Fourier Transform

Fourier Transform
• We have seen that periodic signals can be
represented with the Fourier series
• Can aperiodic signals be analyzed in terms of
frequency components?
• Yes, and the Fourier transform provides the
tool for this analysis
• The major difference w.r.t. the line spectra of
periodic signals is that the spectra of
aperiodic signals are defined for all real
values of the frequency variable  not just
for a discrete set of values
 Frequency Content of the
Rectangular Pulse
x(t )

xT (t )

x(t )  lim xT (t )
T 
 Frequency Content of the
Rectangular Pulse – Cont’d

• Since xT (t ) is periodic with period T, we

can write

xT (t )  
k 
ck e jk 0 t
, t

where
T /2
1

 jk o t
ck  x (t )e dt , k  0, 1, 2,
T T / 2
 Frequency Content of the
Rectangular Pulse – Cont’d
• What happens to the frequency components
of xT (t ) as T   ?
• For k  0
1
c0 
T
• For k  0
2  k 0  1  k 0 
ck  sin   sin   , k  1, 2,
k 0T  2  k  2 
 0  2 / T
 Frequency Content of the
Rectangular Pulse – Cont’d

plots of T | ck |
vs.   k 0
for T  2,5,10
 Frequency Content of the
Rectangular Pulse – Cont’d
• It can be easily shown that
 
lim Tck  sinc  ,   
T 
 2 
where
sin( )
sinc( ) 

 Fourier Transform of the
Rectangular Pulse
• The Fourier transform of the rectangular
pulse x(t) is defined to be the limit of Tck
as T   , i.e.,
 
X ( )  lim Tck  sinc  ,   
T   2 
| X ( ) | arg( X ( ))
 Fourier Transform of the
Rectangular Pulse – Cont’d
• The Fourier transform X ( ) of the
rectangular pulse x(t) can be expressed in
terms of x(t) as follows:

1
ck   x(t )e  jk o t
dt , k  0, 1, 2,
T 
x(t )  0 for t  T / 2 and t  T / 2
whence

Tck   x(t )e  jkot dt , k  0, 1, 2,
 Fourier Transform of the
Rectangular Pulse – Cont’d

• Now, by definition X ( )  lim Tck and,

T 
since k 0   as T  


 x(t )e
 j t
X ( )  dt ,   
• The inverse Fourier transform of X ( ) is

1
 X ( )e
j t
x(t )  d , t  
2
 The Fourier Transform in the
General Case
• Given a signal x(t), its Fourier transform
X ( )is defined as


 x(t )e
 j t
X ( )  dt ,   
• A signal x(t) is said to have a Fourier
transform in the ordinary sense if the above
integral converges
 The Fourier Transform in the
General Case – Cont’d
• The integral does converge if
1. the signal x(t) is “well-behaved”
2. and x(t) is absolutely integrable, namely,


 | x(t ) | dt  
• Note: well behaved means that the signal
has a finite number of discontinuities,
maxima, and minima within any finite time
interval
Example: The DC or Constant Signal

• Consider the signal x(t )  1, t  

• Clearly x(t) does not satisfy the first
requirement since
 

 | x(t ) | dt   dt 

• Therefore, the constant signal does not have
a Fourier transform in the ordinary sense
• Later on, we’ll see that it has however a
Fourier transform in a generalized sense
 Example: The Exponential Signal

• Consider the signal x(T )  e  bt u (t ), b  

• Its Fourier transform is given by


 e
 bt  j t
X ( )  u (t )e dt

 t 
1
 e  ( b  j ) t
dt    e  ( b  j ) t

0
b  j  
t 0
 Example: The Exponential Signal –
Cont’d
• If b  0 , X ( ) does not exist
• If b  0 , x(t )  u (t ) and X ( ) does not
exist either in the ordinary sense
• If b  0 , it is
1
X ( ) 
b  j
amplitude spectrum phase spectrum
1  
| X ( ) | arg( X ( ))   arctan  
b2   2 b
 Example: Amplitude and Phase
Spectra of the Exponential Signal
10 t
x(t )  e u (t )
 Rectangular Form of the Fourier
Transform
• Consider


 x(t )e
 j t
X ( )  dt ,   
• Since X ( ) in general is a complex
function, by using Euler’s formula

  
X ( )   x(t ) cos( t ) dt  j    x(t )sin( t ) dt 

             
R ( ) I ( )
X ( )  R ( )  jI ( )
Polar Form of the Fourier Transform

• X ( )  R ( )  jI ( ) can be expressed in
a polar form as
X ( ) | X ( ) | exp( j arg( X ( )))

where
| X ( ) | R ( )  I ( )
2 2

 I ( ) 
arg( X ( ))  arctan  
 R ( ) 
 Fourier Transform of
Real-Valued Signals

• If x(t) is real-valued, it is
 Hermitian
X ( )  X ( ) symmetry
• Moreover

X ( ) | X ( ) | exp(  j arg( X ( )))
whence
| X ( ) || X ( ) | and
arg( X (  ))   arg( X ( ))
 Fourier Transforms of
Signals with Even or Odd Symmetry

