5

Fourier Transforms

5.1 INTRODUCTION

In Chapter 4 we discussed about the use of Fourier series to analyse periodic signals. Using
exponential formn of Fourier series, any continuous-time periodic signal x() can
represented as a linear combination of complex exponentials and the Fourier coefic1eaß
(spectrum) are discrete. Fourier series can deal only with the periodic signals. Ths S
major drawback of Fourier series. However, all the naturally produced signals whica
processing will be in the form of nonperiodic or aperiodic signals. Therefore, the applicabiliy
of the Fourier series is limited.
In this chapter, we discuss about Fourier transform which is mostly used to analys
aperiodic signals and can be used to analyse periodic signals also. So it overcomes the
limitation of Fourier series.
Fourier transform is a transformation technique which transforms signals tromth
continuous-time domain to the corresponding frequency andwhich
vice versa,
domain
applies for both periodic as well as aperiodic signals. Fourier transform can be developedb
and
findingThe
the Fourier series of a periodic
Fourier transform derived infunction and
this chapter then tendingtheT tocontinuous-time
is called
infiniy Founie
Founi
transform (CTFT) to distinguish it from its
transform (DTFT), discrete Fourier several yariants like discrete-timesakeo
convenience we shall call it here only astransform (DFT) etc. However, for the
Fourier
The Fourier transform is an extremely usefultransform.
mathematical tool and 1S extensively
in the analysis of linear time-invariant (LTI) systems, cryptography, signal analysis,sig
spect
processing, astronomy, etc. Several applications spread
communication employ Fourier transform. ranging from RADAR to

296
 Fourier Transforms 297

FOURIERTRANSFORM
TH REPRESENTATION OF
FUNCTIONS NON-PERIODIC
Chapter4, we
had derived the Fourier series of a periodic function.
Now, we shall try to
amVe
n atthe Fourier transform through Fourier series. For this purpose, consider an arbitrary
reriodicfunction x() with a fundamental periodT as shown in Figure 5.1(a). We know that
spectrum of such a periodic signal will be a discrete one with spectral lines at frequencies
the with the adjacent spectral lines separated
ao o ..,
by afrequency interval of fo =1/T.
the amplitude spectrum of signal x() is given in Figure 5.1(b).
Aplotof
4f)

-2 2 T
-T
(a)

Afot

2/t 3/T
-1/t 1/t
-3/t -2/t
(b)

Aft
fo

3/T
2/T
1/T
-3/t -2/t -1/t
(c)

Af,t
allh
3/ 4/t f
2/t
(0 1/T
-3/t -2/t -1/7
(d) Spectrum when
(d)
Spectrum when T is doubled,
Figure 5.1 (a) Signal, (b) Its amplitude spectrum,
(c)
is increased
by four
times.
T
298 Signals and Systems
fundamental period T of the periodic signal
Suppose we now increase the appear as
say double the period. Then the amplitude spectrum will will be as shown in
When the period is increased by four times, the spectrum the generalin shown
Observe that the amplitude of the spectrum is decreased but
unchanged.
As can be seen from Figure 5.1[(c) and (d)], an increase in
shaFipegurereTI5.ÄNl
fundamental periobedTressi
in a spectrum in which the spectral lines become closer and closer. This is to
since fo = 1/T and so, as Tgoes on increasing, fo goes on decreasing and the
adjacent spectral lines becomes smaller and smaller. spaci n g e xpet
bet e
w Aec
On the other hand, as the pulses in x() are separated by Tseconds. as T
interval between adjacent pulses becomes longer and longer. Finally, as
adjacent pulses become separated by an infinite amount of time,
T incre
tends toase
i.e. x(t) infi ty
s, the
non-periodic function. At the same time, separation between the spectral lines a
becomes
a becomes,
infinitesimally small, or, in other words, a non-periodic signal will be
spectrum. The envelope of the spectrum depends only upon the pulse shape having
but notcontunn
inuox
period of repetition.

