SIGNALS AND SYSTEMS

For
Graduate Aptitude Test in Engineering
By
Dr. N. Balaji
Professor of ECE
JNTUK
Session: 8
Topic : Discrete Time Fourier Transform
Date : 20.05.2020

By
Dr. N. Balaji
Professor of ECE
 Syllabus

Continuous-time signals: Fourier series and Fourier transform

representations, sampling theorem and applications; Discrete-Time
Signals: Discrete-Time Fourier Transform (DTFT), DFT, FFT, Z-
transform, interpolation of discrete-time signals;
LTI systems: definition and properties, causality, stability, impulse
response, convolution, poles and zeros, parallel and cascade
structure, frequency response, group delay, phase delay, digital
filter design techniques.

3
Discrete Time Fourier Transform
5
 Fourier Transforms
Time Frequency
Fourier Series (FS) Continuous periodic Discrete infinite ck
x(t)
~
Fourier Transform (FT) Continuous infinite Continuous infinite
x(t) X(W)
Discrete-Time FT Discrete infinite x[n] Continuous periodic
(DTFT) X(ejw)
Discrete FT (DFT) Discrete finite/pdc Discrete finite/pdc
~
x[n] X[k]
6
 The Discrete-Time Fourier transform

Definition of DTFT
The discrete-time Fourier transform (DTFT) X  e 
jw
of a
discrete sequence x[n] is defined by
Summation (not integral)
 Discrete (normalized)
X e jw
   x  n e  jw n
frequency variable w
n 
Argument is ejw, not plain w
As can be seen from the definition, The discrete-time Fourier transform
(DTFT) X  e  of a sequence x[n] is a function of the normalized angular
jw

frequency.

7
 DTFT Example

8
 The Discrete-Time Fourier Transform
Derivation: (Analogous to CTFT except 𝒋𝝎𝒏 𝒋(𝝎 𝟐𝝅)𝒏
)
• -aperiodic and (for simplicity) of finite duration
𝑵
• N is large enough so that = 0 if
𝟐
𝑵
• for and periodic with N
𝟐

9
for any n as
 The Discrete-Time Fourier Transform
, DTFS Synthesis Equation

DTFS Analysis Equation

periodic in with period

10
 DTFT Derivation (Home Stretch)

11
 DT Fourier Transform Equations

– Analysis Equation
– DTFT

– Synthesis Equation
– Inverse DTFT

12
 Transform Requirements
 Synthesis Equation:
 Finite Integration Interval, X must be finite
 Analysis Equation:
 Need conditions analogous to CTFT, e.g.

— Finite energy

— Absolutely summable

13
 Examples
 Unit Samples

14
 Decaying Exponential

Infinite sum formula

15
 Example 1: 1st Order System, Decay Power

 Calculate the DT Fourier transform of the signal:

x [ n ]  a n u [ n ], | a | 1 stable system
 Therefore:

X (e jw
)  
n  
a n u [ n ]e  jw n

1
 
n0
( ae  jw
)n 
1  ae  jw

16
 Example 2: Rectangular Pulse
 Consider the rectangular pulse
N1=2
1 | n | N
x[n ]   1

0 | n | N 1

 and the Fourier transform is

N1
X (e jw
)  
n N1
e jw n

2 N1
 e jw N 1

m0
e jw m

jw ( N 1  1 / 2 )
e 1  e  jw ( 2 N 1  1 )

e jw (1 / 2 ) 1  e  jw
e jw ( N 1  1 / 2 )  e  jw ( N 1  1 / 2 )

e jw / 2  e  jw / 2
sin( w ( N 1  1 / 2 ))

sin( w / 2 )
17
 Rectangular Pulse

FIR LPF

18
 Ideal DT LPF

19
 DTFTs of Periodic Functions
 Complex Exponentials
Recall CT result:

What about DT:

a) We expect an impulse (of area 2) at wwo
b) But X(ejw) must be periodic with period 2
In fact

Note: The integration in the synthesis equation is over 2π period,

only need X(ejw) in one 2π period. Thus,

20
 DTFT of General Periodic Functions Using FS
FS

by
superposition

21
 DTFT of Sine Function

22
 DTFT of DT Unit Sample Train

— Also periodic impulse train – in the frequency domain!

23
23
 Properties of the DT Fourier Transform
 Periodicity and Linearity

— Different from CTFT

24
24
 Time and Frequency Shifting

— Important implications in DT because of periodicity

Example

25
 Time Reversal and Conjugate Symmetry

26
 Time Expansion
7) Time Expansion Time scale in CT is infinitely fine
Recall CT property:
But in DT: x[n/2] makes no sense
x[2n] misses odd values of x[n]
But we can “slow” a DT signal down by inserting zeros:
k — an integer ≥ 1
x(k)[n] — insert (k - 1) zeros between successive values

Insert two zeros

in this example

27
 Time Expansion (continued)

— Stretched by a factor
of k in time domain

— compressed by a factor
of k in frequency domain

28
 The Convolution Property

29
 Convolution Property

30
 `
 Convolution: x[n]  g[n] h[n]
 X(e )  n g[n] h[n]e
jw   jwn

 n 
k g[k]h[n  k] e 
 jwn

 k g[k]e  jwk
n h[n  k]e  jw (nk )

 G(e jw )  H (e jw )
Convolution
g[n] h[n]  G(e jw )H(e jw ) becomes
multiplication

31
31
 Multiplication Property or DTFT modulation

Periodic convolution

32
 DTFT and convolution Multiplication Property

Modulation: x[n]  g[n]  h[n]


