Lecture-17-18 Revision-Chap#01

Signals & Systems

MT - 252
Lecture # 17-18: Fourier series representation of discrete-time
periodic signals

Dr. Hafiz Zia Ur Rehman

[email protected]

Department of Mechatronics Engineering

Air University, Islamabad
 Lecture-17-18 Revision-Chap#01

Fourier Series Representation DT periodic Signals

❑ There are some important differences that should be considered before the start
of Fourier series representation of discrete time signals.
❑ In particular, the Fourier series representation of a discrete-time periodic signal is
a finite series, as opposed to the infinite series representation required for
continuous time periodic signals.
❑ Consequently, there is no mathematical issues of convergence (as in CT Fourier
series Convergence).
❑ The following properties of CT exponential are different in the DT case:
➢ CT:- The larger the magnitude of 𝝎𝒐 , the higher is the rate of oscillation in the signal.
➢ CT:- Complex exponential, i.e., 𝒆𝒋𝝎° 𝒕 is periodic for any value of 𝝎𝒐 .
➢ CT:- All harmonically related complex exponentials 𝑒 𝑗𝑘𝝎° 𝑡 , are distinct.
➢ DT:-If 𝜔0 is increased from 0 𝑡𝑜 2𝜋, the rate of oscillations first increases (from 0 to 𝜋) and
then decreases (from 𝜋 to 0) .
➢ DT:- The signal with frequency 𝝎𝟎 is identical to signals with frequencies 𝝎𝟎 ± 𝟐𝝅 , 𝝎𝟎 ±
𝟒𝝅, and so on.
➢ DT:- There are only N distinct periodic exponentials, i.e., k + N  n  = k  n 
 Lecture-17-18 Revision-Chap#01

Fourier Series Representation DT periodic Signals

❑ The set of all discrete-time complex exponential signals that are periodic with period N
is given by
2
k  n  = e
jk n
jk0 n
=e N
, k = 0, 1, 2,
2𝜋
❑ All these signals have fundamental frequencies that are multiples of and thus are
𝑁
harmonically related.
❑ Consider the discrete-time complex exponential:
j (0 + 2 ) n
e = e j0 n e j 2 n  e j0 n  e j 2 n = cos ( 2 n ) + j sin ( 2 n )

❑ This is a consequence of the fact that discrete-time complex exponentials which differ in
frequency by a multiple of 𝟐𝝅 are identical. In fact, the signal with frequency 𝝎𝟎 is
identical to signals with frequencies 𝝎𝟎 ± 𝟐𝝅 , 𝝎𝟎 ± 𝟒𝝅, and so on……..
❑ Specifically, 𝜙0 𝑛 = 𝜙N 𝑛 , 𝜙𝟏 𝑛 = 𝜙N+1 𝑛 , and, in general,

k  n  = k + rN  n  , r, N  Z 3
 Lecture-17-18 Revision-Chap#01

Fourier Series Representation DT periodic Signals

❑ Consider the representation of more general periodic sequences in

terms of linear combinations of the sequences 𝝓𝒌 [𝒏]. Such a linear combination
has the form
2
x  n  =  akk  n  =  ak e jk0 n =  ak e
jk n
N

k k k

❑ Therefore, Since the sequences 𝝓𝒌 [𝒏] are distinct only over a range of N
successive values of 𝒌, the summation in eq. (above) need only include terms
over this range.
❑ Thus, the summation is on k, as k varies over a range of N successive integers,
beginning with any value of k. We indicate this by expressing the limits of the
summation as 𝒌 = 𝑵 . That is,
2
x  n =  akk  n  =  
jk n
ak e jk0 n = ak e N

k = N  k = N  k = N 
 Lecture-17-18 Revision-Chap#01

Fourier Series Representation DT periodic Signals

❑ Thus, the summation is on k, as k varies over a range of N successive integers,

beginning with any value of k. We indicate this by expressing the limits of the
summation as 𝒌 = 𝑵 . That is,
2
x  n =  akk  n  =  
jk n
ak e jk0 n = ak e N

k = N  k = N  k = N 

❑ For example,
➢ 𝒌 could take on the values 𝒌 = 0, 1, … , 𝑁 − 1, or 𝒌 = 1, … , 𝑁 or 𝑘 = 2, 4, … , 𝑁 +
1, and so on …..
➢ In either case, the same set of complex exponential sequences appears in the
summation on the right-hand side of the above equation.

