70 Continuous-Time Fourier Transform (Chapter 6)

6.10 For each differential/integral equation below that defines a LTI system with input x and output y, find the
frequency response H of the system. (Note that the prime symbol denotes differentiation.)
(a) y′′ (t) + 5y′ (t) +Ry(t) + 3x′ (t) − x(t) = 0;
t
(b) y′ (t) + 2y(t) + −∞ 3y(τ)dτ + 5x′ (t) − x(t) = 0; and
(c) y (t) + 5y (t) + 6y(t) = x′ (t) + 11x(t).
′′ ′

Answer (b).
First, we differentiate the given equation with respect to t. This yields

( dtd )2 y(t) + 2 dtd y(t) + 3y(t) + 5( dtd )2 x(t) − dtd x(t) = 0.

Taking the Fourier transform of both sides of the above equation, we obtain

F{( dtd )2 y(t) + 2 dtd y(t) + 3y(t) + 5( dtd )2 x(t) − dtd x(t)} = 0
⇒ F{( dtd )2 y(t)} + 2F{ dtd y(t)} + 3F{y(t)} + 5F{( dtd )2 x(t)} − F{ dtd x(t)} = 0
⇒ ( jω)2Y (ω) + 2( jω)Y (ω) + 3Y (ω) + 5( jω)2 X(ω) − ( jω)X(ω) = 0
⇒ − ω 2Y (ω) + j2ωY (ω) + 3Y (ω) − 5ω 2 X(ω) − jωX(ω) = 0
⇒ [−ω 2 + j2ω + 3]Y (ω) = [5ω 2 + jω]X(ω).

Therefore, the frequency response H(ω) of the system is

Y (ω) 5ω 2 + jω
H(ω) = = .
X(ω) −ω 2 + 2 jω + 3

Copyright c 2012–2020 Michael D. Adams Edition 2020-04-15

 71

6.11 For each frequency response H given below for a LTI system with input x and output y, find the differential
equation that characterizes the system.
jω
(a) H(ω) = ; and
1 + jω
jω + 12
(b) H(ω) = .
− jω 3 − 6ω 2 + 11 jω + 6
Answer (b).
From the given frequency response H(ω), we can write

Y (ω) jω + 12
=
X(ω) − jω 3 − 6ω 2 + 11 jω + 6
⇒ [− jω 3 − 6ω 2 + 11 jω + 6]Y (ω) = [ jω + 12 ]X(ω)
⇒ − jω 3Y (ω) − 6ω 2Y (ω) + 11 jωY (ω) + 6Y (ω) = jωX(ω) + 12 X(ω)
⇒ − jω 3Y (ω) − 6ω 2Y (ω) + 11 jωY (ω) + 6Y (ω) − jωX(ω) − 12 X(ω) = 0.

Taking the inverse Fourier transform of both sides of the above equation, we obtain

F−1 {− jω 3Y (ω) − 6ω 2Y (ω) + 11 jωY (ω) + 6Y (ω) − jωX(ω) − 21 X(ω)} = 0

⇒ F−1 {( jω)3Y (ω)} + 6F−1 {( jω)2Y (ω)} + 11F−1 { jωY (ω)} + 6F−1 {Y (ω)}
− F−1 { jωX(ω)} − 21 F−1 {X(ω)} = 0
⇒ ( dtd )3 y(t) + 6( dtd )2 y(t) + 11 dtd y(t) + 6y(t) − dtd x(t) − 12 x(t) = 0.

Edition 2020-04-15 Copyright c 2012–2020 Michael D. Adams

72 Continuous-Time Fourier Transform (Chapter 6)

6.12 Consider the LTI system with impulse response

h(t) = δ (t) − 300 sinc 300πt.

Using frequency-domain methods, find the response y of the system to the input

x(t) = 21 + 34 cos 200πt + 12 cos 400πt − 14 cos 600πt.

Answer.
Let X(ω), Y (ω), and H(ω) denote the Fourier transforms of x(t), y(t), and h(t), respectively.
We begin by finding the frequency response H(ω) of the system. This process yields
π
H(ω) = F{δ (t)} − F{ 300
π sinc 300πt}
ω
= 1 − rect( 600 π)
(
1 |ω| > 300π
=
0 otherwise.

Next, we determine the Fourier transform X1 (ω) of the input signal. We have

X1 (ω) = F{ 21 } + 34 F{cos 200πt} + 12 F{cos 400πt} − 41 F{cos 600πt}

= πδ (ω) + 43 π[δ (ω − 200π) + δ (ω + 200π)] + 21 π[δ (ω − 400π) + δ (ω + 400π)]
− 41 π[δ (ω − 600π) + δ (ω + 600π)].

