Ecc302a 2

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 24

Discrete Fourier Transform

1
Faculty of Engineering & Technology © Ramaiah University of Applied Sciences
Lecture 2
Properties of DFT
Lecture delivered by:Vasanthavalli.S

2
Faculty of Engineering & Technology © Ramaiah University of Applied Sciences
Objectives
At the end of this lecture, student will be able to:

• State the importance of analyzing signal in frequency domain


• Describe DFT and its frequency domain representation of discrete time
signal
• Describe the properties of DFT

3
Faculty of Engineering & Technology © Ramaiah University of Applied Sciences
Topics

• Fourier Analysis of Discrete Time Signals


• The Discrete Time Fourier Transform (DTFT)
• Development of DFT from DTFT
• DFT-Analysis and Synthesis
• DFT Properties

4
Faculty of Engineering & Technology © Ramaiah University of Applied Sciences
Fourier Analysis of Discrete Time Signals

5
Faculty of Engineering & Technology © Ramaiah University of Applied Sciences
Fourier Analysis of Discrete Time Signals

6
Faculty of Engineering & Technology © Ramaiah University of Applied Sciences
Fourier Analysis of Discrete Time Signals

For a discrete time sequence we define two classes of Fourier Transforms:


• DTFT (Discrete Time FT) for sequences having infinite duration
DTFT
Infinite,aperiodic x[n] continuous X(w),periodic with 2

• DFT (Discrete FT) for sequences having finite duration.


DFT
Finite{aperiodic or periodic} x[n] discrete X(K),periodic with N
N

7
Faculty
©M. S. of
Ramaiah
Engineering
University
& Technology
of Applied Sciences © Ramaiah University of Applied Sciences
The Discrete Time Fourier Transform (DTFT)
• Given a sequence x(n) having infinite duration, we define the DTFT as
follows:

X ( )  DTFT  x (n)   x ( n ) e  j n

n 


x (n)  IDTFT  X ( ) 
1
2 

X ( )e j n d

x(n) X ( )

….. …..

N 1 n   

discrete time continuous frequency 8


Faculty
©M. S. of
Ramaiah
Engineering
University
& Technology
of Applied Sciences © Ramaiah University of Applied Sciences
Development of DFT from DTFT
• Limitation of DTFT in practice:
• Infinite number of arithmetic operations
• Infinite number of points in creating frequency domain (continuous)and
so it can’t be computed by a digital computer.
• So making as discrete frequency domain, sampling is performed at N
equally spaced frequency points in the period 0≤ω≤2𝜋
2𝜋𝑘
• The sampling frequency are denoted as ωk= ;for k=0,1,2,…N-1
𝑁
• Now DTFT becomes DFT. The DFT sequences consisting of N samples
of DTFT.

9
Faculty of Engineering & Technology © Ramaiah University of Applied Sciences
Discrete Fourier Transform (DFT)
• Discrete Fourier Transform: N point DFT is defined as

for k=0,1,2,….N-1
X[K] → DFT harmonic function of x[n] and k → harmonic index
• Inverse Discrete Fourier Transform: Discrete time signal x(n) is expressed as

for n=0,1,2,…N-1

• The DFT and IDFT pair

10
Faculty of Engineering & Technology © Ramaiah University of Applied Sciences
DFT - Introducing Roots of Unity
−𝑗2𝜋
• WN=𝑒 Nth root of unity (W NN=𝑒 −𝑗2Π =1) This notation makes
𝑁

handling the exponential easier. This is sometimes called the "twiddle


factor.“
𝑁−1 𝑁−1
−𝑗2𝜋𝑛
𝑒 𝑁 = 𝑊𝑘𝑛
𝑛=0 𝑛=0
• Where WNk = cos(2k/N)-j sin(2k/N)

11
Faculty of Engineering & Technology © Ramaiah University of Applied Sciences
Twiddle factor (N=8)

1 1
- √2+j√2 j −𝑗2𝜋(2)
𝑒 8 =j
−𝑗2𝜋(1) −𝑗𝜋 𝜋 𝜋
=cos 4 −𝑗 𝑠𝑖𝑛 4
1 1
𝑒 8 =𝑒 4

+j
√2 √2
-1 1

1 1
1 1 −j√2
√2
- √2- -j√2
-j

12
Faculty of Engineering & Technology © Ramaiah University of Applied Sciences
DFT-Analysis and Synthesis

 j (2 / N )
• Phase factor or Twiddle factor WN  e

 N 1
• DFT Analysis  x(n)WNkn 0  n  N 1
X (k )   n 0

0 otherwise

N 1
 X (k )WNkn 0  n  N 1
• DFT Synthesis x ( n)   k  0

0 otherwise

x(n) 
 X (k )
DFT

13
Faculty of Engineering & Technology © Ramaiah University of Applied Sciences
DFT Properties
• Linearity
• Periodicity
• Time reversal
• Circular time shift
• Circular frequency shift
• Conjugation
• Multiplication
• Symmetry
• Circular convolution
• Parseval’s Theorem