5.2.1 Derivation of the Fourier Transform of a

the Fourier Series of a Periodic Signal Non-periodic Signal fron
Let x() be a non-periodic function and, x(t) be
is given by periodic with period T, and let their relatin
x(t) = T’o
Li x,(t)
The Fourier series of a
periodic signal x() is:
(0)= nco
Cçemag
where 1 7/2
C,= T x() eMo dt and a, = 2
-TI2
(The term Ch represents the
amplitude of the component of
TI2 frequency n@:l
TC, = p()e na dt
Let na = 0 at T ’ --TI2
Founi
o, As T ’ oo,
spectrumn becomes continuous. We have on =
(2t/T) ’0 and thex()discrete
Thus, as T ’ o, Further, the summation
TI2 becomes integral and
TC, =T’L
-TI2
dt=x(t) e-ja dt= X(@)

[X(0) represents the frequency spectrum of x() and isdt called the Spectral density function/
 Fourier Transforms 299
the Fourier transform or the Fourier integral of x(t).
X(o) is called
Hence,

iM
M
2=--0

X(@) TC, X(0)

T

X(0) T=7
nb = ,
27

Lt 1
Lt x,(t) = T’o
x(t) =T-’o X(0) emag!
2x n=-o0

represented by d.
As T ’ oo, an = (2t/T) becomes infinitesimally small and may be
integration.
Also the summation becomes
1
x(t)= |X(0)e da
27

Hence, x(i) is called the inverse Fourier transform of X().

The equations
X(0) = |x()e-jon dt

1
and x(t)= |X(0)e'o do
27

pair and can be denoted as:

a0) and x(1) are known as Fourier transform
X(0) = F[x()]
and x() = F[X(0)]
The other transform pair is:
notation that can be used to represent the Fourier
x(t) 4 X(0)
53 MAGNITUDE AND PHASE REPRESENTATION OF FOURIER

The TRANSFORM transform is the tool used to

analyse
magnitude and
the transformed signal.
phase representation of the Fourier
Therefore, X(0) can be written as:
In X( o) is a Complex valued function of
o.
general,
X(@) = Xp (@) + jX, (@)
where Xplo) is the real part of X(0) and X,(w) is the
imaginary part of X(o).
 Signals and Systems
300
X(0) is given by
The magnitude of
|x(o))=\X;(o)'+ X,(0)'
X(o) is given by
and the phase of tan-! (a)
X(0)=
Xp (@)
The plot of |X(0) versus ois known as amplitude spectrum, and the plot of \X{
versus @
is known as phase spectrum. The amplitude spectrum and phase spectrumtogethe

iscalled frequency spectrum.

EXISTENCE OF FOURIER TRANSFORMS

The Fourier transform does not exist for all aperiodic functions. The conditions for a funa
xt) to have Fourier transform, called Dirichlet's conditions, are:
1. x() is absolutely integrable over the interval -o to o, that is

2. x() has a finite number of discontinuities in every finite time interval. Furher, catl
of these discontinuities must be finite.
3. x() has a finite number of maxima and minima in every
finite time interval.
Almost all the signals that we come across in physical
conditions except possibly the absolute integrability problenms satisfy all tne au
condition.
Dirichlet's condition is asufficient condition but not necessary condition. This means
Fourier transform will definitely exist for functions which satisfy these conditions. Onth
other hand, in some cases, Fourier transform can be found with the use of impulses evenfo
functions like step function,
condition. sinusoidal function, etc. which do not satisfy the convergen
5.5 FOURIER TRANSFORMS OF
5.5.Y Impulse Function S(t) STANDARD SIGNALS
Given x(t)= S(0),
for t =0
|0 for t#0
Then
X(@) = x() ejar dt
=S() e ja dt
lts 1
 Fourier Transforms 301

F[S()] = 1 or S) FT

Hence,the Fourier transform of a unit impulse function is unity.

