ssm

Contact me in order to access the whole complete document. Email: [email protected]

mtt
WhatsApp: https://wa.me/message/2H3BV2L5TTSUF1 Telegram: https://t.me/solutionmanual

Chapter 2
bb99
Problem 2.1
Let g(t) G(f ). Then
88@
1 f
g(at) G( ) (1)
|a| a
@
To prove this property, we note that
Z ∞
F [g(at)] = g(at) exp(−j2πf t)dt
−∞
ggm
Set τ = at. There are two cases that can arise, depending on whether the scaling factor a
is positive or negative. If a > 0, we get

1 ∞
Z    
f
maa

F [g(at)] = g(τ ) exp −j2π τ dτ

a −∞ a
 
1 f
= G
a a
On the other hand, if a < 0, the limits of integration are interchanged so that we have the
multiplying factor −(1/a) or, equivalently, 1/|a|. This completes the proof of Eq. (1).
iill..cc

Note that the function g(at) represents g(t) compressed in time by a factor a, whereas
the function G(f /a) represents G(f ) expanded in frequency by the same factor a. Thus,
the scaling property states that the compression of a function g(t) in the time domain is
equivalent to the expansion of its Fourier transform G(f ) in the frequency domain by the
same factor, or vice versa.
oom

1
m

complete document is available on https://unihelp.xyz/ *** contact me if site not loaded

PROBLEM 2.2 CHAPTER 2.

For the special case when a = −1, we readily find from Eq. (1) that

g(−t) G(−f ) (2)

In words, if a function g(t) has a Fourier transform given by G(f ), then the Fourier
transform of g(−t) is G(−f ).

Problem 2.2
(a) This property follows from the relation defining the inverse Fourier Transform by writ-
ing it in the form:
Z ∞
g(−t) = G(f ) exp(−j2πf t)df
−∞

and then interchanging t and f .

(b) To prove this property, we take the Fourier transform of g(t − t0 ) and then set τ =
(t − t0 ) to obtain
Z ∞
F [g(t − t0 )] = exp(−j2πf t0 ) g(τ ) exp(−j2πf τ )dτ
−∞

The time-shifting property states that if a function g(t) is shifted in the positive direc-
tion by an amount t0 , the effect is equivalent to multiplying its Fourier transform G(f ) by
the factor exp(−j2πf t0 ). This means that the amplitude of G(f ) is unaffected by the time
shift, but its phase is changed by the linear factor −2πf t0 .
(c) We note that
1
cos(2πfc t) = [exp(j2πfc t) + exp(−j2πfc t)]
2
and
 
t
Arect AT sinc(f T ) (1)
T
Therefore, applying the frequency-shifting property to the Fourier-transform pair of Eq.
(1), we get the desired result
AT
G(f ) = {sinc[T (f − fc )] + sinc[T (f + fc )]}
2

2
PROBLEM 2.2 CHAPTER 2.

1
In the special case of fc  , we may use the approximate result
T
 AT

 2
sinc[T (f − fc )], f > 0



G(f ) ' 0, f =0



 AT
sinc[T (f + fc )], f < 0

2

3
PROBLEM 2.3 CHAPTER 2.

Problem 2.3
(a) To prove this property, we first denote the Fourier transform of the product g1 (t)g2 (t)
by G12 (f ), so that we may write

g1 (t)g2 (t) G12 (f )

where Z ∞
G12 (f ) = g1 (t)g2 (t) exp(−j2πf t)dt
−∞

For g2 (t), we next substitute the inverse Fourier transform

Z ∞
g2 (t) = G2 (f 0 ) exp(−j2πf 0 t)df 0
−∞

in the integral defining G12 (f ) to obtain

Z ∞Z ∞
G12 (f ) = g1 (t)G2 (f 0 ) exp[−j2π(f − f 0 )t]df 0 dt
−∞ −∞

Define λ = f − f 0 . Then, interchanging the order of integration, we obtain

Z ∞ Z ∞
G12 (f ) = dλG2 (f − λ) g1 (t) exp(−j2πλt)dt
−∞ −∞

The inner integral is recognized simply as G1 (λ), so we may write

Z ∞
G12 (f ) = G1 (λ)G2 (f − λ)dλ
−∞

which is the desired result. This integral is known as the convolution integral expressed
in the frequency domain, and the function G12 (f ) is referred to as the convolution of
G1 (f ) and G2 (f ). We conclude that the multiplication of two signals in the time domain
is transformed into the convolution of their individual Fourier transforms in the frequency
domain. This property is known as the multiplication theorem.
In a discussion of convolution, the following shorthand notation is frequently used