14
Faculty of Engineering & Technology © Ramaiah University of Applied Sciences
Linearity
• DFT of a linear weighted combination of two or more signal is equal to
linear weighted combination of DFT of individual signals.

x1 (n)  X 1 (k )
DFT

x1 ( n)
Duration N1
n
0 N11

x2 (n) Duration N2
n
0 N21
x2 (n)  X 2 (k )
DFT

ax1 (n)  bx2 (n)  aX 1 (k )  bX 2 (k )


DFT

15
Faculty of Engineering & Technology © Ramaiah University of Applied Sciences
Periodicity ,Time Reversal
• Periodicity:
If x(n) and X(K) are N point DFT pair then,
x(n +N)=x(n); for all n
X(k+N)=X(k); for all k

• Time reversal:
Reversing the N point sequence in time is equal to reversing the
DFT sequence
x (n) DFT X (k)
x (N- n) DFT X (N- k)

16
Faculty of Engineering & Technology © Ramaiah University of Applied Sciences
Circular Time and Frequency Shift
• Circular time shift : If a discrete time signal is circularly shifted in time by
−𝒋𝟐𝝅𝒌𝒎
m units then its DFT is multiplied by𝒆 𝑵

−𝑗2𝜋𝑘𝑚
x( (N- m))N DFT X (k)𝑒 𝑁

𝒋𝟐𝝅𝒎𝒏
• Circular frequency shift : If a discrete time is multiplied 𝒆 𝑵 its DFT is
circularly shifted by m units

𝒋𝟐𝝅𝒎𝒏
x (n) 𝒆 𝑵 DFT X( (k- m))N

17
Faculty of Engineering & Technology © Ramaiah University of Applied Sciences
Conjugation , Multiplication

• Conjugation: Let x(n) be a complex N point sequence and x*(n) be its


conjugate sequence

x* (n) DFT X* (N-k)

• Multiplication: DFT of product of two discrete time sequences is equal to


1
DFT of individual sequences scaled by a factor
N

DFT 1
x1 (n) x2(n) X1 (K) N X2 (K)
N

18
Faculty of Engineering & Technology © Ramaiah University of Applied Sciences
Symmetry Property

X(N-k)=X*(k)
• Example:
For N=8 point DFT
X(8-2)=X*(2)
X(6)=X*(2)

19
Faculty of Engineering & Technology © Ramaiah University of Applied Sciences
Circular Shift of a Sequence

~x1 (n)  x((n  m)) N 0  n  N 1


x1 (n)  
0 otherwise

x(n)

n
0 N
~
x ( n)

n
0 N ~
x1 (n)  ~
x (n  m)

n
0 N 20
Faculty of Engineering & Technology © Ramaiah University of Applied Sciences
Circular or Periodic Convolution

• Folding and shifting (rotating)operations performed in circular fashion


by computing index of one of the sequences by modulo N operation
• It can be performed with non periodic sequences by periodically
extending them
• Circular convolution of two sequences: at least one of the sequences
should be periodic
• Circular convolution can be performed of two finite sequences of same
number of samples
• Circular convolution of two finite sequences with different number of
sample : Convert the smaller size sequence to the length of larger size
sequence by appending zeros(zero padding)

21
Faculty of Engineering & Technology © Ramaiah University of Applied Sciences
Circular Convolution cont’d
• Circular convolution : If x1(n)and x2(n)are two periodic sequences
x3(n)= 𝑁−1
𝑚=0 x1(m)x2( (n- m))N or
x3(n)= 𝑁−1
𝑚=0 x2(m)x1( (n- m))N where m – dummy variable

22
Faculty of Engineering & Technology © Ramaiah University of Applied Sciences
Circular Convolution , Parseval’s Theorem

• Circular convolution : The DFT of circular convolution of two


sequences is equal to product of their individual DFT

x1 (n) N x2 (n) DFT X1(K) X2(K)

• Parseval’s theorem:
𝑁−1 𝑁−1
DFT
𝑥1 𝑛 𝑥2 ∗ 𝑛 X1(K) X2∗(K)
𝑛=0 𝑘=0

23
Faculty of Engineering & Technology © Ramaiah University of Applied Sciences
Summary
• The DFT of a discrete time signal can be obtained by sampling the DTFT of
a signal
• The Fourier transform of a discrete time signal is periodic with 2π
• DFT supports only circular convolution
• Appending zeros to a sequence in order to increase its length is called zero
padding
• In N point DFT of a finite duration sequence x(n) of length L , the value of N
should be N≥L
• The N point DFT of a L point sequence will have N periodicity
• The term W in phase factor represents a complex number 1∠-2π
• Convolving two signal in time domain is equivalent to multiplying their
spectra in frequency domain
24
Faculty of Engineering & Technology © Ramaiah University of Applied Sciences

You might also like