X(0)=1 for all o
X
(@) =0 for all o
heimpulse function with its magnitude and phase spectra are shown in Figure 5.2.
|X(o) |
4 x(1)
4X(O)

0 0

(a) (b) (c)

Figure 5.2 (a) Impulse function, (b) Its magnitude spectrum, (c) Its phase spectrum.

dt =e0 ie. S(t - h) eiao

Similarly, F[Bt -)]=|t-to) e

552 Single-sided Real Exponential Function eult)

for t>0
Given x(t) = e u(t), for t <0

Then X(@) = x(t) ejor dt = e u()e dt

e-(a+jo):
-Je"e ar dt =ea+jox,dt =|
-(a+j@) <a+ jo)
0 0

0-1 1
<a+ jw) a+ jO
-at
u() FT
or a+ jo
Fle
u(t)]= a+ jw

Now, X(0) =
a-jo
(4-jo)
a+ jo (a+ jo) - tan-!O
a- jo
a +o
A(0)=-tan-!O for all w
1
|X(@)= Va'ta
 Signals and Systemns
302 with its
exponential function magnitude and
Figure 5.3shows the
4x() =eu)
single-sided
4|X(0)|
4\X(0)
ohase speAR
1/a

+/2

(a)
(b) (c)
Figure S8 (a) One-sided exponential function, (b) Its amplitude spectrum, and (c) Its phase spectrum

eal|
55.3 Double-sided Real Exponential Function
Given x) = el
for ts0
e-ai) for t>0
=e a(-)
u-t) +e ult)
=e"u-t) + e "u(t)

X(@) = x()eia dt
0
dt +ee jan dt =ea-ja dt+ |e(a+tjo dt
e-(a-ja)r e (a+jo): o0
dt + ea+jo): dt = +
a- jo) Jo -<a+ j®)b
+ 1 1 2a
-(a- jo) -(a+ jo) a-
jo a+ jo
F(e altl)s2a
or eal1| 2a
a +o
2a
for all w
and a'+o
A X(0) =0 for all o
two-sided
Figure 5.4. exponential function
and its are
shown

amplitude and phase spectra

 Fourier Transforms 303
4x()= ede|
t|X(%)|
2la 4LX(0)

0
0
(a) (b) (c)
Figure5.4 (a) Two-sided exponential function, (b) Its magnitude spectrum, and (c) Its phase spectrum.

eloot
s54 Complex Exponential Function
- Eod fhe Fourier transform of complex exponential function ejao, consider finding the
inverse Fourier transform of (O o). Let
X(0) = (0 - a)

x() =FX(0)]= F[B(O- a)]= 277 X(0)ea" do

-eja
27 27

or F(28 o-o,)]=ea
F80- 0)]= 27
F[e]=2rö(0- a,)
lejar FT 226(0- 0))|

555 Constant Amplitude (1) Fourier transform directly, So

its
Sncethe x(1) = 1 is not absolutely integrable, we cannot find Fourier transform of S(0).
inverse
determined through
Fourier
We know thattransform of x(t) = 1 is

x)=
(X(@)e do
27
|1 for W=0
Consider X(o) = o(0) for w#0

Fö0))
x() = FX(0)]=
27 21
21
 a
consider
to
duration
function.
us T/2
Letsmall constant
5.5. (b)
the
=1
F'2r80)]
Figure
extend the
dt
-T/2
of ss
2AS(0)
FÊA]= )ejn 2z8(x) 2rö(w) in we
section 2
|0
elsewhere
2zÓ(w)
F[l]= 2zS(o)
=
F[1] shown
If Small
1=
x(t) t. Therefore.
FU]= = = duration for
or dt dt is (b)
[e-ix |ej function x(t)=
Lt function,
27 = 1=
x(t)
1 =
F'S(0)) X(0) of function.
-00 say,
constant Constant
Another
alternate
method function,
original
Systems a (a)
of constant a)
5.5
and
method
Alternate the
waveform
Figure
back
Signals the
get
Similarly. We
have Theof
will
section
where
304 Let Let we
or
 Fourier Transforms 305
detinition,the
Fourier transform of x() is:
By
X(0)=}F[x()]= F Lt rect rect
t/2 e jn]/2
= Lt | ()e dt = Lt
-t/2 T’jw J-t/2
eJOTI2)j(T/2) 1
= Lt = Lt 2sin[o( t/2)|
-jo T-’o