❑ This Equation is referred to as the discrete-time Fourier series and the

coefficients 𝒂𝒌 as the Fourier series coefficients.
 Digital Image Processing, 3rd ed.
Lecture-17-18 Revision-Chap#01

Gonzalez & Woods

Discrete-time Fourier series
www.ImageProcessingPlace.c
om
Chapter 10
2
x  n =  k  n  =Segmentation
akImage  k 
jk n
jk0 n
a e = ak e N

k = N  k = N  k = N 
❑ If 𝒌 is in the range from 0 to 𝑵 − 1,

x  n  = a00  n  + a11  n  + + aN −1N −1  n 

❑ If 𝒌 is in the range from 1 to 𝑵,

x  n  = a11  n  + a22  n  + + aN N  n 
❑ If 𝑘 is in the range from −1 to 𝑵 − 2 (write yourself in the same manner as above)

x  n  = a−1−1  n  + a00  n  + + aN − 2N − 2  n 

❑ If we consider more than 𝑵 sequential values of 𝒌, the values 𝒂𝒌 repeat
periodically with period 𝑵, i.e.,
k  n  = k + rN  n  , r, N  Z

© 1992–2008 R. C. Gonzalez & R. E. Woods

 Digital Image Processing, 3rd ed.
Lecture-17-18 Revision-Chap#01

Gonzalez & Woods

www.ImageProcessingPlace.c
Procedure for determiningChapter
the coefficients
10
𝒂𝒏 om

❑ By following steps parallel Image to those used in continuous time, it is possible

Segmentation
to obtain a closed-form** expression for the coefficients 𝒂𝒌 in terms of
the values of the sequence 𝒙[𝒏].
❑ Assuming that a given discrete time periodic signal can be represented
with the series of eq.
2
x  n  =  ak e , jk0 n
0 =
k = N  N
❑ For determining the coefficients 𝒂𝒌 , multiplying both sides of above eq.
by 𝒆−𝒋𝒓𝝎°𝒏 ,
x  n  e − jr0n =  ak e jk0n .e − jr0n
k = N 

❑ Summing over N terms on both sides,

 x  n  e − jr0n =   j ( k − r )0 n
ak e
n = N  n = N  k = N 
**A closed-form expression is an expression that can be evaluated/solved in a finite number of steps/operations.

© 1992–2008 R. C. Gonzalez & R. E. Woods

 Digital Image Processing, 3rd ed.
Lecture-17-18 Revision-Chap#01

Gonzalez & Woods

www.ImageProcessingPlace.c
Procedure for determiningChapter
the coefficients
10
𝒂𝒏 om

❑ Interchanging the orderImage Segmentation

of summation on the right-hand side,
 j ( k − r )0 n 
 x  n e
n = N 
− jr0 n
=  ak   e
k = N   n = N 


❑ Rewriting the integral inside ‘[ ]’ using Euler's formula,


n = N 
e
j ( k − r )0 n
= 
k = N 
cos ( k − n ) 0 n + j 
k = N 
sin ( k − n ) 0 n

❑ The innermost sum on non the right-hand side of is zero, unless 𝒌 − 𝒓 is

zero or an integer multiple of 𝑵. Therefore, if we choose values for 𝒓
over the same range as that over which 𝒌 varies in the outer summation,
the innermost sum on the right-hand side equals 𝑵 if 𝒌 − 𝒓 and 0 if 𝒌 ≠
𝒓, i.e.,
N , if k − r = 0,  N , 2 N , 
 e
j ( k − r )0 n
=
k = N  0, otherwise

© 1992–2008 R. C. Gonzalez & R. E. Woods

 Digital Image Processing, 3rd ed.
Lecture-17-18 Revision-Chap#01

Gonzalez & Woods

www.ImageProcessingPlace.c
Procedure for determiningChapter
the coefficients
10
𝒂𝒏 om

❑ The right-hand side thenImage Segmentation

reduces to 𝑵𝒂𝒓 , and the final expression will be,
1
ar =  x 
n e − jr0 n

N n = N 
❑ This provides a closed-form expression for obtaining the Fourier series
coefficients, and the discrete-time Fourier series pair (synthesis
equation and analysis equation) will be:

x  n =  ak e jk0 n Synthesis Equation

n = N 

1 2
ak =  x  
n e − jk0 n
, where  0 = Analysis equation
N n = N  N

The set of coefficients 𝒂𝒌 are often called the Fourier series coefficients or
the spectral coefficients of 𝒙 𝒏 .
© 1992–2008 R. C. Gonzalez & R. E. Woods
 Digital Image Processing, 3rd ed.
Lecture-17-18 Revision-Chap#01

