DISCRETE FOURIER TRANSFORM

(DFT)

Dr Malaya Kumar Hota (Prof., SENSE, VIT University)

Discrete-Time Fourier Transform
(DTFT)

DR. MALAYA KUMAR HOTA (PROF., SENSE) 2

DTFT

DR. MALAYA KUMAR HOTA (PROF., SENSE) 3

DTFT

DR. MALAYA KUMAR HOTA (PROF., SENSE) 4

Discrete Fourier Transform (DFT)
The DTFT of a sequence x(n) is
(i) Continuous function of frequency .
(ii) Periodic with period 2.

DTFT is not computationally convenient representation.

One period of DTFT

DTFT DFT
Sample at N equidistance point

DR. MALAYA KUMAR HOTA (PROF., SENSE) 5

Discrete Sequence
1

0.9

0.8

0.7

0.6
x(n)

0.5

0.4

0.3

0.2

0.1

0
0 5 10 15 20 25 30 35
n

DR. MALAYA KUMAR HOTA (PROF., SENSE) 6

DTFT to DFT
Relation between DTFT & DFT
5

4.5

3.5
|X(e )|and|X(k)|

2.5
j

1.5

0.5

0
0 0.2 0.4 0.6 0.8 1 2
freq

N-1
DR. MALAYA KUMAR HOTA (PROF., SENSE) 7
Discrete Fourier Transform (DFT)
Let x(n) is a finite duration sequence of length L  N .
i.e., x(n)  0 for n  0 & also for n  L.
L 1
X ( )   x(n)e  jn
n 0
L 1
X (k )  X ( )   2k   x(n)e  jn
N n 0 2k

N
L 1  j 2kn
  x ( n )e N

n 0

DR. MALAYA KUMAR HOTA (PROF., SENSE) 8

Discrete Fourier Transform (DFT)
N - point DFT of x(n) for N  L is
N 1  j 2kn
X ( k )   x ( n )e N

n 0

where k  0,1, 2, ......., N  1

If N > L, then Spectral display will be better due to zero padding.

If N < L, then time domain aliasing takes place.

DR. MALAYA KUMAR HOTA (PROF., SENSE) 9

Discrete Fourier Transform (DFT)
N 1  j 2kn
X ( k )   x ( n )e N
where k  0,1, 2, ......., N  1
n 0

N 1
X (k )   x(n)WNkn where k  0,1, 2, ......., N  1
n 0

Where by defination
 j 2
WN  e N

Phasefactor or Twiddle factor

Symmetry property Periodicity property

N
k
WN 2
 W k WNk  N  WNk
N

DR. MALAYA KUMAR HOTA (PROF., SENSE) 10

Discrete Fourier Transform (DFT)
A finite duration sequence of length L is given by

1 , 0  n  L 1
x ( n)  
0 , otherwise

Determine the N-point DFT of the sequence for N ≥ L.

DR. MALAYA KUMAR HOTA (PROF., SENSE) 11

Discrete Fourier Transform (DFT)

N 1  j 2kn
X ( k )   x ( n )e N
where k  0,1, 2, ......., N  1
n 0
L 1  j 2kn
  (1)e N

n 0
 j 2kL
1 e N
  j 2k
1 e N

DR. MALAYA KUMAR HOTA (PROF., SENSE) 12

Discrete Fourier Transform (DFT)
 jkL
 jNkL  jkL

e N e e N 
 
X (k )   jk  jk  jk

N 
 N 

e e
N
e 
 
 kL 
sin    jk ( L 1)
  N 
e N
 k 
sin  
N 

DR. MALAYA KUMAR HOTA (PROF., SENSE) 13

Inverse Discrete Fourier Transform
(IDFT)
IDFT of X (k ) is
N 1 j 2kn
1
x ( n) 
N
 X ( k )e
k 0
N

where n  0,1, 2, ......., N  1

IDFT of X (k ) is
N 1
1
x ( n) 
N
 X (k )W
k 0
 kn
N

where n  0,1, 2, ......., N  1

DR. MALAYA KUMAR HOTA (PROF., SENSE) 14

DFT as a Linear Transformation
 x(0)   X (0) 
 x(1)   X (1) 
   