Since the system is LTI, we can write

Y (ω) = X1 (ω)H(ω)
= 21 π[δ (ω − 400π) + δ (ω + 400π)] − 14 π[δ (ω − 600π) + δ (ω + 600π)].

Taking the inverse Fourier transform of Y (ω), we obtain

y(t) = F−1 {Y (ω)}

= 21 F−1 {π[δ (ω − 400π) + δ (ω + 400π)]} − 14 F−1 {π[δ (ω − 600π) + δ (ω + 600π)]}
= 12 cos 400πt − 14 cos 600πt.

Copyright c 2012–2020 Michael D. Adams Edition 2020-04-15

 73

6.13 Consider the LTI system with input v0 and output v1 as shown in the figure below, where R = 1 and L = 1.

R
i
+ +

v0 L v1

− −

(a) Find the frequency response H of the system.

(b) Determine the magnitude and phase responses of the system.
(c) Find the impulse response h of the system.

Answer (a).
From basic circuit analysis, we can write

v1 (t) = L dtd [ R1 [v0 (t) − v1 (t)]]

L d L d
⇒ v1 (t) = R dt v0 (t) − R dt v1 (t)

Taking the Fourier transform of the preceding equation, we have

V1 (ω) = RL F{ dtd v0 (t)} − RL F{ dtd v1 (t)}

L
⇒ V1 (ω) = R jωV0 (ω) − RL jωV1 (ω)
⇒ [1 + RL jω]V1 (ω) = L
R jωV0 (ω)

V1 (ω)
H(ω) =
V0 (ω)
L
jω
R
=
1 + RL jω
jω
= 1+ jω .

Answer (b).
Taking the magnitude of the frequency response H(ω), we obtain
| jω|
|H(ω)| =
|1 + jω|
|ω|
=√ .
1 + ω2
Taking the argument of the frequency response H(ω), we obtain

arg H(ω) = arg jω − arg(1 + jω)

= π
2 sgn ω − tan−1 ω.

As an aside, we note that

(
π
2 ω >0
arg jω =
− π2 ω <0
π
= 2 sgn ω.

Edition 2020-04-15 Copyright c 2012–2020 Michael D. Adams

74 Continuous-Time Fourier Transform (Chapter 6)

Answer (c).
Taking the inverse Fourier transform of the frequency response H(ω) yields

h(t) = F−1 { jω( 1+1jω )}

d −1 1
= dt F { 1+ jω }
d −t
= dt [e u(t)]
−t
= −e u(t) + δ (t)e−t
= −e−t u(t) + δ (t).

Copyright c 2012–2020 Michael D. Adams Edition 2020-04-15

 75

6.18 Consider the system shown below in Figure A with input x and output x̂, where this system contains a LTI
subsystem with impulse response g. The Fourier transform G of g is given by
(
2 |ω| ≤ 100π
G(ω) =
0 otherwise.

Let X, X̂, Y , and Q denote the Fourier transforms of x, x̂, y, and q, respectively.

X1 (ω)
sin 1000π t sin 1000π t
1

x y q x̂
× × g

(a) ω
−100π 0 100π
(b)

(a) Suppose that X(ω) = 0 for |ω| > 100π. Find expressions for Y , Q, and X̂ in terms of X.
(b) If X = X1 where X1 is as shown in Figure B, sketch Y , Q, and X̂.

Answer (a).
From the system block diagram, we have

Y (ω) = F{x(t) sin 1000πt}

= F{ 21j [e j1000π t − e− j1000π t ]x(t)}
= 1
2 j F{e
j1000π t
x(t)} − 21j F{e− j1000π t x(t)}
1 1
= 2 j X(ω − 1000π) − 2 j X(ω + 1000π), (5.1)
Q(ω) = F{y(t) sin 1000πt}
= F{ 21j [e j1000π t − e− j1000π t ]y(t)}
= 1
2 j F{e
j1000π t
y(t)} − 21j F{e− j1000π t y(t)}
1 1
= 2 j Y (ω − 1000π) − 2 j Y (ω + 1000π), and (5.2)
X̂(ω) = G(ω)Q(ω). (5.3)

