Question 1: Compute the 8-point DFT of the following sequence using a FFT algorithm.

x[n] = {2, 3, 4, 1, 1, 4, 3, 2}.

Report your computations (upto 2 decimal places) in the following table. Specify whether you have used Decimation-
in-Time or Decimation-in-Frequency algorithm by fill-in-the-blank given below the table. (8 M)
Answer:

k=0 k=1 k=2 k=3 k=4 k=5 k=6 k=7

Stage-2 10 1−i −4 1+i 10 −1 + i 4 −1 − i
Stage-3 20 1+0.4142i −4 − 4i 1+2.4142i 0 1−2.4142i −4 + 4i 1−0.4142i
X[0] X[1] X[2] X[3] X[4] X[5] X[6] X[7]

Table 1: Decimation-in-Time.

Stage-2 10 10 −4 −4i 1−i 1.4142i 1+i 1.4142i

Stage-3 20 0 −4 − 4i −4 + 4i 1+0.4142i 1−2.4142i 1+2.4142i 1−0.4142i
X[0] X[4] X[2] X[6] X[1] X[5] X[3] X[7]

Table 2: Decimation-in-Frequency.

Question 2: Consider the following anticausal discrete-time system.

(i) Find the transfer function H(z) of this anticausal system; clearly specify the ROC. (5 M)
Answer: Notice that
k
V (z) = X(z) − z −1 V (z),
2
k −1
Y (z) = V (z) + z V (z).
3
1 1+ k
3 z −1
Therefore, V (z) = k
X(z); subsequently, Y (z) = k
X(z). Now, since the system is anticausal,
1+ 2 z −1 1+ 2 z −1
1+ k
3 z −1 n |k| o
H(z) = k
, ROC = z ∈ Ce : 0 ≤ |z| < . (1)
1+ 2 z −1 2

(ii) For what range of k ∈ R is this system stable? (3 M)

Answer: A discrete-time LTI system is BIBO stable if and only if the ROC of its transfer function contains
the unit circle. Therefore, by (1), the range of k for stability is |k| > 2; in other words,
k ∈ (−∞, −2) ∪ (2, ∞).
(iii) Find y[n] if k = 6 and x[n] = (0.5)n for all n ∈ Z. (4 M)
1 + 2z −1
Answer: For k = 6, we have H(z) = . By the eigenfunction property,
1 + 3z −1
5
y[n] = H(0.5) (0.5)n = (0.5)n ∀n ∈ Z.
7
Remark 1. Note that the z-transform does not exist for the given two-sided input sequence.

Question 3: Let y[n] := x[2n]. Find Y (eiω ) in terms of X(eiω ), where (8 M)

F F
−→ X(eiω )
x[n] ←− and −→ Y (eiω ).
y[n] ←−

Answer:
∞
X ∞
X
Remark 2. For an arbitrary sequence g[·], g[2n] and g[m] need not be equal. However, the equality
n=−∞ m=−∞
holds if g[m] = 0 when m is odd.

Consider a sequence v[·] given by

 1 + (−1)n   1 + eiπn 
v[n] = x[n] = x[n].
2 2
Therefore,
X(eiω ) + X ei(ω−π)


iω X(eiω ) + X(−eiω )
V (e ) = = (2)
2 2
Notice that (
x[n], if n is even
v[n] = (3)
0, if n is odd.
Subsequently, v[2n] = x[2n] for all n ∈ Z.
∞
X ∞
X
−iωn
iω
Y (e ) = y[n] e = v[2n] e−iωn
n=−∞ n=−∞
| {z }
g[2n]
∞
X iωm
= v[m] e− 2 (follows from Remark 2)
m=−∞
iω
iω

iω
X e2 +X −e2
=V e 2 = (refer Eq. (2)).
2
 Question 4: Let x[·] and y[·] denote the input and output, respectively. Either prove or disprove the following
assertions.
(i) The following system is linear time-invariant (4 M)
 n
X
x[k], if n ≤ 0





k=−∞
y[n] := 0 n
 X X
x[k] + k x[k], if n ≥ 1.




k=−∞ k=1

Answer: Notice that

δ[n] 7→ u[n]
δ[n − 2] 7→ 2u[n − 2].
Therefore, the given system is time-variant.
Remark 3. Note that the given system is indeed linear.
 n  n
X X
x1 [k], if n ≤ 0 x2 [k], if n ≤ 0

 


 

 
k=−∞ k=−∞
x1 [n] 7→ y1 [n] = 0 n and x2 [n] 7→ y2 [n] = 0 n
 X X  X X
x1 [k] + k x1 [k], if n ≥ 1 x2 [k] + k x2 [k], if n ≥ 1

 


