Fourier Transform

(Part 3)

1
 The Fourier Series
The Fourier series is a mathematical tool used for
analyzing periodic functions.

Any periodic signal f(t) ,maybe expressed as

summation of simple sines and cosines of frequencies
that are integer multiples of the fundamental
frequency of f(t).
 The Fourier Series
n 
f (t )   F (n) e
n  
jnt

T /2
1

 jnt
F ( n)  f (t ) e dt
T T / 2
 The Fourier Transform
The Fourier Transform is named in honour of French
mathematician, Joseph Fourier, is a certain linear
operator that maps functions to other functions.

The Fourier Transform of a signal x(t) can be thought

of as that signal in the “frequency domain”
 The Fourier Transform

x(t) X ( fs )

How does Fourier Transform differ from Laplace

Transform and z-Transform in its application?

Laplace time
Z - Transform time
Fourier frequency
 Fourier Transform
The function X ( ) is called the Fourier Transform of
x(t):

X ( )  F{x(t )}   x(t )e  j t
dt


X ( )
The Inverse Fourier Transform of is:
1 
 X ( )e jt d
1
x(t )  F { X ( )} 
2 
 The Fourier Transform
The continuous Fourier transform is defined as
follows: 

 f (t ) e
 j t
F ( )  dt

And the inverse continuous Fourier Transform, is
defined as:

1

j t
f (t )  F ( ) e d
2 
CFT→DFT
 From DFT to FFT
Fourier Transform = can be divided into FOUR categories.

Fourier Aperiodic Signal extend to positive and

Transform negative infinity without repeating
Continuous
in pattern.

Periodic Signal extend to positive and

Fourier negative infinity and repeats itself
Series Continuous
in a pattern.

Aperiodic Discrete point signal extend to

DTFT positive and negative infinity
Discrete
without repeating in pattern.

DFT Periodic Discrete point signal extend to

positive and negative infinity and
Discrete
repeats itself in a pattern.
 Derivation of DFT
Fourier Transform Pair:



 j 2ft
X(f )  x (t ) e dt (1)



1
 X(f )e
j 2ft
x(t )  df (2)
2 
 Derivation of DFT
Consider if x(t) is periodic and repeating:
1 T0

 j 2kf 0t
X (kf 0 )  x(t )  e dt (3)
T0 0

x(t )  k   X (kf 0 )  e
 j 2kf 0t (4)

= (3) and (4) represent (1) and (2)

 Derivation of DFT
Now consider x(t) is sampled N times per period with
sampling period T:

xnT    x(t )   (t  nT )
n  
(5)
 Derivation of DFT
Replace (3) with (5)
1  T0
X (kf 0 ) 
T0

n  
0
x(t )   (t  nT )  e  jk 2f 0t
dt (6)

Since
1 N 1
X kf 0    x[ nT ]  e  jk 2f 0 nT
(7)
N n 0
 Derivation of DFT

Fig1: Sampled sequence for DFT

 Derivation of DFT
1 T0  NT
Since f0 
T0
1 (8)
f 0T 
N
N 1
X [ k ]   x[n]  e ( j 2nk ) / N (9)
n 0

1 N 1
x[n] 
N
 X
n 0
[ k ]  e ( j 2nk ) / N
(10)
 Derivation of DFT
Substitution by W:

W nk
N e  ( j 2nk ) / N

 e 
 ( j 2 ) / N nk

DFT becomes:
N 1
X [k ]   x[n]  W nk
N
(12)
n 0
 Derivation of DFT
and WN is the Nth root of unity:
 j ( 2 / N )
WN  e
the Inverse DFT (IDFT) is:
1 N 1
x[n] 
N
 X [k ] W
n 0
 kn
N

n  0,1.., N  1
 Properties of DFT
DFT variables n and k must be restricted to the range
0  n , k  N the DFT shifts x[n  n0 ] or X [k  k0 ] imply
x[n  n0or
]mod N X [k  kwhere
0 ]mod N the modulo

notation:
[m]mod N  m  iN

for some integer i such that:

0  [m]mod N  N
 Properties of DFT
For example, if x[n]   [n  3] then,

x[n  4]mod 6   [n  7]mod 6

  [ n  7  6]
  [n  1]
The DFT shift is also known as circular shift.
 Properties of DFT
Linearity:

a1 x1[n]  a2 x[n]  a1 X 1[k ]  a2 X 2 [k ]

Time Shifting:
kn0
x[n  n0 ]mod N  W N X [k ]
 j ( 2 / N )
(WN  e )
 Properties of DFT
Frequency Shifting:

WN kn0 x[n]  X [k  k0 ]mod N

Conjugation:
kn0
x[n  n0 ]mod N  W N X [k ]
 Properties of DFT
Time Reversal:

x[n]mod N  X [k ]mod N

Duality:

X [n]  Nx[ k ]mod N

 Properties of DFT
Circular Convolution:

x1[n]  x2 [n]  X 1[k ] X 2 [k ]

where
N 1
x1[n]  x2 [n]   x1[i ] x2 [n  i ]mod N
i 0
 Properties of DFT
Multiplication:
1
x1[n]x2 [n]  X 1[k ]  X 2 [k ]
N
where
N 1
X 1[k ]  X 2 [k ]   X 1[i ] X 2 [k  i ]mod N
i 0
 Properties of DFT
Additional Properties:

