ X (e )e dω
2π − π