Gonzalez & Woods

www.ImageProcessingPlace.c
Example-01: om
Calculate the Fourier series coefficients 𝒂𝒌 for 10
Chapter the following discrete-time periodic
signal Image Segmentation
𝟐𝝅
Note that 𝒙[𝒏] periodic only if ‘ ’ is an integers or a ratio of
x  n  = sin 0 n 
𝝎𝒐
𝟐𝝅
integers. When is an integer 𝑵 , it is periodic with
𝝎𝒐
fundamental period 𝑵.
Solution:-
 2 
x  n  = sin 0 n  , 1 − jk   n
ak =
N
 x  n e
n = N 
 N 

x  n =
2j
(
1 j0 n − j0 n
e −e , ) Comparing with synthesis eq., by
inspection that

 2   2  1 1
1 j  N  n 1 − j  N  n a1 = , a−1 = −
x  n = e − e 2j 2j
2j 2j
The remaining coefficients over the
2 2
1 j (1) n 1 j ( −1) n interval of summation <N>, i.e.,
x  n = e N
− e N
one time period are zero.
2j 2j
1 1
These coefficients repeat with period N; aN +1 = , aN −1 = −
2j 2j

© 1992–2008 R. C. Gonzalez & R. E. Woods

 Digital Image Processing, 3rd ed.
Lecture-17-18 Revision-Chap#01

Gonzalez & Woods

x  n  = A cos (0 n +  )
www.ImageProcessingPlace.c
1 1 om
a1 = , a−1 = − Chapter 10
2j 2j Image Segmentation  2 
N0 = m  
2 
 0
Let 0 = , then
5  2 
N0 = m    5. where ‘m’ is the smallest integer
 2  that makes N as an integer.
 5

The Fourier series coefficients for this example with 𝑁 = 5 are illustrated in Figure
below:

𝟐𝝅
Fourier coefficients for 𝒙 𝒏 = 𝐬𝐢𝐧 5
𝒏 .

The fact that they repeat periodically is indicated. However, only one period is utilized in the synthesis equation.

© 1992–2008 R. C. Gonzalez & R. E. Woods

 Digital Image Processing, 3rd ed.
Lecture-17-18 Revision-Chap#01

Gonzalez & Woods

𝟐𝝅 www.ImageProcessingPlace.c
Consider now the case when is a ratio ofomintegers, i.e., when
𝝎𝒐 Chapter 10
Image
 Segmentation
2
0 = M  
 N0 
Assuming that 𝑴 and 𝑵 do not have any common factors, 𝒙[𝒏] has a fundamental
period of 𝑵. Again expanding 𝒙[𝒏] as a sum of two complex exponentials,
 2   2 
1 jM  N  n 1 − jM  N  n 1 1
x  n = e − e aM = , a− M = −
2j 2j 2j 2j

For N=5, a2 = a−3 .

𝟐𝝅
Fourier coefficients for 𝒙 𝒏 = 𝐬𝐢𝐧 𝟑 𝒏.
5

© 1992–2008 R. C. Gonzalez & R. E. Woods

 Digital Image Processing, 3rd ed.
Lecture-17-18 Revision-Chap#01

Gonzalez & Woods

www.ImageProcessingPlace.c
Example-02: Find the Fourier series coefficient
om 𝑎𝑘 for the following signal:
Chapter 10
Image  2Segmentation
  2   4  
x  n  = 1 + sin   n + 3cos   n + cos  + 
 N   N   N 2
Solution:- Expand 𝑥 𝑛 directly in terms of complex exponentials,

1  j  N  n − j  N  n  3  j  N  n − j  N  n  1  j  N n + 2  − j  N n + 2  
 2   2   2   2   4    4  

x  n = 1 +  e −e  + e +e  + e +e 

2j  2   2  
    
 3 1  j  N  n  3 1  − j  N  n  1 j  2   j 2 N  n  1 − j  2   − j 2 N  n
 2   2     2     2 

x  n = 1 +  +  e +  − e + e e + e e
 2 2 j   2 2 j   2   2 
   
 
j 
   
e 2
= cos   + j sin    j
a0 = 1 2 2
 
1 j  2  1
3 1  3 1 a2 = e  j
a1 =  +   − j 2 2 ak = 0, k  0, 1, 2
2 2j 2 2  
3 1  3 1
1 j  2 
a−2 = e
1
− j or ak = 0, k  2
a−1 =  −   + j 2 2
2 2j 2 2