= Lt
sin[ o(r/2)]| Lt Sa t/2
o t/2) :27 Lt

T/2
Il'ing the sampling property of the delta function Sa
- we get

X(0) = F Lt rect

556 Signum Function sgn(t)

The signum function is denoted by sgn() and is defined by
for t >0
sgnt0) = for t<0
Fourier transform.
Ins Tunction is not absolutely integrable. So we cannot directly find its obtain the
ereiore, let us consider the function e a sgn(t) and substitute the limit a ’ 0 to
bove sgn (t).
Given x(1)= sgn(t)= a-’0 Lt [eut) -eu(-t)]
Lt e a'l sgn(t) = a-’0

dt
X(0) =F[sgn()] =

= Lt
a’0
jo u(i) dt - (e"e jaru-t) dt
p-(a+jo) -jo" dt
Lt
= Lt
a’0
[etaja) dt - [ee-j" dt a’00
0

e-0
e la-ja)r
e (a+jo L
a-’0 -(a+ jo) <a-j ) |
= Lt
a’0
|L-(a+ j0) Jo -(a- jo)
2
Lt 1 1
-jw jo
a-’0| a+ j) a- jo jo
 Signals and Systems
306
FIsgn()]=jo

FT
2
sgn()
or
jo
for w<0 and
for w> 0
and X(@) =,2 2
Figure5.
phase spectra.
shows the signum function and its magnitude and
Figure 5.6 4|X(0) |
sgn () +LX(0)
u2

-n2

(a) (b) (c)

Figure 5.6 (a) Signum function, (b) Its amplitude spectrum, (c) Its phase spectrum.

5.5.7 Unit Step Function u)

The unit step function is defined by
1 for t 0 Consider
for t <0 defined
Since the unit step function is not absolutely integrable, we cannot directly find its ouriet
transform. So express the unit step function in terms of
signum function as:
1,1
ut) =sgn(t)
2
Given 1
x()=u() =Il+sgn())
X(0) =

We know that -;{FUJ+ F(sgn()}

F(l]=2rÕ( 0) and F[sgn(t)] = 2
jo

jo jo
 Fourier Transforms 307
F(u()] = rö() +
jo
u(t) FT 1
’ nSlw) +
of
jo
= 00 at @ and is equal to 0 at ) =
= 0
-oo and o= 0,
5.7shows the unit step function and its
Figure spectrum.
0.5 sgn(t) x(t) = u(t) 4X(0)
1
0.5 +0.5 sgn(t)
0.5

-0.5

(a) (b) (c)

Figure 5.7 (a) 0.5 sgn (), (b) u(), (c) Spectrum of u().

or rect
j58 Aectangular Pulse (Gate pulse)
function and is
Onsider a rectangular pulse as shown in Figure 5.8. This is called a unit gate
defined as.
4 II(¢/)

1.0

-tI2 0
function.
Figure 5.8 Unit gate

[1 for s tI2
otherwise

t/2
J(1)e-jar dt =
-j@ ]2
-t/2
308 Signals and Systems
gJo(th2)-pj0(tl2)
e jo(ti2)-jo(tl2)

-jo jo
ejo(r|2) -e jv(r/2) |sin o(t/2)
2j t/2)
ot2)
=tsinc
2

= TSa
2

sinc that is
2

rect T sinc

Figure 5.9 shows the spectra of the gate function.