Substituting the expression for Y (ω) from (5.1) into (5.2), we have
1 1
Q(ω) = 2 j Y (ω − 1000π) − 2 j Y (ω + 1000π)
1 1 1
= 2 j [ 2 j X([ω − 1000π] − 1000π) − 2 j X([ω − 1000π] + 1000π)]
− 21j [ 21j X([ω + 1000π] − 1000π) − 21j X([ω + 1000π] + 1000π)]
= − 14 X(ω − 2000π) + 41 X(ω) + 14 X(ω) − 41 X(ω + 2000π)
1 1 1
= 2 X(ω) − 4 X(ω − 2000π) − 4 X(ω + 2000π). (5.4)

Combining (5.3) and (5.4) and using the fact that X(ω) = 0 for |ω| > 100π, we have

X̂(ω) = G(ω)Q(ω)
= 2( 12 X(ω))
= X(ω).

Edition 2020-04-15 Copyright c 2012–2020 Michael D. Adams

76 Continuous-Time Fourier Transform (Chapter 6)

Answer (b).
The frequency spectra of the various signals are plotted below, where ωc = 1000π and ωb = 100π.

Y (ω )
j
2

ω
−2ωc −ωc − ωb −ωc −ωc + ωb −ωb ωb ωc − ωb ωc ωc + ωb 2ω c

− 2j

(a)
Q(ω )
1
2

1
4

ω
−2ωc − ωb −2ωc −2ωc + ωb −ωb ωb 2ω c − ω b 2ω c 2ωc + ωb

− 14

(b)
X̂(ω )

ω
−2ωc −ωc −ωb ωb ωc 2ω c

(c)

Copyright c 2012–2020 Michael D. Adams Edition 2020-04-15

 77

6.19 Consider the system shown below in Figure A with input x and output y. Let X, P, S, H, and Y denote the
Fourier transforms of x, p, s, h, and y, respectively. Suppose that
∞
n 1 ω
∑



p(t) = δ t − 1000 and H(ω) = 1000 rect 2000π .
n=−∞

(a) Derive an expression for S in terms of X. Derive an expression for Y in terms of S and H.
(b) Suppose that X = X1 , where X1 is as shown in Figure B. Using the results of part (a), plot S and Y . Indicate
the relationship (if any) between the input x and output y of the system.
(c) Suppose that X = X2 , where X2 is as shown in Figure C. Using the results of part (a), plot S and Y . Indicate
the relationship (if any) between the input x and output y of the system.
p

x s y
× h

(a)
X1 (ω) X2 (ω)

1 1

ω ω
−1000π 1000π −2000π 2000π
(b) (c)
Answer (a).
1
Since p(t) is periodic with period T = 1000 (and frequency ω0 = 2000π), it can be expressed in terms of a
Fourier series as
∞
p(t) = ∑ ck e j2000π kt .
k=−∞

Using the Fourier series analysis equation, we compute ck as

−1 Z 1/2000
ck = 1
1000 δ (t)e− j2000π kt dt
−1/2000
= 1000.

Combining the above equations, we have

∞
p(t) = 1000 ∑ e j2000π kt .
k=−∞

From the system block diagram, we have

s(t) = x(t)p(t)
!
∞
j2000π kt
= x(t) 1000 ∑ e
k=−∞
∞
= 1000 ∑ x(t)e j2000π kt .
k=−∞

Edition 2020-04-15 Copyright c 2012–2020 Michael D. Adams

78 Continuous-Time Fourier Transform (Chapter 6)

Taking the Fourier transform of both sides of the preceding equation, we obtain
∞
S(ω) = 1000 ∑ X(ω − 2000πk).
k=−∞

From the system block diagram, we have

y(t) = s(t) ∗ h(t).

Taking the Fourier transform of both sides of the preceding equation, we obtain

Y (ω) = H(ω)S(ω)
(
1
S(ω) |ω| < 1000π
= 1000
0 otherwise
 ∞
∑ X(ω − 2000πk) |ω| < 1000π


= k=−∞

0 otherwise.

Answer (b).
We observe that X(ω) = 0 for |ω| < 1000π. Therefore, the copies of the original spectrum of X(ω) in S(ω) do
not overlap. A plot of S(ω) is provided below. The plot of Y (ω) is identical to that of X(ω) = X1 (ω) given in
the problem statement.

S(ω)

1000

ω
−1000π 0 1000π

Since X(ω) = Y (ω), the input and output of the system are identical.

Answer (c).
We observe that X(ω) 6= 0 for some ω satisfying |ω| > 1000π. Therefore, the copies of the original spectrum
in S(ω) overlap, resulting in aliasing. The plots of S(ω) and Y (ω) are given below.