Let us define, xN    , XN  
   
   
 x( N  1)  X ( N  1)

1 1 1  1 
1 W 1 WN2  WNN 1 
 N

WN  1 WN2 WN4  WN2 ( N 1) 

 
| | |  | 
1 WNN 1 WN2 ( N 1)    WN( N 1)( N 1) 
DR. MALAYA KUMAR HOTA (PROF., SENSE) 15
DFT as a Linear Transformation

N - point DFT, X N  WN .x N

1
xN 
*
IDFT, WN . X N
N

DR. MALAYA KUMAR HOTA (PROF., SENSE) 16

Example
Find the DFT of the given sequence

1
 , 0n2
x ( n)   3
 0 , elsewhere
Solution:
1 
 3
x3   1 
 3
1 
 3 

DR. MALAYA KUMAR HOTA (PROF., SENSE) 17

Example

W30 W30 W30  1 1 1 

 0    j 4 / 3 
W3  W3 W3 W3   1
1 2
e  j 2 / 3 e 
 0 2 4 1 e  j 4 / 3 e  j 8 / 3 
W
 3 W 3 W 3  
1 1 1 
 
 1  0.5  j 0.866   0.5  j 0.866  
 
1  0.5  j 0.866   0.5  j 0.866  

DR. MALAYA KUMAR HOTA (PROF., SENSE) 18

Example

1 
X 3  W3 x3   0 
 0 

DR. MALAYA KUMAR HOTA (PROF., SENSE) 19

Example
Find the DFT of the given sequence

x(n)  {0,1, 2,3}

Solution:

0
1 
x4   
 2
 
 3
DR. MALAYA KUMAR HOTA (PROF., SENSE) 20
Example

W40 W40 W40 W40  1 1 1 1

 0  
W4 W4 W4 W4  1  j  1 j 
1 2 3

W4   0   
6
W4 W4 W4 W4  1  1
2 4
1  1
 
W 0 9
W4 W4 W4  1
3 6 j  1  j
 4

DR. MALAYA KUMAR HOTA (PROF., SENSE) 21

Example

 6 
 2  j 2 
X 4  W4 x 4   
 2 
 
  2  j 2 

DR. MALAYA KUMAR HOTA (PROF., SENSE) 22

Find the N-point DFT
 1 n0
x[n]  
 0 otherwise
N 1
X [k ]   x[n]WN k  0,1,2,.......,( N  1)
kn

n 0

 1

Find the N-point DFT

 1 nm
y[n]  
 0 otherwise

DR. MALAYA KUMAR HOTA (PROF., SENSE) 23

Find the N-point DFT
 1 nm
y[n]  
 0 otherwise

N 1
Y [k ]   x[n]WN k  0,1,2,......., ( N  1)
kn

n 0

 WNkm

DR. MALAYA KUMAR HOTA (PROF., SENSE) 24

Relationship of DFT with Z-transform

X ( z)   x (n ) z
n 
n

z  re j
On unit circle, r  1 i.e., z  e j
( Z  transform becomes DTFT )
2 k
j
ze N
( Z  transform becomes DFT )

X (k )  X ( z ) z e j 2N k

DR. MALAYA KUMAR HOTA (PROF., SENSE) 25

Relationship of DFT with Z-transform
Let x(n) be a finite duration sequence of length N.
N 1
X ( z )   x(n) z  n IDFT
n 0
N 1
1 N 1
j 2 kn / N   n
   X ( k ) e z
n 0  N k 0 
N 1 N 1
1
X ( z) 
N
 X (k ) e
k 0 n 0
j 2 kn / N
z n

1 N 1
 1  zN 
  X (k ) 
 j 2 k / N 1 
N k 0  1 e z 
1  z  N N 1 1  zN N 1
X (k )
 
X (k )
N k 0 1  e j 2 k / N z 1
X ( z) 
N

k 0 1  e
j 2 k / N 1
z

DR. MALAYA KUMAR HOTA (PROF., SENSE) 26

Relationship of DFT with DTFT

X (k )  X ()   2 k
N

1  e  j N N 1
X (k )
X ( ) 
N

k 0
 2 k 
j   
1 e  N 

 Interpolation formula

DR. MALAYA KUMAR HOTA (PROF., SENSE) 27