G12 (f ) = G1 (f ) ? G2 (f )

Accordingly, we may use the following symbolic form:

g1 (t)g2 (t) G1 (f ) ? G2 (f ) (1)

4
PROBLEM 2.4 CHAPTER 2.

Note that convolution is commutative, that is,

G1 (f ) ? G2 (f ) = G2 (f ) ? G1 (f )

which follows directly from Eq. (1).

(b) This result follows directly by combining duality property and time-domain multipli-
cation property. We may thus state that the convolution of two signals in the time domain
is transformed into the multiplication of their individual transforms in the frequency do-
main. This property is known as the convolution theorem. Its use permits us to exchange a
convolution operation for a transform multiplication, an operation that is ordinarily easier
to manipulate. Accordingly, we may write:

g1 (t) ? g2 (t) = G1 (f )G2 (f )

where ? denotes convolution.

(c)
According to the convolution theorem:
Z t
g1 (τ )g2 (τ − t) d τ G1 (f )G2 (f )
−∞

Suppose we complex conjugate the G2 (f ), yielding G∗2 (f ), by definition if g2 (t) G2 (f ),

then g2∗ (t) G∗2 (f ). Therefore, we may go on to write:
Z ∞
g1 (τ )g2∗ (τ − t) d τ G1 (f )G∗2 (f )
-∞

which confirms the correlation theorem, entry 13 in Table 2.1.

Problem 2.4
This result follows from noting that the energy intensity |g(t)|2 may be expressed as the
product of two time functions, namely, g(t) and its complex conjugate g ∗ (t). The Fourier
transform of g ∗ (t) is equal to G∗ (−f ), by virtue of complex conjugation property. Then,
applying the multiplication theorem to the product g(t)g ∗ (t) and evaluating the result for
f = 0, we obtain the relation
Z ∞ Z ∞
∗
g(t)g (t)dt = G(λ)G∗ (λ)dλ
−∞ −∞

5
PROBLEM 2.5 CHAPTER 2.

which is equivalent to Z ∞ Z ∞
2
|g(t)| dt = |G(λ)|2 dλ
−∞ −∞

Let g (f ) denote the squared amplitude spectrum of the signal g(t) , as shown by

g (f ) = |G(f )|2

The quantity g (f ) is referred to as the energy spectral density of the signal g(t). To explain
this meaning of the definition, suppose g(t) denotes the voltage of a source connected to a
1-ohm load resistor. Then the quantity
Z ∞
|g(t)|2 dt
−∞

equals the total energy delivered by the source. According to Rayleigh’s theorem, this
energy equals the total area under the g (f ) curve. If follows, therefore, that the function
g (f ) is a measure of the density of the energy contained in g(t) in joules per Hertz. Note
that since in the special case of a real-valued signal the amplitude spectrum is an even
function of f , the energy spectral density of such a signal is symmetrical about the vertical
axis passing through the origin.

Problem 2.5
By the differentiation property:

 
dg(t)
F = j2πf G(f )
dt
1
= τ
[H(f ) exp(j2πf τ ) − H(f ) exp(−j2πf τ )]

2j
= τ
H(f ) sin(2πf τ )

6
PROBLEM 2.6 CHAPTER 2.

But H(f ) = τ exp(−πf 2 τ 2 )

1
∴ G(f ) = πf
exp(−πf 2 τ 2 ) sin(2πf T )

= exp(−πf 2 τ 2 ) sin(2πf
πf
T)

= 2T exp(−πf 2 τ 2 )sinc(2πf T )

lim G(f ) = 2T sinc(2πf T )

τ →0

Problem 2.6
(a)
Z ∞
G(f ) = g(t) exp(−j2πf t)dt
−∞
Z 0 Z ∞
= g(t) exp(−j2πf t)dt + g(t) exp(−j2πf t)dt
−∞ 0
Z 0 Z 0
= g(t) cos(2πf t)dt − jg(t) sin(2πf t)dt
−∞ −∞
Z ∞ Z ∞
+ g(t) cos(2πf t)dt − jg(t) sin(2πf t)dt
0 0