 

k=−∞ k=1 k=−∞ k=1
 n
X

ax1 [k] + bx2 [k] , if n ≤ 0





k=−∞
x[n] := ax1 [n] + bx2 [n] 7→ y[n] = 0 n = ay1 [n] + by2 [n].
 X
X

ax1 [k] + bx2 [k] + k ax1 [k] + bx2 [k] , if n ≥ 1




k=−∞ k=1

(ii) y[n] := n1 x[n] u[n−1] is a BIBO stable1 time-variant system. (4 M)

Answer: Note that
δ[n] 7→ 0
1
δ[n − k] 7→ δ[n − k], where k ≥ 1.
k
Therefore, the given system is time-variant.
Take an arbitrary bounded input sequence x : Z → C. Since x[·] is bounded, there exists c ∈ [0, ∞) such that
|x[n]| ≤ c for all n ∈ Z. Let v[n] := n1 u[n − 1]. Notice that |v[n]| ≤ 1 for all n ∈ Z. Therefore,
|y[n]| ≤ c, ∀n ∈ Z;
consequently, the given system is BIBO stable.

1
A system is said to be BIBO stable if every bounded input produces a bounded output.
 F
−→ H(eiω ).
Question 5: Consider a discrete-time LTI system with impulse response h[·]. Let h[n] ←−
(i) Using the following information, show that h[·] is finite duration. (7 M)
(a) The system is causal. (b) H(eiω ) = H ∗ (e−iω ).
(c) DTFT of the sequence g[n] := h[n + 1] is real-valued.
Answer:
(a) Since the system is causal, h[n] = 0 for n < 0.
F
(b) Recall that h∗ [n] ←−
−→ H ∗ (e−iω ). Therefore, H(eiω ) = H ∗ (e−iω ) implies that h[·] is a real-valued sequence.
F
−→ X(eiω ), where x[·] is a real-valued sequence, we have
(c) Recall that for x[n] ←−

F
F
−→ Re X(eiω ) −→ iIm X(eiω ) .



Ev x[n] ←− and Od x[n] ←−
Now, since G(eiω ) is real-valued, we can conclude that g[n] is an even sequence. Furthermore, since
g[n] := h[n + 1], where h[·] is a causal sequence, we have g[n] = 0 for n < −1.
Subsequently, it follows that h[n] = {h[0], h[1], h[2]} is a sequence of length 3. Thus,
2
X
iω
H(e ) = h[n] e−iωn . (4)
n=0

(ii) Subsequently, using the following information, find h[·]. (5 M)

Z π
1
iπ
(a) H(e ) = 0. (b) H(eiω ) dω = 2.
2π −π
Answer: Since g[n] := h[n + 1] is an even sequence, we have h[0] = h[2].
(a) Since H(eiπ ) = 0, it follows from (4) that h[0] − h[1] + h[2] = 0.
Z π
1
(b) Using H(eiω ) dω = 2, we get h[0] = 2.
2π −π
Therefore, h[n] = {2, 4, 2}.
 2
Question 6: Let x[n] = 1+√ 3i
{1, 0, 1, 2, 1, 0} be a sequence of length 6. We are interested in computing the
DTFT of x[·] at the following frequencies:
π 2πk
ωk = + for k = 0, . . . , 3.
3 4
(i) Construct a 4-point sequence y[·] from x[·] such that Y [k] = X(eiωk ) for k = 0, . . . , 3, where
F DF T
−→ X(eiω )
x[n] ←− and y[n] ←−
−→ Y [k];
use DTFT properties, and relation among DTFT, DFS & DFT. Explain why your constructed sequence y[·]
satisfies the desired property. (8 M)
iπ π

iω i(ω− )
Answer: Let v[n] := e 3 x[n]. Thus, V (e ) = X e 3 ; and hence,
i2πk

X(eiωk ) = V e 4 for k = 0, . . . , 3. (5)
We can obtain X(eiωk ) for k = 0, . . . , 3 from N = 4 uniformly spaced samples of DTFT of v[·]. This sampling
of V (eiω ) leads to periodic extension of v[n]; and the corresponding periodic sequence with period 4 is given by
∞
X
ỹ[n] = v[n + 4r]. (6)
r=−∞

Note that the undersampling (N = 4) of a 6-point sequence v[·] leads to aliasing.

The finite length sequence corresponding to ỹ[·] can be obtained as follows:
{y[0], y[1], y[2], y[3]} = {v[0], v[1], v[2], v[3]} + {v[5], v[6], 0, 0}. (7)
Thus, if follows from (5)-(7) that y[n] = {2, 0, 1, 2} is the desired 4-point sequence.

(ii) Using DFT of y[·], compute X(eiωk ) for k = 0, . . . , 3. (4 M)

Answer:
X(eiω0 )
        
Y [0] 1 1 1 1 2 5
X(eiω1 ) Y [1] 1 −i −1 i  0 1 + 2i
X(eiω2 ) = Y [2] = 1 −1 1 −1 1 =  1  .
        

X(eiω3 ) Y [3] 1 i −1 −i 2 1 − 2i

— Cheers —