4|X(0) |

210 2T

Mainlobe 4n
T
(a)
4 LX(0)

2r 4n 67 81
8T
4
210 0

(b)
Figure 5.9 (a)
Amplitude spectrum, and (b)) Phase spectrum of II(#).
The amplitude
At w= 0,spectrum is
sinc (0t2)obtained as follows:
i.e. at =1.
Therefore, |X(o)l at o = 0 is equal to t. At (wr2)=
2nt
-,n=1,2, ..., sinc
 Fourier Transforms 309
spectrum is:
phase
p
1he
X(@) =0 if sinc(orI2) >0
=tn if
sinc(ot/2)
<0
[eamplituderesponse hetween the first two zero crossings is known as main lobe and the
for
rtonsof the response O<-2Tt) and o> (2rd) are known as
side lobes. From the
smplitudespectrum, we can find that majority of the energy of the
MALnlobe. The first zero crossing occurs at w= (2 T/t) or at f= (1/0)signal
Hz. isAs contained
the width inof the
Rtangularpulse is made longer, the main lobe becomes narrower. The phase spectrum 1s
dd function of 0. If the amplitude spectrum is positive, then phase is zero, and if the
nnlitude spectrumis negative, then the phase is -T or 1.

559 Triangular Pulse

?)
Consider the triangular pulse as shown in Figure 5.10. It is defined as:

-T/2
function.
Figure 5.10 Triangular

1 for
tl2
for r0cr<s
elsewhere

2|r| for le|<;

otherwise

X(@) = F[x())=} t/2

 Signals and Systems
310

---)a t/2 2t ,-jx

2t e dt
0 0
t/2
+e Jo]dt
= |e+e j dt - 0
0

2t cos M dt
=2cos cx dt - |
sin @t 7/2 t/2
sin ax 4 r Sin 0t
=2 - -J 0
dt

q/2 sin ot qt/2

Isin ot
+
Jo
4 4
2sinar2 2 a'z
COS
WT
2
4 2sin2 r]|
|1- cos
2 4
8 \2 sin? (wtI4)
Sinc?@r
or 4 (ori4)? 4

4
or
FT
sinc?
2 4
Figure 5.11 shows the
amplitude spectrum of a triangular
pulse.
4|X(0)|

-8rlt -4nIt
Figure 5.11 4tlt 8tlt
Amplitude spectrum of a
triangular puise.
 Fourier Transforms 311
s&10 ossine Wave COS ht

Given
ontO
2aTARSO
x(t) =cOs @ht

xIo)= F[x)) =F(cos @y]=F(ete

)+F(e )=2rö(0 0) +2nö(0 +a,)]
Flcos ah] = n[S(0 + a) + 8(0- a)] or cos a,t T[8la + on) + S(o- a,)
Figure 5.12shows the cosine wave and its amplitude and phase spectra.
4 x(0) 4|X(0) | 4LX(0)

-ab0

(a) (b) (c)

Figure 5.12 (a) Cosine wave, (b) Its magnitude spectrum, (c) Its phase spectrum.
S5.11 Sine Wave sin abt
Given x(t) = sin !

Then F[x()]= X(0) = F[sin ayt]=F 2j

2r8(0 + )l
2j

jr[B(o-a-0+ an)
Ö(0+ a)] or sin @ht
jn[8(0 - ab) +
and phase spectra.
Fguxe 5.13 amplitude
shows the sine wave signal and its |X(@)| 4LX(@)

4x() = sin a

0
-x2
(c)
spectrum.
(b) phase
(a) spectrum, (c) Its
amplitude
(b)Its
Figure Sinusoidal signal,
 FOURIER
312
Signals and Systems
CONTINUOUS TIME TRANSFORM
PROPERTIES OF properties. These properties
5,6 ofimportant are useiul to
transform has a number well as for deducing general frequency
The Fourier
Fourier transform pairs as effect of various time domain operations domän
deriving
These also help to find the properties are discussed as follows:
relationships. important
domain. Some of the
frequency
56.1 Linearity Property
states that the Fourier transform of a weighted sum of two signals is
The linearity property of their individual Fourier
transforms.
sum
equal to the weighted
() ’ X{(@) and x, () FT X,(0)
ie. If

Then ax (t) + bx,(t)

FT
aX{(@) + bX,(w)
where aand b are constants.