If g(t) is even, then g(t) = −g(t). Hence

Z 0 Z ∞
g(t) cos(2πf t)dt = g(t) cos(2πf t)dt
−∞ 0
Z 0 Z ∞
g(t) sin(2πf t)dt = − g(t) sin(2πf t)dt
−∞ 0
Z ∞
and G(f ) = 2 g(t) cos(2πf t)dt, which is purely real.
0
If, on the other hand, g(t) is odd, then g(t) = −g(−t). Hence
Z 0 Z ∞
g(t) sin(2πf t)dt = g(t) sin(2πf t)dt
−∞ 0

7
PROBLEM 2.6 CHAPTER 2.

Z 0 Z ∞
g(t) cos(2πf t)dt = − g(t) cos(2πf t)dt
−∞ 0
Z ∞
and G(f ) = −2j g(t) sin(2πf t)dt, which is purely imaginary.
0
(b) The Fourier transform of g(t) is defined by
Z ∞
G(f ) = g(t) exp(−j2πf t)dt
−∞

Differentiating both sides of this relation n times with respect to f :

Z ∞
dn G(f ) n
n
= (−j2π) tn g(t) exp(−j2πf t)dt (1)
df −∞

That is,
n
dn G(f )

n j
t g(t)
2π df n
(c) Putting f = 0 in Eq. (1), we get
Z ∞  n
n j
t g(t)dt = G(n) (0)
−∞ 2π
dn G(f )
where G(n) (f ) =
df n
∗
(d) Since g2 (t) G∗2 (−f ), it follows that
Z ∞
∗
g1 (t)g2 (t) G1 (λ)G∗2 (λ − f )dλ
−∞

From this result we deduce the Fourier transform

Z ∞
∗
F [g1 (t)g2 (t)] = g1 (t)g2∗ (t) exp(−j2πf t)dt
Z−∞
∞
= G1 (λ)G∗2 (λ − f )dλ (2)
−∞

Setting f = 0 in Eq. (2), we get the desired relation

Z ∞ Z ∞
∗
g1 (t)g2 (t)dt = G1 (λ)G∗2 (λ)dλ
−∞ −∞

8
PROBLEM 2.8 CHAPTER 2.

Problem 2.7
d
The Hilbert transform of g(t) is, by definition:
dt
Z ∞  
1 d 1
g(t) dτ
π -∞ dt t−τ

Interchange t with τ and vice versa:

1 ∞ d
Z  
1
g(t) dt
π -∞ d t τ −t
1 ∞ d
Z  
1
=− g(t) dt
π -∞ d t t−τ

d
which is immediately recognized as inverse Hilbert transform of g(t).
dt
d d
Conversely, we may say that ĝ(t) is the Hilbert transform of g(t), where ĝ(t) is
dt dt
the Hilbert transform of g(t).

Problem 2.8
We are given
∞
X
δT0 (t) = δ(t − mT0 )
m=−∞

a) Taking the Fourier transform of both sides:

∞
X
F [δT0 (t)] = F [δ(t − mT0 )]
m=−∞

Since F [δ(t)] = 1, then

F [δ(t − mT0 )] = exp (− j 2πf T0 )

= cos(2πmf T0 ) − j sin(2πmf T0 )
= cos(2πmf T0 )

9
PROBLEM 2.8 CHAPTER 2.

where the imaginary term is zero. Moreover,

∞
X 1 1
cos(2πmf T0 ) = δ(f ) for − ≤f ≤
m=−∞
2 2

Therefore,

F [δ1 (f )] for T0 = 1 and f0 = 1

Now introducing the scaling factor T0 into play, we have

1
F [δT0 (t)] = δf (f ) = f0 δf0 (f )
T0 0
Equivalently, we may write
1
δT0 (t) δf (f )
T0 0
b) from (2.34) in the text, we have
∞
X 1 1
δ(f ) = cos(2πmf ) − ≤f ≤
m=−∞
2 2
∞
X
= exp(j 2πmf )
m=−∞

