EE1101 : Signals and Systems JAN–MAY 2020

Tutorial 8 Solutions

t t
1. (a) Given f1 (t) = 2 rect and f2 (t) = rect Using the convolution property of Fourier Transform,
4 2

g(t)
G(jω) = F1 (jω)F2 (jω)
4 = 8 sinc(2ω/π)2 sinc(ω/π)
f1(t) f2(t)
= 16 sinc(2ω/π) sinc(ω/π)
2 1

* =
-2 0 2 t -1 0 1 t -3 -1 0 1 3 t
(d) The magnitude spectrum of G(jω) is plotted using
 matlab in Figure 1. Phase spectrum is plotted in Fig-
2(t + 3)
 , −3 ≤ t ≤ −1 ure 2.

4 , −1 < t ≤ 1
g(t) = f1 (t) ∗ f2 (t) =
2(3 − t)
 ,1 < t ≤ 3 16

0 , |t| ≥ 3


14
(b) Using the analysis equation,
Z∞ 12
G(jω) = g(t)e−jωt dt
−∞ 10
Z−1 Z1
|G(j!)|

−jωt
= 2(t + 3)e dt + 4e−jωt dt 8

−3 −1
Z3 6
−jωt
+ 2(3 − t)e dt
4
1
cx−1


the formulae xecx = c2 ecx and
R
RUsing 2
ecx = 1c ecx ,
−1  −jωt 1
(−2jωt − 2 − 6jω)e−jωt

4e 0
G(jω) = + -8 -6 -4 -2 0 2 4 6 8
−ω 2 −3 −jω −1 !
3
(2jωt + 2 − 6jω)e−jωt

+ Figure 1: Magnitude spectrum of G(jω) in Q2 generated using
−ω 2 1 matlab
(2 + 4jω)ejω − 2ej3ω 8 sin ω
= +
ω2 ω
(2 − 4jω)e−jω − 2e−j3ω
+
ω2 ∠G(jω)
4 cos(ω) − 8ω sin(ω) − 4 cos(3ω) 8 sin ω
= + π
ω2 ω
4[cos(ω) − cos(3ω)]
= − 7π 5π 3π π
ω2 2 − 2 − 2 −2 ω
8 sin(2ω) sin(ω) π 3π 5π 7π
= 2 2 2 2
ω2
= 16 sinc(2ω/π) sinc(ω/π)
−π
sin(πx)
where sinc(x) = πx .
(c) We know that
t  ωτ  Figure 2: Phase spectrum of G(jω) in Q2
rect ←→ τ sinc
τ 2π
t
f1 (t) = 2 rect ←→ 2(4) sinc(2ω/π) Note: ∠G(jω) can take either π or −π for values of ω
4 for which G(jω) is negative. Here we have taken π for
t
f2 (t) = rect ←→ 2 sinc(ω/π) when ω is positive and −π for when ω is negative to
2 imply the point that phase spectrum is an odd function
for real-valued time-domain signals.
2. The given signal x(t) can be written as, 4. (a)

x(t) = x1 (t) ∗ x2 (t) Z∞

t X(jω) = x(t)e−jωt dt
where, x1 (t) = rect(t) and x2 (t) = rect −∞
2
Z∞
= e−a|t| e−jωt dt
x1(t) x2(t) x(t) −∞
1 1 1
Z0 Z∞
* = = at −jωt
e e dt + e−at e−jωt dt
-0.5 0 0.5 t -1 0 1t -1.5 -0.5 0 0.5 1.5 t −∞ 0
(a−jω)t 0
∞
e(−a−jω)t
  
e
= +
Using the convolution property of Fourier Transform, a − jω −∞ −a − jω 0
1 1
= +
X(jω) = X1 (jω)X2 (jω) a − jω a + jω
ω ω 2a
= sinc 2 sinc = 2
2π  π a + ω2
 ω ω
= 2 sinc sinc
π 2π (b) From the duality property of Fourier transform, we
know that