© 1992–2008 R. C. Gonzalez & R. E. Woods

 Digital Image Processing, 3rd ed.
Lecture-17-18 Revision-Chap#01

Gonzalez & Woods

e  a 
www.ImageProcessingPlace.c
om
Chapter k10
Image Segmentation

m ak 

© 1992–2008 R. C. Gonzalez & R. E. Woods

 Digital Image Processing, 3rd ed.
Lecture-17-18 Revision-Chap#01

Gonzalez & Woods

www.ImageProcessingPlace.c
om
Chapter 10
Image Segmentation
ak

ak

© 1992–2008 R. C. Gonzalez & R. E. Woods

 Digital Image Processing, 3rd ed.
Lecture-17-18 Revision-Chap#01

Gonzalez & Woods

www.ImageProcessingPlace.c
om
Chapter 10
Image Segmentation

© 1992–2008 R. C. Gonzalez & R. E. Woods

 Digital Image Processing, 3rd ed.
Lecture-17-18 Revision-Chap#01

Gonzalez & Woods

www.ImageProcessingPlace.c
om
Chapter 10
Image Segmentation

© 1992–2008 R. C. Gonzalez & R. E. Woods

 Digital Image Processing, 3rd ed.
Lecture-17-18 Revision-Chap#01

Gonzalez & Woods

www.ImageProcessingPlace.c
Example-03: om
Suppose we are given the following information10about a signal 𝒙 𝒏 :
Chapter
Image Segmentation
1. 𝒙 𝒏 is real and even signal.
2. 𝒙 𝒏 has period 𝑵 = 10 and Fourier coefficients 𝒂𝒌 .
3. 𝒂11 = 5.
4.
1 9
   = 50.
2
x n
10 n =0

Show that 𝒙[𝒏] = 𝑨 𝒄𝒐𝒔(𝑩𝒏 + 𝑪), and specify numerical values for the constants A,
B, and C.

© 1992–2008 R. C. Gonzalez & R. E. Woods

 Digital Image Processing, 3rd ed.
Lecture-17-18 Revision-Chap#01

Gonzalez & Woods

Solution:- www.ImageProcessingPlace.c
om
Chapter 10
Image
Fact-1: Since 𝒙 𝒏 is real and even,Segmentation
therefore, 𝒂𝒌 will be,
ak = a− k , e.g., a1 = a−1
Fact-2: Period is given as
N = 10 e.g., a1 = a11
Fact-3: Given:
a11 = 5 e.g., a−1 = a1 = a11 = 5

Fact-4: Given:
1 9
 x  n  = 50.
2

10 n =0
Using Parseval’s relation,

1
 x  n =  a
2
=50.
2
k
N N  N 

© 1992–2008 R. C. Gonzalez & R. E. Woods

 Digital Image Processing, 3rd ed.
Lecture-17-18 Revision-Chap#01

Gonzalez & Woods

1 9
Solution:- www.ImageProcessingPlace.c
  
2
om x n = 50.
Chapter 10 10 n =0

Using Parseval’s relation,Image Segmentation 1

 x  n =  a
2
=50.
2
k
N N  N 
8

 ak = 50,
2

n =−1
8
a−1 + a0 + a1 +  ak = 50,
2 2 2 2

n=2

8
( 5) + a0 + ( 5 ) +  ak = 50,
2 2 2 2

n=2
8
a0 +  ak = 0, ak = 0 for k = 2,3,
2 2
,8.
n=2

Using synthesis equation,

 2   2 
jk  n 8 jk  n
1
N
a e
N 
k
 0 
=  ak e
−1
 0 
,

© 1992–2008 R. C. Gonzalez & R. E. Woods

 Digital Image Processing, 3rd ed.
Lecture-17-18 Revision-Chap#01

Gonzalez & Woods

Solution:- www.ImageProcessingPlace.c
om
Chapter 10
Using synthesis equation,Image Segmentation

 2   
jk  n 8
1 jk   n

N
a e
N 
k
 0 
=  ak e
−1
5
,

   
j n − j n
= a−1e 5
+ a1e 5
,
   
j n − j n
= 5e 5
+ 5e 5
,
 j  5  n − j  5  n 
e   +e   
= 5 2 ,
 2 
  A = 10,

  B= ,
= 10 cos   n. 5
5
C=0
© 1992–2008 R. C. Gonzalez & R. E. Woods
 Digital Image Processing, 3rd ed.
Lecture-17-18 Revision-Chap#01