S(ω) Y (ω)

1000 1

ω ω
−1000π 0 1000π −1000π 0 1000π

We can see that Y (ω) does not in any way resemble X(ω). The input x(t) and output y(t) are not related, except
in the sense that the output is an extremely distorted version of the input, with the distortion being caused by
aliasing.

Copyright c 2012–2020 Michael D. Adams Edition 2020-04-15

 79

6.28 A function x is bandlimited to 22 kHz (i.e., only has spectral content for frequencies f in the range [−22000, 22000]).
Due to excessive noise, the portion of the spectrum that corresponds to frequencies f satisfying | f | > 20000 has
been badly corrupted and rendered useless. (a) Determine the minimum sampling rate for x that would allow
the uncorrupted part of the spectrum to be recovered. (b) Suppose now that the corrupted part of the spectrum
were eliminated by filtering prior to sampling. In this case, determine the minimum sampling rate for x.

Answer (a).
Let y denote the signal obtained from x after sampling and reconstruction. Let X and Y denote the Fourier
transforms of x and y, respectively. The spectrum X has a form something like that shown in the figure below.

X(ω )

corrupted corrupted

ω
−44000π 44000π
−40000π 40000π

Since we only wish to be able to recover the uncorrupted part of the spectrum of x, it does not matter if aliasing
occurs in the range of frequencies where the spectrum has already been corrupted. If we choose the sampling
rate to be 84000π, we obtain the spectrum shown in the figure below, after impulse sampling.

Y (ω )

aliasing aliasing

ω
−84000π −44000π 44000π 84000π
−40000π 40000π

From this plot, we can see that aliasing only occurs in the corrupted part of the spectrum. Thus, we can recover
the uncorrupted part of the spectrum by lowpass filtering. Using a lower sampling rate would cause aliasing to
occur in the uncorrupted part of the spectrum. Thus, the minimum sampling rate required is 84000π rad/s (or
equivalently, 42 kHz).

Answer (b).
Since the corrupted part of the spectrum has been removed (i.e., set to zero), the new signal is bandlimited to
frequencies in the range [−40000π, 40000π]. From the sampling theorem, we must sample the signal at

ωs = 2(40000π)
= 80000π.

Thus, a sampling rate of 80000π rad/s (or equivalently, 40 kHz) is required.

Edition 2020-04-15 Copyright c 2012–2020 Michael D. Adams

80 Continuous-Time Fourier Transform (Chapter 6)

6.101 (a) Consider a frequency response H of the form

M−1
∑k=0 ak ω k
H(ω) = N−1
,
∑k=0 bk ω k

where ak and bk are complex constants. Write a MATLAB function called freqw that evaluates a function of
the above form at an arbitrary number of specified points. The function should take three input arguments:
1) a vector containing the ak coefficients;
2) a vector containing the bk coefficients; and
3) a vector containing the values of ω for which to evaluate H(ω).
The function should generate two return values:
1) a vector of function values; and
2) a vector of points at which the function was evaluated.
If the function is called with no output arguments (i.e., the nargout variable is zero), then the function should
plot the magnitude and phase responses before returning. (Hint: The polyval function may be helpful.)
(b) Use the function developed in part (a) to plot the magnitude and phase responses of the system with the
frequency response
16.0000
H(ω) = .
1.0000ω 4 − j5.2263ω 3 − 13.6569ω 2 + j20.9050ω + 16.0000
For each of the plots, use the frequency range [−5, 5].
(c) What type of ideal frequency-selective filter does this system approximate?

Answer (a).
The freqw function can be implemented with the code below.

Listing 5.1: freqw.m

function [ freqresp , omega ] = freqw ( ncoefs , dcoefs , omega )

freqresp = polyval ( ncoefs , omega ) ./ polyval ( dcoefs , omega );

% If no output arguments were specified, plot the frequency response.

if nargout == 0

% Compute the magnitude response as a unitless quantity.

magresp = abs ( freqresp );

% Compute the phase response.

phaseresp = angle ( freqresp );

% On the first of two graphs, plot the magnitude response.

subplot (2 , 1, 1);
plot ( omega , magresp );
title ( ’ Magnitude Response ’);
xlabel ( ’ Frequency ( rad /s) ’);
ylabel ( ’ Magnitude ( unitless ) ’);

% On the second of two graphs, plot the phase response.

subplot (2 , 1, 2);
plot ( omega , phaseresp );