Introducing the scaling factor f0 = 1/T0 , we may go on to write

∞
X ∞
X
δ(t − mT0 ) = f0 exp(j 2πnf0 t)
m=−∞ n=−∞

According to the duality theorem, we have

∞
X ∞
X
T0 exp(j 2πmf T0 ) = δ(f − nf0 ) (1)
m=−∞ n=−∞

b) Putting T0 = 1, (1) reduces to

∞
X ∞
X
exp(j 2πmf ) = δ(f − n)
m=−∞ n=−∞

10
PROBLEM 2.10 CHAPTER 2.

Problem 2.9
We are given
∞
X
x(t) = x(nT0 )δ(t − nT0 ).
n=−∞

Transforming this expression into the frequency domain, we get

X(f ) = F [x(t)]
∞
X 1
= x(nT0 ) exp(− j 2πnf0 ) where f0 = .
n=−∞
T0

Therefore, the Fourier transform of the output y(t) is

Y (f ) = H(f )X(f )
∞
X
= H(f ) x(nT0 ) exp(− j 2πnf0 ).
n=−∞

Thus, applying Parseval’s theorem in Table 2.1, average power is given by

Pav = |Y (f )|2
X∞
= |x(nT0 )|2 · |H(f )|2
n=−∞

(Correction: the m in the first summation in the text should be n.)

Problem 2.10
The system is said to be stable if the output signal is bounded for all bounded input signals;
we refer to this as the bounded input-bounded output (BIBO) stability criterion, which is
well suited for the analysis of linear time invariant systems. Let the input signal x(t) be
bounded, as shown by
|x(t)| ≤ M
where M is a positive real finite number. Substituting x(t) in
Z ∞
y(t) = x(τ )h(t − τ )dτ
−∞

11
PROBLEM 2.11 CHAPTER 2.

we get
Z ∞
|y(t)| ≤ M |h(τ )|dτ
−∞

It follows, therefore, that for a linear time-invariant system to be stable, the impulse re-
sponse h(t) must be absolutely integrable. That is, the necessary and sufficient condition
for BIBO stability is
Z ∞
|h(τ )|dτ ≤ ∞
−∞

Problem 2.11
Property 1
The signal g(t) and its Hilbert transform ĝ(t) have the same amplitude spectrum.

To prove this property, we observe that the Fourier transform of ĝ(t) is equal to −jsgn(f )
times the Fourier transform of g(t), and since the magnitude of −sgn(f ) is equal to one
for all f , the g(t) and ĝ(t) will both have the same amplitude spectrum. As a corollary to
this property, we may state that if a signal f (t) is band limited, then its Hilbert transform
ĝ(t) will also be band limited.

Property 2
If ĝ(t) is the Hilbert transform of g(t), then the Hilbert transform of ĝ(t) is −g(t).

To prove this property, we note that the process of Hilbert transformation is equivalent
to passing g(t) through a two-port device with a transfer function equal to −jsgn(f ). A
double Hilbert transformation is therefore equivalent to passing g(t) through a cascade of
two such devices. The overall transfer function of such a cascade is equal to

[− j sgn(f )]2 = −1 for all f

The resulting output is thus −g(t); that is, the Hilbert transform of ĝ(t) is equal to −g(t).
This results is subject to the requirement that G(0) = 0, where G(0) is the Fourier trans-
form of g(t) evaluated at f = 0.

Property 3
A signal g(t) and its Hilbert transform ĝ(t) are orthogonal.

12
PROBLEM 2.11 CHAPTER 2.

To prove this property, we use a special case of the multiplication theorem. In particu-
lar, for a signal g(t) multiplied by its Hilbert transform ĝ(t) we may write
Z ∞ Z ∞
g(t)ĝ(t) d t = G(f )Ĝ(−f ) d f (1)
−∞ −∞

which can be rewritten as:

Z ∞ Z ∞
g(t)ĝ(t) d t = j sgn(f )G(f )G(−f ) d f
−∞ −∞
Z ∞
= j sgn(f )G(f )G∗ (f ) d f (2)
−∞
Z ∞
= j sgn(f ) |G(f )|2 d f
−∞

where the second line, we have used the fact that for a real-valued signal G(−f ) = G∗ (f ).
The integrand in the right-hand side of (2) is an odd function of f , being the product of the
odd function sgn(f ) and the even function |G(f )|2 . Hence, the integral is zero, yielding
the final result
Z ∞
g(t)ĝ(t) d t = 0 (3)
−∞

This shows that the energy signal g(t) and its Hilbert transform ĝ(t) are orthogonal over
the entire interval (−∞, ∞)

