Signals and Systems: Solution for Assignment 3

Problem 4.21. [20pt] Compute the Fourier transform of each of the following signals:

(a) e−αt cos ω0 t u(t), α > 0

 

Sol )
Z ∞  −αt
cos ω0 t u(t)e−jωt dt

e
−∞
Z ∞
= e−αt cos ω0 te−jωt dt
0
∞
ejω0 t + e−jω0 t −jωt
Z
= e−αt · ·e dt
0 2
1 ∞ h −αt j(ω0 −ω)t
Z i
= e ·e + e−αt · e−j(ω0 +ω)t dt
2 0
 
1 1 ∞ 1 ∞
= · e(−α+j(ω0 −ω))t − · e−(α+j(ω0 +ω))t
2 −α + j(ω0 − ω) 0 α + j(ω0 + ω) 0
 
1 1 1
= +
2 α − j(ω0 − ω) α + j(ω0 + ω)
1 1
= + .
2(α − jω0 + jω) 2(α + jω0 + jω)

 sin πt  h sin 2π(t−1) i

(b) πt π(t−1)

sin πt sin 2πt

Sol ) Let x1 (t) = πt and a(t) = πt such that
(
sin πt F 1 |ω| < π
x1 (t) = ←
→ X1 (jω) =
πt 0 |ω| > π,
(
sin 2πt F 1 |ω| < 2π
a(t) = ←
→ A(jω) = .
πt 0 |ω| > 2π

Using time shifting property of CTFT,

(
F e−jω |ω| < 2π
x2 (t) = a(t − 1) ←
→ X2 (jω) = e−jω A(jω) = .
0 |ω| > 2π

Using multiplication property of CTFT,

   ∞
sin πt sin 2π(t − 1)
Z
F 1 1
x1 (t)x2 (t) = ←
→ X1 (jω) ∗ X2 (jω) = X1 (jθ)X2 (j(ω − θ)) dθ.
πt π(t − 1) 2π 2π −∞

To perform the convolution of the two Fourier transforms, we may consider the condition of ω where
the overlap between the transforms exists as below: From Fig. 1, we can characterize four cases of the
convolution between X1 (jθ) and X2 (j(ω − θ)).

1
 (i) −3π < ω < −π (ii) −π < ω < π (iii) π < ω < 3π

Figure 1: Overlaps between X1 (jθ) and X2 (j(ω − θ)) depending on the conditions of ω.

(i) −3π < ω < −π (Fig. 1i),

Z 2π+ω
1 −jω 2π+ω jθ
Z
1 1 −j(ω−θ)
X1 (jω) ∗ X2 (jω) = e dθ = e e dθ
2π 2π −π 2π −π
1 −jω jθ 2π+ω 1 −jω h j(2π+ω) i
= e ·e = e e − e−jπ
2πj −π 2πj
1 −jω jω 1
= e (e + 1) = (1 + e−jω ).
2πj 2πj

(ii) −π < ω < π (Fig. 1ii),

Z π
1 −jω π jθ
Z
1 1
X1 (jω) ∗ X2 (jω) = e−j(ω−θ) dθ = e e dθ
2π 2π −π 2π −π
1 −jω jω π 1 −jω jπ
= e · e −π = e (e − e−jπ ) = 0.
2π 2πj

(iii) π < ω < 3π (Fig. 1iii),

Z π
1 −jω π
Z
1 1 −j(ω−θ)
X1 (jω) ∗ X2 (jω) = e dθ = e ejθ dθ
2π 2π −2π+ω 2π −2π+ω
1 −jω jθ π 1 −jω h jπ i
= e ·e = e e − ej(ω−2π)
2πj −2π+ω 2πj
1 −jω j
= e (−1 − ejω ) = (1 + e−jω ).
2πj 2π
1
(iv) Otherwise, there is no overlap thus 2π X1 (jω) ∗ X2 (jω) = 0.
 h 2π(t−1) i
To sum up, the CTFT of sinπtπt sinπ(t−1)

is
 1
−jω ) −3π < ω < −π
1  2πj (1 + e

j
X1 (jω) ∗ X2 (jω) = 2π (1 + e−jω ) π < ω < 3π .
2π 
0 otherwise


2
Problem 4.24. [30pt] Determine which, if any, of the real signals depicted in Fig. 2 have Fourier transforms
that satisfy each of the following conditions:

Figure 2

(1) Re{X(jω)} = 0

Sol )
X(jω) + X ∗ (jω)
Re{X(jω)} = = 0. (1)
2
F
Using conjugation property and time reversal property of CTFT, x∗ (−t) ←
→ X ∗ (jω).
Using linearity property of CTFT in Eq. (1),

x(t) + x∗ (−t)
= 0,
2
x(t) = −x(−t), (x(t) is real.)

thus x(t) is odd. Signal (a) satisfies this condition.