Could solve if g[n] was just sinusoids...
 1  
X(e )  n    G(e )e d  h[n]e jwn
jw j jn

2 
1 

2
 
G(e j
)  n
 h[n]e  j (w  )n

d
1 
 g[n]  h[n] 
2
 
G(e j
)H (e j (w  )
)d
Dual of convolution in time

33
33
 Discrete Time Fourier Transform

F
• A few points
• DTFT is periodic in frequency with period of 2

• X[n] is a discrete signal

• DTFT allows us to find the spectrum of the discrete signal as viewed
from a window
34
 DT Fourier Transforms
1. W is in radian and it is between 0 and 2 in each discrete time
interval
2. This is different from wwhere it was between –  and +
3. Note that X(W) is periodic

35
 Common Fourier Transform Pairs

𝛿 𝑛 1
𝛿 𝑛−𝑛 e
𝑥 𝑛 =1 2𝜋𝛿 𝛺 , 𝛺 ≤ 𝜋

𝑒 2𝜋𝛿 𝛺 − 𝛺 , 𝛺 , 𝛺 ≤𝜋

cos 𝛺 𝑛 𝜋[𝛿 𝛺 − 𝛺 + 𝛿 𝛺 + 𝛺 ], 𝛺 , 𝛺 ≤𝜋

sin 𝛺 𝑛 −𝑗𝜋[𝛿 𝛺 − 𝛺 − 𝛿 𝛺 + 𝛺 ], 𝛺 , 𝛺 ≤𝜋

𝑢 𝑛 1
𝜋𝛿 𝛺 + , 𝛺 ≤𝜋
1−e
𝛺 𝛿 𝛺 − 𝑘𝛺 =
36
 −𝒖 −𝒏 − 𝟏 𝟏
−𝝅𝜹 𝜴 + 𝒋𝜴
, 𝜴 ≤𝝅
𝟏−𝐞
𝒂𝒏 𝒖 𝒏 , 𝒂 < 𝟏 𝟏
𝟏 − 𝒂𝒆 𝒋𝜴

−𝒂𝒏 𝒖 −𝒏 − 𝟏 , 𝒂 > 𝟏 𝟏
𝟏 − 𝒂𝒆 𝒋𝜴

𝒏 + 𝟏 𝒂𝒏 𝒖 𝒏 , 𝒂 < 𝟏 𝟏
𝟏 − 𝒂𝒆 𝒋𝜴 𝟐

𝒂𝒏, 𝒂 <𝟏 𝟏 − 𝒂𝟐
𝟏 − 𝟐 𝐚𝐜𝐨𝐬 𝜴 + 𝒂𝟐
1 | n | N1 𝟏
x[n]   𝐬𝐢𝐧 𝜴 𝑵𝟏 + 𝟐
0 | n | N1 𝜴
𝐬𝐢𝐧 𝟐
𝐬𝐢𝐧 𝑾𝒏 1 0 | W | W
,𝟎 < 𝒘 < 𝝅 X [W ]  
𝝅𝒏
0 W | W | 
37
 Properties of DTFT

 Remember:
 For time scaling note that m>1  Signal spreading

38
 Properties of DTFT

𝟏 𝟏

𝟐 𝟐

Periodicity:

Linearity: 𝟏 𝟐 𝟏 𝟐

Time Shifting: 𝟎
𝒋𝜴𝒏𝟎

Frequency Shifting: 𝒋𝜴𝟎 𝒏

𝟎

∗
Conjugation:
39
 Properties of DTFT

Frequency differentiation:

First difference : 𝒋𝜴

Accumulation: 𝒌
𝒋𝜴

Convolution: 𝟏 𝟐 𝟏 𝟐

Multiplication: 𝟏 𝟐
𝟏 𝟐

Real Sequence : 𝒆 𝒐

Parseval’s relation 𝟏 𝟏 𝟐
𝟏 𝟐
40 𝟐𝝅
 Problem on DTFT
+ + with its Fourier Transform
Then the value of at w= is

a) 9 b) 5 c) d) 7

Solution: Option (d)

=2+1+1+1+2

4+2+1=7
41
 Problem on DTFT

A discrete time signal given by =

0 , elsewhere

If Fourier transform is then the value of integral is

a) b) 170 c) d) 2

42
 Solution to the Problem

43
 Problem on DTFT

44
 Solution

45
 Solution to the Problem

46
 Solution to the Problem

47
 Problem and Solution

X (e jw
)  
n  
a n u [ n ]e  jw n

1
 
n0
( ae  jw
)n 
1  ae  jw

48
 Problem

49
 Solution to the Problem

50
 Problem

51
 Solution

52
 Problem and Solution

53
 Problem

54
 Solution

55
 Problem and Solution

56
 Problem and Solution

57
 Problem and Solution

58
 Problem

59
 Solution

60
 Problem

61
 Solution

62
 Problem

63
 Solution

64
 Sources, References and Acknowledgement
i) Lecture slides of Michael D. Adams

ii) Lecture slides of Prof. Paul Cuff

iii) A.V. Oppenheim, A.S. Willsky and S. H. Nawab, “Signals and Systems”, PHI, 2nd Edition, 1997.

iv) GATE Previous Questions with Solutions, Nodia Publications

v) MIT Open Courseware http://ocw.mit.edu

vi) https://www.aceenggacademy.com/

vii) https://www.madeeasy.in/

 Disclaimer: The material presented in this presentation is taken from various standard
Textbooks and Internet Resources and the presenter is acknowledging all the authors.
65 Acknowledgement : Lecture slides of Michael D. Adams and Prof. Paul Cuff