13
PROBLEM 2.12 CHAPTER 2.

Problem 2.12
a)

sin(t)
g(t) =
t
The Hilbert transform of g(t) is

1 ∞ g(τ )
Z
ĝ(t) = dτ
π -∞ t − τ
Z ∞
1 sin(τ )
= dτ
π -∞ τ (t − τ )
1 ∞ 1
Z  
1
= + sin(τ ) d
πt -∞ τ t−τ
1 ∞ sin(τ ) 1 ∞ sin(τ )
Z Z
= dτ + dτ
πt -∞ τ πt -∞ t − τ
It is well known that:
Z ∞
sinc(y) d y = π
-∞

therefore
Z ∞
1 sin(τ ) 1
dτ =
πt -∞ τ t
Likewise, it should be noted that that
Z ∞ Z ∞
sin(τ ) sin(t − τ )
dτ = d tau
-∞ t − τ -∞ τ
Z ∞ Z ∞
cos(τ ) sin(τ )
= sin(t) d τ − cos(t) dτ
-∞ τ -∞ τ
= − π cos(t)

We thus obtain the Hilbert transform

1
ĝ(t) = (1 − cos(t))
t

14
PROBLEM 2.12 CHAPTER 2.

b)

g(t) = rect(t)

The Hilbert transform of g(t) is

Z ∞
1 1
ĝ(t) = P dτ
π -∞ t − τ

where P denotes the “principal value of”. When t < −1/2 the singularity in the integrand
is below the range of integration and the significant value of t − τ are negative. We then
have
1 1/2
ĝ(t) = −[ln(t − τ )]−1/2
π  
1 t − 1/2 1
= − ln , t<− (1)
π t + 1/2 2

When t > 1/2, the singularity is above the range of integration, and the significant values
of t − τ are positive. The corresponding value of ĝ(t) is
1 1/2
ĝ(t) = −[ln(t − τ )]−1/2
π  
1 t − 1/2 1
= − ln , t> (2)
π t + 1/2 2
1 1
For the case when − < t < , we write
2 2
"Z #
t− Z 1/2
1 dτ dτ
ĝ(t) = lim +
π →0 −1/2 t − τ t+ t − τ
1 n o
= lim [− ln (t − τ )]t−
−1/2 + [− ln (t − τ )]1/2
t+
π →0   
1  t − 1/2
= lim − ln( ) − ln −
π →0 t + 1/2 
1
1 −t 1 1
= − ln 2 , − <t< (2)
π 1 2 2
−t
2

15
PROBLEM 2.12 CHAPTER 2.

We may finally combine the results of Eqs (1) and (2) by expressing the Hilbert transform
ĝ(t), for all t as follows
1 t − 1/2
ĝ(t) = − ln
π t + 1/2
c)
g(t) = δ(t)
The Hilbert transform of the delta function δ(t) is
1 ∞ δ(τ )
Z
ĝ(t) = dτ
π -∞ t − τ
Using the shifting property of the delta function, we get the desired result
1
ĝ(t) =
πt
d)
1
g(t) =
1 + t2
The Hilbert transform of g(t) is
1 ∞
Z
1
ĝ(t) = 2
dτ
π -∞ (1 + τ ) (t − τ )
Z ∞ 
1 t+τ 1
= + dτ (3)
1 + t2 -∞ 1 + τ 2 t − τ
But (see appendix 6)
Z ∞
1
2
d τ =π
-∞ 1 + τ
Z ∞
τ
2
d =0
-∞ 1 + τ
Z ∞
1
d =0
-∞ t − τ

Therefore, Eq. (3) reduces to

t
ĝ(t) =
1 + t2

16
PROBLEM 2.13 CHAPTER 2.

Problem 2.13
Consider the Fourier-transform pair
1
exp(t)u(−t) (1)
1 − j2πf
where

 0, t>0
1
exp(t)u(−t) = , t=0
 2
exp(t), t < 0

Applying the duality property of the Fourier transform to Eq. (1):

1
exp(f )u(−f ) = G(f ) (2)
1 − j2πt
where

 exp(−f ), f > 0
1
G(f ) = , f =0
 2
0, f <0

From Eq. (2), it follows therefore that the inverse Fourier transform of G(f ) is the complex
time-function:
1
g(t) = (1)
1 − j2πt

The real and imaginary parts of g(t) are as follows:

1
Re[g(t)] =
1 + (2πt)2
2πt
Im[g(t)] =
1 + (2πt)2
From Table 2.3 of the textbook, we see that these two real-time functions are the Hilbert
transform of each other, except for the minor change of 2πt being replaced simply by t.