When x[n] is real, let x[n]  xe [ n]  xo [ n]

where xe [n] and xo [n] are the even and odd
components of x[n] respectively. Let,
x[n]  X [k ]  A[k ]  jB[k ]
| X [k ] | e j [ k ]
 Properties of DFT
Additional Properties:

Then, X [k ]mod N  X  [k ]

xe [n]  Re{ X [k ]}  A[k ]

xo [n]  j Im{ X [k ]}  jB[k ]

 Properties of DFT
Additional Properties:


From X [k ]mod N  X [k ]
A[ k ]mod N  A[k ]
B[ k ]mod N   B[k ]

| X [ k ] |mod N | X [k ] |

 [k ]mod N   [k ]
 Properties of DFT
Parseval’s Relation:
N 1
1 N 1


n 0
2
| x[n] | 
N
 | X [k ] |
n 0
2
 Matrix Interpretation of DFT
W = root of unity
6
W 8
5
W 8 W 7
8

W 4 W 80  W 88
8

1
W 3
8
W 8

2
W 8

Fig 2: Argand Diagram of 8-point DFT

 Matrix Interpretation of DFT
By matrix notation:

X [0] WN0 WN0 WN0 WN0 x[0]

X [1] WN0 WN1 WN2 WN( N 1) x[1]

X [2] = WN0 WN2 WN4 WN2( N 1) x[2]

( N 1) 2
X [ N  1] W0
N W( N 1)
N W 2( N 1)
N
W N
x[ N  1]

(13)
 Matrix Interpretation of DFT
Each row consists of the samples of complex sine
waves of different frequencies

DFT output = result of correlating input samples with

complex sine waves having frequencies equal to
multiples of “fundamental”.
Matrix Interpretation of DFT
X [0] x[0]

X [1] x[1]

X [2] x[2]

X [3] = x[3]

X [4] x[4]

X [5] x[5]

X [6] x[6]

X [7] x[7]

Fig 3: Matrix Representation of 8-point DFT

 Example 1
Compute the DFT of the four‐point sequence:

x[n] = ( 0 1 2 3 )
 Argand Diagram

Fig 4: Argand Diagram of 4 point DFT

 Solution
The matrix of four‐point sequence, W4 may be
expressed as:
 Solution
Then,
 Example 2
Figure 1 shows the 4 point sequence of signal
x[n]={2,0,-1,3}
x[n]
3
2
1
n
0 1 2 3
-1

N 1
(a) Find the DFT by X [k ]   x[n] e  j 2kn / N
n 0
(b) Illustrate the matrix interpretation of DFT
 Solution (a)
N 1
Applying X [k ]   x[n] e  j 2kn / N
n 0
3
X [0]   x[n]  2  0  1  3  4
n 0
3
X [1]   x[n] e  jn / 2  x[0]  x[1] e  j / 2  x[2] e  j  x[3] e  j 3 / 2
n 0
 33 j
3
X [2]   x[n] e  jn  x[0]  x[1] e  j  x[2] e  j 2  x[3] e  j 3
n 0
 2
3
X [3]   x[n] e  j 3n / 2  x[0]  x[1] e  j 3 / 2  x[2] e  j 3  x[3] e  j 9 / 2
n 0
 33 j
41
42
 Solution (b)
W40 W40 W40 W40  1 1 1 1  1 1 1 1 
 0   1 2 3  
W W 1
W42 W43  1 W4 W4 W4  1  j  1 j 
W4   40 4
 
W4 W42 W44 W46
1 W4 W4 W4  1  1 1  1 
2 0 2

 0 9  3 2 1  
W4 W4
3
W46 W4   1 W4 W 4 W 4  1 j  1  j 

 4 
3  3 j 
X 4  W4 x4   
 2 
 
3  3 j 
 Matrix Interpretation of DFT
X  WN x

If we assume the inverse of WN exists, then:

x  WN1 X
1 
x  WN X
N
1 1 
W N  WN
N
 Matrix Interpretation of IDFT
example: W41

1 1 1 1 
 
1 1 j 1  j 
W41  
4 1 1 1 1 
 
1  j  1 j 

 Example 3
Compute the IDFT of X[k] given as follows:

 X [0]  6 
 X [1]   2  j 2
  
 X [2]   2 
   
 X [3]  2  j 2
 Solution
The IDFT x[n] of X[k] is:
 x[0] 1 1 1 1  6 
 x[1]  1 j  1  j   2  j 2
 1
 
 x[ 2] 4 1  1 1  1   2 
    
 x[ 3]  1  j  1 j    2  j 2

 0  0 
   
1  4  1 
 
4  8  2
   
12 3
TQ

48
 Quiz 2g
Figure 1 shows the 4 point sequence of signal
x[n]={1,2,3,4}
x[n]
4
3
2
1
n
0 1 2 3
-1

N 1
(a) Find the DFT by X [k ]   x[n] e  j 2kn / N
n 0
(b) Illustrate the matrix interpretation of DFT