Proof: By definition,

F[ax, () + bx, ()]= [ax, (t) +bx,(t)] e dt

=Ja,(t) e dt + bx, (t)e-jr dt91 i

=a x()e i dt +b x0)e i dt
=aX{ (@) +bX,(0)
ax (1) + bx, (t) FT

¢aX{ (0) + bX, (0)

56.2 Time Shifting
The time shifting Property
modified by a linearproperty states
phase shift ofthat if a signal x() is spectrum.
If slope -oto, i.e. shifted by to sec, the

Then x{t) ¢F’ X(0)

Proof: By definition,
X(0)
F|x -
t,)]=x(t- t)e io dt
 t - to
Fourier Transforms 313
dt = dp
=p

F[x(!- ,)]=x(p)e jop+0) dp

| x(p)e ioP dp
-jolo X(0)
x(t -to) -jjo
oMo X(o)
FT
Similarly, X(@)
This property has a very important implication. That is
leiot
X(0)|=|X()|
and ejoo X(@) =e jalo +|X(o) =-@to t|X(@) pniso& srmiT
s shows that shifting a function by to results in multiplying its Fourier transform by e,
hs, there is no change in magnitude spectrum but the phase spectrum is linearly shifted.

Frequency Shifting Property (Multiplication ( by an Exponential)

x() by eo0
Fresulequency shifing property states that the multiplication of atime domain signal
ts in the frequency spectrum shifted by @, i.e.
x()4 ’X(0)
Then
Progf: By definition,

elx) X(@- )

Sirlad'y. e ja x(t)
X(0+ b)
 Signals and Systems
314
Property
5.6.4 Time Reversal
property states that
The time reversal
x)4X(0)
If
x(-) X(-0)
Then
Proof: By definition,
F[x(-))= [x(-)e-io dt
[x(-)1, we have
Replacing rby in the RHS of the above expression for F

F[x(-)] = x(1) e dt =x()e-i-or dt = X(-0)

x(-)4X(-o)
5.6.5 Time Scaling Property
Let x(at) is a compressed version of x(t) when a>1or expanded version of x() when a<i.
If
x()¬ X(0)
Then x(at) FT 1

Proof: By definition,
)
F[x(a)] = x(at) e-ior dt
Let
at = p

a
and dt = dp
a

F(x(at)]=x(p)e ja(pla) 4p
a

CASE 1
tx(p) ei(ala) Pdp
When a > 0,

F(x(a)) =- x(p) e-Kai)rdp =x

 <0, Fourier Transforms 315
2
When a
CaSE
1
F[at]= - a ( x(p) e'(aa) Pdp =í
-a x(p) e--(ala)lP dn

--)1
X
lal
FT 1
x(at) 4

566 Differentiation in Time Domain Property

s Aiffenentiation in time domain property states that
in is equivalent to the multiplication of its the differentiation of a function in time
Fourier transform by a factor jo. ie.
x(t) X(0)
d
Then
dt
x(i)¢ joX(@)
Prof: By definition,
1
x()=
27 |X(@) e do
Differentiating both sides w.r.t. t, we have
d 1 d
dt
x(t) =
27 dt JX(0) elu do
1 1
27 |X(@)[el"]
dt
dw= 27
| X(0) joe do

|X(0)ej" dw
277

= joF[X(@)]

jo X(0)

dx(),
dt
F FT , joX(@)
316 Signals and Systems
FT (jo)" X(0)
d"x),
d1"
In general,
Frequency Domain Property
in
5.6.7 Diferentiation that the multiplication of a
domain property states
in frequency
The differentiation differentiation of its Fourier transform in frequency domain, signal xh
ie.
by t is equivalent to
x(1)4 X(0)
If
d
Then
tx() jdo X(@)
Proof: By definition,

K(0) = x) ejar dt
Differentiating both sides w.r.t. W, we have

d d
do [X(@)]= do dw -(e an) dt=x()(-j)e i dt

dt =-jF[tx()]