• Even signal: x(t )  x( t )


X ( )  2  x(t ) cos( t ) dt
0

• Odd signal: x(t )   x (t )


X ( )   j 2  x(t )sin( t ) dt
0
 Example: Fourier Transform of the
Rectangular Pulse
• Consider the even signal

• It is  / 2
2 2   
X ( )  2  (1) cos( t ) dt   sin( t )  t 0  sin 
t  / 2

0
   2 
  
  sinc  
 2 
Example: Fourier Transform of the
Rectangular Pulse – Cont’d
  
X ( )   sinc  
 2 
 Example: Fourier Transform of the
Rectangular Pulse – Cont’d

amplitude
spectrum

phase
spectrum
 Bandlimited Signals

• A signal x(t) is said to be bandlimited if its

Fourier transform X ( ) is zero for all   B
where B is some positive number, called
the bandwidth of the signal
• It turns out that any bandlimited signal must
have an infinite duration in time, i.e.,
bandlimited signals cannot be time limited
 Bandlimited Signals – Cont’d

• If a signal x(t) is not bandlimited, it is said

to have infinite bandwidth or an infinite
spectrum
• Time-limited signals cannot be
bandlimited and thus all time-limited
signals have infinite bandwidth
• However, for any well-behaved signal x(t)
it can be proven that lim X ( )  0
 
whence it can be assumed that
| X ( ) | 0   B
B being a convenient large number
 Inverse Fourier Transform

• Given a signal x(t) with Fourier transform

X ( ), x(t) can be recomputed from X ( )
by applying the inverse Fourier transform
given by

1
 X ( )e
j t
x(t )  d , t  
2
• Transform pair
x(t )  X ( )
Properties of the Fourier Transform

x(t )  X ( ) y (t )  Y ( )
• Linearity:
 x(t )   y (t )   X ( )   Y ( )
• Left or Right Shift in Time:

x(t  t0 )  X ( )e  j t0
• Time Scaling:
1  
x(at )  X  
a a
Properties of the Fourier Transform

• Time Reversal:
x(t )  X ( )
• Multiplication by a Power of t:
n
d
t x(t )  ( j )
n n
X ( )
d n
• Multiplication by a Complex Exponential:

x(t )e j0t  X (   0 )
Properties of the Fourier Transform
• Multiplication by a Sinusoid (Modulation):
j
x(t )sin( 0t )   X (   0 )  X (   0 ) 
2
1
x(t ) cos( 0t )   X (   0 )  X (   0 ) 
2
• Differentiation in the Time Domain:
n
d
n
x(t )  ( j ) X ( )
n

dt
 Properties of the Fourier Transform
• Integration in the Time Domain:
t
1
 x( )d  j X ( )   X (0) ( )
• Convolution in the Time Domain:

x(t )  y (t )  X ( )Y ( )
• Multiplication in the Time Domain:

x(t ) y (t )  X ( )  Y ( )
Properties of the Fourier Transform
• Parseval’s Theorem:
1
 x(t ) y(t )dt  2  X

( )Y ( )d

1
if y (t )  x (t )  | x (t ) | dt   | X ( ) | d
2 2


2 
• Duality:

X (t )  2 x(  )
Properties of the Fourier Transform -
Summary
 Linearity of the Fourier
Transform
• The Fourier transform is a linear function of x(t)
F
x1 (t )  X 1 ( j )
F
x2 (t )  X 2 ( j )
F
ax1 (t )  bx2 (t )  aX 1 ( j )  bX 2 ( j )
• This follows directly from the definition of the
Fourier transform (as the integral operator is linear)
& it easily extends to an arbitrary number of signals
• Like impulses/convolution, if we know the Fourier
transform of simple signals, we can calculate the
Fourier transform of more complex signals which are
a linear combination of the simple signals
 Fourier Transform of a Time Shifted
•
Signal
We’ll show that a Fourier transform of a signal which has a simple
time shift is:
F {x(t  t0 )}  e  jt0 X ( j )
• i.e. the original Fourier transform but shifted in phase by –wt0

• Proof
• Consider the Fourier transform synthesis equation:

x(t )  2  X ( j )e jt d 
1


x(t  t0 )  1
2  
X ( j )e j ( t t0 ) d

  e 

 jt0
 2
1
X ( j ) e jt d 


• but this is the synthesis equation for the Fourier transform

• e-jw0tX(jw)
 Example: Linearity & Time Shift
• Consider the signal (linear sum of two time
shifted rectangular pulses)
x(t )  0.5 x1 (t  2.5)  x2 (t  2.5) x1(t)