if x(t) ←→ X(ω)
2 sin(3(ω − 2π))
3. (a) Given X(jω) = then X(t) ←→ 2π · x(−ω).
(ω − 2π)
t  ωτ  Thus from part (a), we have
2  ωτ 
rect ←→ τ sinc = sin (1)
τ 2π ω 2 2a
←→ 2π · e−a|ω|
a2 + t2
Comparing with standard form in (1), 1
=⇒ ←→ π · e−|ω|
1 + t2
t 2
rect ←→ sin(3ω)
6 ω
Using Parseval’s theorem we have,
Z∞  2 Z∞ h i2
Using the frequency shifting property of Fourier Trans- 1 1
dt = π · e−|ω| dω
form, 1 + t2 2π
−∞ −∞
t 2 sin(3(ω − 2π)) Z0 Z∞
ej2πt rect ←→ π 2ω π
6 (ω − 2π) = e dω + e−2ω dω
2 2
−∞ 0
2ω
0 ∞
π e−2ω
 
π e
= +
(b) Given X(jω) = cos(4ω + π3 ) 2 2 −∞ 2 −2 0
π
=
2
π
X(jω) = cos(4ω + )
3
1 h j(4ω+ π ) π
i 5. a) Z∞
= e 3 + e−j(4ω+ 3 )
2 X(jω) = x(t)e−jωt dt
 ej π3   e−j π3 
= e j4ω
+ e−j4ω −∞
2 2 Z∞
X(0) = x(t)e−j(0)t dt
−∞
We know that, Z∞
X(0) = x(t)dt
δ(t) ←→ 1 −∞

Z0 Z1
By using time shifting property, X(0) = 1dt + (−t + 1)dt
−1 0
δ(t + 4) ←→ ej4ω Z2 Z3
δ(t − 4) ←→ e−j4ω + (t − 1)dt + 1dt
1 2
By using linearity property,
 ej π3   e−j π3  =⇒ X(0) = 3
x(t) = δ(t + 4) + δ(t − 4)
2 2
b) Z∞ F −1 {Re{X(jω)}}
1
x(t) = X(jω)ejωt dω
2π
−∞ 1
Z∞
1 0.5
x(0) = X(jω)ejω(0) dω
2π
−∞ -3 -2 -1 0 1 2 3 t
Z∞
1
x(0) = X(jω)dω 6. For the given signal, T0 = 1, ω0 = 2π.
2π
−∞
(a) Fourier series coefficients Pn of function p(t) :
Z∞ 1
X(jω)dω = 2πx(0) 1
Z4
−∞ Pn = p(t)e−jnω0 t dt, n 6= 0
T0
Z∞ −1
4
=⇒ X(jω)dω = 2π 1
Z0 Z4
−∞
= (1 + 4t)e−j2πnt dt + (1 − 4t)e−j2πnt dt
−1 0
4
2
Z∞ 4 sin (πn/4)
c) 1 = , n 6= 0.
x(t) = X(jω)ejωt dω π 2 n2
2π Z4
1

−∞ 1
Z∞ P0 = p(t)dt = .
1 4
x(1) = X(jω)ejω(1) dω −1
4
2π
−∞
Note: This can be solved using convolution property as
Z∞
1 well. Convolution of rect function with itself will give
x(1) = X(jω)ejω dω triangular function. In Fourier domain that manifests
2π
−∞ as sinc2 () function.
Z∞
X(jω)ejω dω = 2πx(1)
−∞
Z∞
=⇒ X(jω)ejω dω = 2π.0
−∞
Z∞
=⇒ X(jω)ejω dω = 0
−∞

d)
Using Parseval’s theorem,
Z∞ Z∞
1 2 Figure 3: Sketch for Pn vs. n in Q6(a)
|X(jω)| dω = |x(t)|2 dt
2π
−∞ −∞
Z∞ " Z0 Z1 (b) Fourier transform of function p(t) :
=⇒ |X(jω)|2 dω = 2π 12 dt + (1 − t)2 dt
By using Fourier series expansion,
−∞ −1 0
Z2 Z3 # ∞
X
+ 2
(t − 1) dt + 2
1 dt p(t) = Pn ej2πnt
n=−∞
1 2 ∞
  X
1 1 16π P (jω) = F{Pn ej2πnt }
= 2π 1 + + + 1 =
3 3 3 n=−∞
X∞
= Pn F{ej2πnt }
e) X(jω) + X ∗ (jω)
Re{X(jω)} = n=−∞
2 ∞
F −1 {X(jω)} + F −1 {X ∗ (jω)}
X
F −1 {Re{X(jω)}} = = 2πPn δ(ω − 2πn)
2 n=−∞
x(t) + x∗ (−t) ∞
8 sin2 (πn/4)
=
X
2 = δ(ω − 2πn)
n=−∞
πn2
x(t) + x(−t)
= −→ Ev{x(t)}
2
 Figure 5: Sketch for X(jω) in Q6(c)
Figure 4: Sketch for P (jω) vs. ω in Q6(a)