F[x()]= jX(0)
do
FT d
tx(t)
jX
do (0)
5.6.8 Time
The time
Integration Property
equivalent integration property states that the
to the division domainis
If
of its Fourier integration of a
transform by jo, i.e. function x() in time
x() ’X(0)
Then
xí) dre
Proof: If X(0) = 0, this
case of the
Txjo), jo if X(0) =0
differentiation property can be easily the
in
By definition, proved by using integration by parts
property. as

2r X
(0) ejo dw
 Fourier Transforms 317
dummy yariable T, we have
t by a
1
x(r)= |X(0) ejor do
2

sides over -oo to t, we have

Megratingboth

|X(0) eior dt \do

rchangingthe order of integration, we have

27

eJr
-
dw
27 jo J-oo

1 X(o) eler
do = F-! X(@)|
27 jo jo

F X(0)
jo

x() dr "xX(0)
jo

transform ofx(7) dt
)FO then x(t) is not an energy function, and the Fourier
dudes an impulse function, that is
-X(0)+TX(0) S(0)
jo

e Convol ution Property or Theorem

convolution of two signals in
time

heymainconvol
is ution property
the
Or theorem
states that domain. This is called
spectrain frequency
time equivalent the
to multiplication of their
convolution theorem. x()"X,(o)
X(w) and
Then X, (0) X, (w)
X,() 4
318 Signals and Systems
and Xz(t) is given
that the convolution of two signals xj() by
Proof: We know
()*x)= x(t) x,(t-)dr

F[()* (O]= ||4(t) x,(t-t)dt

e J dt

Interchanging the order of integration, we have

dt

Substituting t - T =p in the second integration, we have

t=p+ T and dt = dp

-o

-jua| |(p) e-jap dp eja dr

=J () X,(o) e jor dt

()eiar
X (@) X,(0)

)*x,()X,(o) X, (0)
5.6.10 Multiplication
The Property or Theorem
domainmultiplication
is property
equivalent to the
or theorem
states that the intin:
called the frequency convolution of their multiplication of two functions
domain, This
If convolution theorem. spectra in the frequency
Then
)X(0) and X, (@)
Proof: We know that x0) xy(0)
FX(0) 2 * X,(@)
F[x()]= X(0) = x() eias dt
 Fourier Transforms 319
F'X(o)]= x)= 1
27 |X(0)ea" do
F[M() x,(0)]= () x,() ejon dt

Interchanging the order of integration, we get

dt d

2
x,()ei@-2) dtdl

1
27 x,(4) X,0-4) da
-21XX{ (@)* X, (0)

1
;(1) x, (0) 20 X,(0)* X,(o)
2rx (1) x, (") ¢ X (0) *X, ()
*; ) x, () X)* X,()

5611 Duality (Symmetry) Property

In spectrum analysis, the duality between the time and the frequency is exhibited. The duality

(symmetry) property states that

x(t) 4 X(0)
Then X() 4 27x(-0)
Proof: By definition,
x() =27 X(@) e da
 or
320
functions,
evenFor i.e. Interchanging
Similarly, Proof Then Modulation
If Property
translated bye 5.6.12
Thif
Systems
Signalsand
modulation
property
states
0I]=F
F[x()cos
up
t
and and
down o,
x(!) x) 2 x) we
in
sin 2
cos frequency 2Tx(-0) have
w,)- xt)
w,!X(O-
w,l+ cos
that, 2rx(-) 27x(1)
27x(-0)
X()4F[X()]=
x(-0)27
@! x(t) by x(-0)
x(0) = =
X|aO 2
, a
2j
7X(o-
x()| =
2 o,signal FT = =|
X()e
wl+ (0)’X
i.e.
X(0)ej X(0)
x() 27x(0)
w) 2
w,) is dt e
+ + multiplied =
(w+ X
Xo+ 2
X(0+ F[X()] dw dw
Q) a, Q,))
cOS@f,its

spectrum ge