• where x1(t) is of width 1, x2(t) is of width 3,

t
centred on zero (see figures)
• Using the FT of a rectangular pulse L10S7
x2(t)
X 1 ( j )  2 sin( / 2)

t
X 2 ( j )  2 sin(3 / 2)
 x (t)
• Then using the linearity and time shift
Fourier transform properties
t

X ( j )  e  j 5 / 2   sin( / 2)  2 sin(3 / 2)  
 
 Fourier Transform of a Derivative
• By differentiating both sides of the Fourier transform
synthesis equation with respect to t:
dx(t )
dt
 1
2 

j X ( j )e jt d 
• Therefore noting that this is the synthesis equation for
the Fourier transform jwX(jw)
dx(t ) F
 j X ( j  )
• This is verydtimportant, becauseit replaces
differentiation in the time domain with
multiplication (by jw) in the frequency domain.
• We can solve ODEs in the frequency domain using
algebraic operations (see next slides)
Convolution in the Frequency Domain
• We can easily solve ODEs in the frequency domain:
F
y (t )  h(t ) * x(t )  Y ( j )  H ( j ) X ( j )
• Therefore, to apply convolution in the frequency domain, we just
have to multiply the two Fourier Transforms.
• To solve for the differential/convolution equation using Fourier
transforms:
1. Calculate Fourier transforms of x(t) and h(t): X(jw) by H(jw)
2. Multiply H(jw) by X(jw) to obtain Y(jw)
3. Calculate the inverse Fourier transform of Y(jw)
• H(jw) is the LTI system’s transfer function which is the Fourier
transform of the impulse response, h(t). Very important in the
remainder of the course (using Laplace transforms)
• This result is proven in the appendix
 Example: Linearity
x(t )  p4 (t )  p2 (t )

 2   
X ( )  4sinc    2sinc  
    
Example: Time Shift
x(t )  p2 (t  1)

    j
X ( )  2sinc   e
 
 Example: Time Scaling

 
p2 (t ) 2sinc  
 

 
sinc  
p2 (2t )  2 

a  1 time compression  frequency expansion

0  a  1 time expansion  frequency compression
 Example: Multiplication in Time

x(t )  tp2 (t )

d     d  sin    cos   sin 

X ( )  j  2sinc     j 2    j2
d     d    2
Example: Multiplication in Time –
Cont’d
 cos   sin 
X ( )  j 2
 2
Example: Multiplication by a Sinusoid
x(t )  p (t ) cos( 0t ) sinusoidal
burst

1   (   0 )    (  0 )  
X ( )   sinc     sinc  
2  2   2 
 Example: Multiplication by a
Sinusoid – Cont’d
1   (   0 )    (  0 )  
X ( )   sinc     sinc  
2  2   2 

 0  60 rad / sec

  0.5
Example: Integration in the Time
Domain

 2|t |
v(t )  1   p (t )
  

dv(t )
x(t ) 
dt
 Example: Integration in the Time
Domain – Cont’d
• The Fourier transform of x(t) can be easily
found to be
        
X ( )   sinc    j 2sin  
  4    4 
• Now, by using the integration property, it is

1  2   
V ( )  X ( )   X (0) ( )  sinc  
j 2  4 
Example: Integration in the Time
Domain – Cont’d
 2   
V ( )  sinc  
2  4 
 Generalized Fourier Transform

• Fourier transform of  (t )

  (t )e
 j t
dt  1   (t )  1
• Applying the duality property
x(t )  1, t    2 ( )
  
generalized Fourier transform
of the constant signal x(t )  1, t  
Generalized Fourier Transform of
Sinusoidal Signals

cos( 0t )     (   0 )   (   0 ) 

sin( 0t )  j   (   0 )   (   0 ) 
Fourier Transform of Periodic Signals

• Let x(t) be a periodic signal with period T;

as such, it can be represented with its
Fourier transform

x(t )  
k 
ck e jk 0t
 0  2 / T
• Since e j0t  2 (   ) , it is
0

X ( )  
k 
2 ck (  k 0 )
 Fourier Transform of
the Unit-Step Function
• Since t
u (t )    ( )d
using the integration property, it is
t
1
u (t )    ( )d    ( )

j
Common Fourier Transform Pairs
 Fourier Transforms are used in
• X-ray diffraction
• Electron microscopy (and diffraction)
• NMR spectroscopy
• IR spectroscopy
• Fluorescence spectroscopy
• Image processing
• etc. etc. etc. etc.