Gonzalez & Woods

Example-04: Each of the two sequences 𝟏 [𝒏] and 𝒙𝟐 [𝒏] has a period 𝑵 = 𝟒, and the
www.ImageProcessingPlace.c
𝒙om
Chapter
corresponding Fourier series coefficients 10
are specified as:
Image Segmentation
x1  n  ⎯→
FS
ak

x2  n  ⎯→
FS
bk
where,
1
a0 = a3 = a1 = 1, and b0 = b1 = b2 = b3 = 1.
2
Using the multiplication property in Table 3.1, determine the Fourier series coefficients
𝒄𝒌 for the signal 𝒈[𝒏] = 𝒙𝟏 [𝒏] 𝒙𝟐 [𝒏].

© 1992–2008 R. C. Gonzalez & R. E. Woods

 Digital Image Processing, 3rd ed.
Lecture-17-18 Revision-Chap#01

Gonzalez & Woods

www.ImageProcessingPlace.c
Solution: om
Chapter 10
The given information is
Image Segmentation
N =4
1 1
a0 = a3 = a1 = a2 = 1; b0 = b1 = b2 = b3 = 1.
2 2
Using Multiplicative property

g  n  = x1  n  x2  n  ⎯→
FS
ck =  ab l k −l
l = N 

3
ck =  al bk −l  a0bk −0 + a1bk −1 + a2bk − 2 + a3bk −3
l =0
3
ck =  al bk −l  bk −0 + 2bk −1 + 2bk − 2 + bk −3
l =0
3
ck =  al bk −l  1 + 2 + 2 + 1 = 6
l =0

© 1992–2008 R. C. Gonzalez & R. E. Woods

 Digital Image Processing, 3rd ed.
Lecture-17-18 Revision-Chap#01

Gonzalez & Woods

www.ImageProcessingPlace.c
Problem: Consider the function om

N −Chapter
1 j  2  kn 10
 
a k  =  e  N 
Image Segmentation
n =0

a) Show that
a  k  = N for k = 0,  N , 2 N , 3N ,

a) Show that a  k  = 0 whenever k is not an integer multiple of N. (use the finite sum
formula)

a) Repeat parts (a) and (b) if

 2 
j  kn
a k  =  e  N 

N 

© 1992–2008 R. C. Gonzalez & R. E. Woods

 Digital Image Processing, 3rd ed.
Lecture-17-18 Revision-Chap#01

Gonzalez & Woods

N −1  2 
Solution:- www.ImageProcessingPlace.c j  kn

a) Let k = pN , p  Z . Then,
om
Chapter 10 a k  =  e  N 

n =0
Image Segmentation
N −1  2  N −1 N −1
j  pNn
a  pN  =  e  N 
= e j 2 pn
 1 = N
n =0 n =0 n =0

b) Using the finite sum formula, we have

1 − e j 2 k
a k  =  2 
= 0, if k  pN , p  Z
j k
1− e  N 

q + N −1  2 
c) Let j  kn
a k  =  e  N 
where q is some arbitrary integer.
n=q

By putting 𝑘 = 𝑝𝑁, we may again easily show that

q + N −1  2  q + N −1 q + N −1
j  pNn
a  pN  =  e  N 
=  e j 2 pn
  1= N
n=q n=q n=q

Now,  2  N −1  2 
j  kq j  kn
a k  = e  N 
e  N  using part ‘b’ a  k  =0, if k  pN , p  Z
n =0

© 1992–2008 R. C. Gonzalez & R. E. Woods

 Lecture-17-18 Revision-Chap#01

References and Reading assignment

◼ The slides are prepared based on the following textbook:
◼ Chapter #3 (Art. 2.1), signals & systems by Alan V. Oppenheim, S. Hamid
Nawab, 2nd edition, Pearson education - prentice hall
◼ Reading Assignment : 3.6, 3.7,3.8 with respective examples.
◼ Special thanks to
◼ Prof. Dr. Muhammad Shehzad Hanif