17
PROBLEM 2.14 CHAPTER 2.

Problem 2.14
To solve the problem in the time domain, we have to be careful: rigorously, speaking, the
Hilbert transform of g(t) is defined by the principal value of the integral in (2.54) in the
text, as shown by:
Z τ −∆t Z ∞ 
1 g(τ ) g(τ )
ĝ(t) = lim dτ + dτ
∆t→0 π −∞ t−τ τ +∆t t − τ

However, an easier way of solving the problem is to work in the frequency domain in two
steps:
Step 1)
Differentiation of ĝ(t), the Hilbert transform of the function g(t), requires multiply-
ing the Fourier transform of the differential d ĝ(t)/ d t, as shown by
 
d
F ĝ(t) = j 2πf Ĝ(f ) (1)
dt
Where ĝ(t) Ĝ(f )
Performing the Hilbert transformation in the frequency domain, we write
Ĝ(f ) = − j sgn(f )G(f ) (2)
Where g(t) G(f ), and sgn(f ) is the signum function

In an inverse manner to (2), we may write

G(f ) = j sgn(f )Ĝ(f ) (3)
Multiplying both sides of Equation (3) by j 2πf , we get
 
j 2πf G(F ) = j 2πf j sgn(f )Ĝ(f ) (4)

Translating the terms in Equation (4) into the time domain, we proceed with the
second step.

Step 2)
It may be noted that the left side of Equation (4) is equivalent to
 
d
j 2πf G(f ) = F g(t) (5)
dt

18
PROBLEM 2.15 CHAPTER 2.

Likewise, on the right of Equation (4) we have

 
  d
j 2πf j sgn(f )Ĝ(f ) = F ĝ(t) (6)
dt

Finally, in light of Equations (5), and (6), we immediately make the statement:
d d
The differential ĝ(t) is the Hilbert transform of g(t).
dt dt
Hence the problem is solved

Problem 2.15
Z ∞
a) The Hilbert transform of the integral g(t) d t is defined by
-∞
Z ∞ Z ∞ 
1 1
g(t) d t d τ
π -∞ t−τ -∞
Z ∞ Z ∞
which, in general, is not equal to the integral ĝ(t) d t, because the integral g(t) d t
-∞ -∞
is the total area under the function g(t). In other words, the integral under the square
brackets is independent of time t.

b) However, if we are permitted to write

1 ∞ ∞ g(t)
Z Z Z ∞ Z ∞ 
1 g(t)
dtdτ = dτ dt
π -∞ -∞ t − τ -∞ π -∞ t − τ
Z ∞
= ĝ(t) d t
-∞
Z ∞ Z ∞
then the integral ĝ(t) d t is equal to the Hilbert transform of the integral g(t) d t.
-∞ -∞
Under this approach, the total area under g(t) is the same as that under ĝ(t).

19
PROBLEM 2.16 CHAPTER 2.

Problem 2.16
sin(πt)
(a) g(t) = sinc(t) =
πt
We note that
1 − cos(πt)
ĝ(t) =
πt
Therefore

g+ (t) = g(t) + jĝ(t)

sin(πt) 1 − cos(πt)
= +j
πt πt
j
= [1 − cos(πt) − j sin(πt)]
πt
j
= [1 − exp(jπt)]
πt
(b)

g(t) = [1 + k cos(2πfm t)] cos(2πfc t)]

k k
= cos(2πfc t) + cos[2π(fc + fm )t] + cos[2π(fc − fm )t]
2 2
Since the Hilbert transform of cos(2πf t) is equal to sin(2πf t), it follows that

k k
ĝ(t) = sin(2πfc t) + sin[2π(fc + fm )t] + sin[2π(fc − fm )t]
2 2
where it is assumed that fc > fm . Therefore,
k k
g+ (t) = exp(j2πfc t) + exp[j2π(fc + fm )t] + exp[j2π(fc − fm )t]
2 2
k k
= [1 + exp(j2πfm t) + exp(−j2πfm t)] exp(j2πfc t)
2 2
= [1 + k cos(2πfm t)] exp(j2πfc t)

20