Copyright c 2012–2020 Michael D. Adams Edition 2020-04-15

 81

title ( ’ Phase Response ’);

xlabel ( ’ Frequency ( rad /s) ’);
ylabel ( ’ Angle ( rad ) ’);

end

Using the freqw function, we can generate the necessary plots with the following few lines of code:
ncoefs = [16];
dcoefs = [1.0000 (-j * 5.2263) ( -13.6569) (j * 20.9050) 16.0000];
freqw ( ncoefs , dcoefs , linspace ( -5 , 5, 500));

Answer (b).
The magnitude and phase responses are shown in the figure below.
Magnitude Response
1.4

1.2
Magnitude (unitless)

0.8

0.6

0.4

0.2

0
−5 −4 −3 −2 −1 0 1 2 3 4 5
Frequency (rad/s)

Phase Response
4

2
Angle (rad)

−2

−4
−5 −4 −3 −2 −1 0 1 2 3 4 5
Frequency (rad/s)

Answer (c).
The system approximates a lowpass filter with a cutoff frequency somewhere in the vicinity of 2 rad/s.

Edition 2020-04-15 Copyright c 2012–2020 Michael D. Adams

82 Continuous-Time Fourier Transform (Chapter 6)

6.103 (a) Use the butter and besself functions to design a tenth-order Butterworth lowpass filter and tenth-order
Bessel lowpass filter, each with a cutoff frequency of 10 rad/s.
(b) For each of the filters designed in part (a), plot the magnitude and phase responses using a linear scale for
the frequency axis. In the case of the phase response, plot the unwrapped phase (as this will be helpful later in
part (d) of this problem). (Hint: The freqs and unwrap functions may be helpful.)
(c) Consider the magnitude responses for each of the filters. Recall that an ideal lowpass filter has a magnitude
response that is constant in the passband. Which of the two filters more closely approximates this ideal behav-
ior?
(d) Consider the phase responses for each of the filters. An ideal lowpass filter has a phase response that is a
linear function. Which of the two filters has a phase response that best approximates a linear (i.e., straight line)
function in the passband?

Answer (a,b).

The frequency responses for the two filters can be plotted using the code shown below.

% Choose a filter type.

filtertype = ’ butterworth ’;
% filtertype = ’bessel’;

% Set the cutoff frequency for the filter.

wc = 10;

% Calculate the transfer function coefficients for the filter.

switch filtertype
case { ’ butterworth ’}
[ tfnum , tfdenom ] = butter (10 , wc , ’s ’);
case { ’ bessel ’}
[ tfnum , tfdenom ] = besself (10 , wc );
end

% Calculate the magnitude and phase responses.

[ freqresp , omega ] = freqs ( tfnum , tfdenom , linspace ( -20 , 20 , 500));
magresp = abs ( freqresp );
phaseresp = unwrap ( angle ( freqresp ));

% Plot the magnitude and phase response (using the unwrapped phase).
clf
subplot (2 , 1, 1);
plot ( omega , magresp );
title ( ’ Magnitude Response ’);
xlabel ( ’ Frequency ( rad /s) ’);
ylabel ( ’ Magnitude ( unitless ) ’);
subplot (2 , 1, 2);
plot ( omega , phaseresp );
title ( ’ Phase Response ’);
xlabel ( ’ Frequency ( rad /s) ’);
ylabel ( ’ Angle ( rad ) ’);

The above code produces the plots shown below.

Copyright c 2012–2020 Michael D. Adams Edition 2020-04-15

 83

Magnitude Response
Magnitude Response
1
1.2

1 0.8

Magnitude (unitless)
Magnitude (unitless)
0.8 0.6

0.6
0.4
0.4
0.2
0.2
0
0 −20 −15 −10 −5 0 5 10 15 20
−20 −15 −10 −5 0 5 10 15 20
Frequency (rad/s)
Frequency (rad/s)
Phase Response
Phase Response
0
0

−5
−5

Angle (rad)
Angle (rad)

−10
−10

−15
−15

−20
−20

−25
−25 −20 −15 −10 −5 0 5 10 15 20
−20 −15 −10 −5 0 5 10 15 20
Frequency (rad/s)
Frequency (rad/s)

(a) (b)

Answer (c).
In the passband, the magnitude response of the Butterworth filter is much flatter than that of the Bessel filter.
As it turns out, Butterworth filters have very flat magnitude responses in the passband.

Answer (d).
The Bessel filter has a phase response that is closer to having linear phase than the Butterworth filter. As it turns
out, Bessel filters tend to have phase responses that are approximately linear in the passband.

Edition 2020-04-15 Copyright c 2012–2020 Michael D. Adams