R∞
(2) −∞ X(jω) dω =0

Sol ) From the synthesis equation of CTFT,

Z ∞
1
x(0) = X(jω) dω = 0.
2π −∞

Signals (a), (b) and (c) satisfy this condition.

(3) X(jω) is periodic

Sol ) Suppose X(jω) is periodic with a period W such that

X(jω) = X(j(ω + W )). (2)

F F
→ X(jω), e−jW t x(t) ←
If x(t) ← → X(j(ω + W )) using frequency shifting property of CTFT.

3
From Eq. (2),

x(t) = e−jW t x(t),

x(t)(1 − e−jW t ) = 0. (3)

To satisfy Eq. (3), x(t) = 0 for all t or W t = 2πk for all integers, i.e.,
X 2π
x(t) = ck · δ(t − k).
W
k∈Z

Therefore, x(t) whose CTFT is periodic has impulse signals only. Thus, signal (b) satisfies this condition.
Alternatively, using duality property of CTFT, if X(jω) is periodic, x(t) is a train of impulses. See
relation of Fourier Series and CTFT in the page 8 of our lecture note. (Either answer is okay.)
FYI, Discrete-Time Fourier Transform (DTFT) is always periodic in frequency. We will discuss DTFT
later in our class.

4
Problem 4.38. [30pt] Let x(t) be any signal with Fourier transform X(jω). The frequency-shift property of
the Fourier transform may be stated as
F
ejω0 t x(t) ←
→ X(j(ω − ω0 )).

(a) Prove the frequency-shift property by applying the frequency shift to the analysis equation
Z ∞
X(jω) = x(t)e−jωt dt.
−∞

Sol )
Z ∞
X(j(ω − ω0 )) = x(t)e−j(ω−ω0 )t dt
Z−∞
∞
F
= x(t)ejω0 t e−jωt dt ←
→ x(t)ejω0 t .
−∞

(b) Prove the frequency-shift property by utilizing the Fourier transform of ejω0 t in conjunction with the
multiplication property of the Fourier transform.

Sol ) Let w(t) = ejω0 t such that

F
w(t) = ejω0 t ←
→ W (jω) = 2πδ(ω − ω0 ).

Using multiplication property of CTFT,

F 1
x(t)w(t) ←
→ X(jω) ∗ W (jω),
2π
F
x(t)w(t) ←
→ X(jω) ∗ δ(ω − ω0 ),
F
x(t)w(t) ←
→ X(j(ω − ω0 )).

5
Problem 4.32. [20pt] Consider an LTI system S with impulse response
sin (4(t − 1))
h(t) = .
π(t − 1)
Determine the output of S for each of the following inputs:
sin 4t
Sol ) Let h1 (t) = πt such that
(
sin 4t F 1 |ω| < 4
h1 (t) = ←
→ H1 (jω) = .
πt 0 |ω| > 4

h1 (t) is the impulse response of an ideal low-pass filter whose passband is in the range |ω| < 4.
Since h(t) = h1 (t − 1), h(t) is the impulse response of an ideal loss-pass filter shifted by one to the right.
Using time shifting property of CTFT,
(
F e−jω |ω| < 4
h(t) = h1 (t − 1) ←→ H(jω) = ,
0 |ω| > 4

thus the LTI system S still has its passband in the range |ω| < 4.

(a) x1 (t) = cos (6t + π2 )

Sol ) Let a1 (t) = cos 6t such that

F
a1 (t) = cos 6t ←
→ A1 (jω) = π [δ(ω − 6) + δ(ω + 6)] .
Using time shifting property of CTFT,
π F π π
x1 (t) = a1 (t + ) ← → ej 12 ω A1 (jω) = ej 12 ω [δ(ω − 6) + δ(ω + 6)] ,
12
thus x1 (t) has no frequency component in the range |ω| < 4.
Then, the CTFT of the output is
Y1 (jω) = H(jω)X1 (jω) = 0 for all ω,
therefore y1 (t) = 0.

sin (4(t+1))
(b) x3 (t) = π(t+1)

sin 4t
Sol ) Let a3 (t) = πt such that
(
sin 4t F 1 |ω| < 4
a3 (t) = ←
→ A3 (jω) =
πt 0 |ω| > 4.
Using time shifting property of CTFT,
(
F jω ejω |ω| < 4
x3 (t) = a3 (t + 1) ←
→ e A3 (jω) =
0 |ω| > 4.
Then, the CTFT of the output is
(
1 |ω| < 4
Y3 (jω) = H(jω)X3 (jω) =
0 |ω| > 4,
sin 4t
and its inverse transform is y3 (t) = πt .
Alternatively, Y3 (jω) = X3 (jω)e−jω and using time shifting property of CTFT, y3 (t) = x3 (t − 1) = sin 4t
πt .
(Either approach is okay.)