(c) y(t) = p(t) · x(t). Then the Fourier transform of y(t) is

1
Y (jω) = {P (jω) ∗ X(jω)}
2π
∞
1 n X o
= 2π Pk δ(ω − 2πk) ∗ X(jω)
2π
k=−∞
∞
X
= Pk X(j(ω − 2πk))
k=−∞
∞
X 4 sin2 (πk/4)
= X(j(ω − 2πk))
π2 k2
k=−∞ Figure 6: Sketch for Y (jω) in Q6(c)
(d) x(t) = sinc(t), has the Fourier transform
  8. Given ,
ω
X(jω) = rect .
2π z1 (t) = x(t) cos(ω1 t) + y(t) cos(ω2 t)
ω1 = 5W
1 ω2 = 7W
Y (jω) = {P (jω) ∗ X(jω)}
2π
∞
X 4 sin2 (πk/4) ω − 2πk
 
= rect .
π2 k2 2π For z1 (t) :
k=−∞

7. (a) z1 (t) = x(t) cos(ω1 t) + y(t) cos(ω2 t)

Z∞  ejω1 t + e−jω1 t   ejω2 t + e−jω2 t 
= x(t) + y(t)
X(jω) = e−at e−jωt dt 2 2
0
∞
e(−a−jω)t

= The frequency shift property of Fourier Transform states
−a − jω 0 that,
1
=
a + jω
f (t)ejω0 t ←→ F (j(ω − ω0 ))
Therefore,

Y (jω) = H(jω)X(jω) Using this property and linearity, Z1 (jω) can be written as,
1
= 1
(a + jω)2 Z1 (jω) = X(j(ω − ω1 )) + X(j(ω + ω1 ))+
2
(b)

Y (j(ω − ω2 )) + Y (j(ω + ω2 ))
−1
 
1 d
=j·
(a + jω)2 dω a + jω For z2 (t) :
Using frequency differentiation property, we have
  Z2 (jω) is Z1 (jω) bandlimited to 4W and 6W
d 1
j· ←→ t · e−at u(t) 1 
dω a + jω Z2 (jω) = X(j(ω − ω1 )) + X(j(ω + ω1 ))
1 2
=⇒ ←→ t · e−at u(t)
(a + jω)2
 Using frequency shifting property of Fourier Transform,
Z1 (jω)
1 
Z3 (jω) = Z2 (j(ω − ω1 )) + Z2 (j(ω + ω1 ))
2
1 1 
2 = X(j(ω − 2ω1 )) + 2X(jω) + X(j(ω + 2ω1 ))
4
1 1 
= X(jω) + X(j(ω − 2ω1 )) + X(j(ω + 2ω1 ))
2 4
-8W -7W -6W -5W -4W 4W 5W 6W 7W 8W ω (2)

For z4 (t) :
Figure 7: Sketch for Z1 (jω)
z4 (t) is the LPF output for the input z3 (t) with cutoff freq
= W . So, only the first summation term from eqn.(2) is
available as output in the form of z4 (t), i.e.:
Z2 (jω)
1 1
Z4 (jω) = X(jω) ⇒ z4 (t) = x(t)
2 2
1
2

Z4 (jω)
1
-6W -5W -4W 4W 5W 6W ω 2

Figure 8: Sketch for Z2 (jω)

1
-W 0 W ω
2 Z2 (j(w − w1 ))

Figure 10: Sketch for Z4 (jω)

1
4

w
−W W 9W 11W

1
2 Z2 (j(w + w1 ))

1
4

w
−11W −9W −W W

Z3 (jw)

1
2

w
−11W −9W −W W 9W 11W

Figure 9: Sketch for Z3 (jω)

For z3 (t) :

z3 (t) = z2 (t) cos(ωt)

1
z2 (t)ejω1 t + z2 (t)e−jω1 t


=
2