26
 Lecture-17-18 Revision-Chap#01

Revision Slides From Chapter-1 Lectures

27
 Lecture-17-18 Revision-Chap#01

Some Basic Signals

◼ Periodicity Property of Discrete-Time Complex Exponential
◼ There are many similarities between continuous-time and discrete-time signals. But
also, there are many important differences. One of them is related with the discrete-
time exponential signal 𝒆𝒋𝝎𝟎 𝒏 .
◼ The following properties were found regarding the continuous-time exponential
signal 𝑒 𝑗𝜔0𝑡 :
1) The larger the magnitude of 𝜔0 , the higher is the rate of oscillations in the signal;
2) 𝑒 𝑗𝜔0 𝑡 is periodic for any value of 𝜔0 .
◼ Let us now see how these properties are different in the discrete-time case:
◼ To see the difference for the first property, consider the discrete-time complex
exponential:

j (0 + 2 ) n j0 n j 2 n j0 n

e =e e e  e j 2 n = cos ( 2 n ) + j sin ( 2 n )

➢ This shows that the exponential at 𝝎𝟎 + 𝟐𝝅 is the same as that at frequency 𝝎𝟎 . This is
very different when compared with the continuous-time exponential case, in which the
signals 𝑒 𝑗𝜔0 𝑡 are all distinct for distinct values of 𝝎𝟎 . 28
 Lecture-17-18 Revision-Chap#01

Some Basic Signals

◼ Periodicity Property of Discrete-Time Complex Exponential
◼ In discrete-time, these signals are not distinct.
➢ In fact, the signal with frequency 𝝎𝟎 is identical to signals with frequencies 𝝎𝟎 ± 𝟐𝝅 ,
𝝎𝟎 ± 𝟒𝝅, and so on……..
➢ Therefore, in considering discrete-time complex exponentials, we need only consider a
frequency interval of size 2π.
➢ The most used 2π intervals are 𝟎 ≤ 𝝎𝟎 ≤ 𝟐𝜋 or the interval −𝝅 ≤ 𝝎𝟎 ≤ 𝝅.

◼ Due to equation 𝒆𝒋𝝎𝟎 𝒏 , as 𝝎𝟎 is gradually increased, the rate of oscillations in the

discrete-time signal does not keep on increasing.
➢ If 𝜔0 is increased from 0 𝑡𝑜 2𝜋, the rate of oscillations first increases and then decreases.
This is shown in the Fig (Next Slide).
➢ Note that for 𝜔0 = 𝜋 or for any odd multiple of 𝜋,

= (e ) = ( −1)
j n j n n
e  e j = cos ( n ) + j sin ( n )

➢ It shows that the signal oscillates rapidly, changing sign at each point in time.
29
 Digital Image Processing, 3rd ed.
Lecture-17-18 Revision-Chap#01

Gonzalez & Woods

www.ImageProcessingPlace.c
om
Chapter 10
Image Segmentation

© 1992–2008 R. C. Gonzalez & R. E. Woods

 Lecture-17-18 Revision-Chap#01

3
1
Some Basic Signals
◼ Periodicity Property of Discrete-Time Complex Exponential
◼ It is also considerable value in discrete-time signal and system analysis to consider
sets of harmonically related periodic exponentials – i.e., periodic exponentials with a
common period N.
 2 
jk  n
k  n  = e  N 
, k = 0, 1, 2,
◼ In the continuous-time case, all of the harmonically related complex exponentials
2𝜋
𝑗𝑘 𝑡
𝑒 𝑇 , are distinct. However, this is not the case in discrete time. Specifically,
 2 
j ( k + N ) n
k + N  n  = e  N 
,
 2 
jk  n
k + N  n  = e  N 
e j 2 n = k  n 

◼ This implies that there are only N distinct periodic exponentials. For example,
 2   4   2 
j n j n j ( N −1) n
0  n  = 1, 1  n  = e  N 
, 2  n  = e  N 
, , N −1  n  = e  N 

are all distinct, and any other is identical to one of these (e.g., 𝜙0 𝑛 = 𝜙N 𝑛 ).
 Digital Image Processing, 3rd ed.
Lecture-17-18 Revision-Chap#01

Gonzalez & Woods

www.ImageProcessingPlace.c
Comparison of the signals
om 𝒆𝒋𝝎𝟎 𝒕 and 𝒆𝒋𝝎𝟎 𝒏
Chapter 10
𝒋𝝎𝟎 𝒕
𝒆 Image Segmentation 𝒆𝒋𝝎𝟎 𝒏

Identical signals for values of 𝜔0 separated

Distinct signals for distinct values of 𝜔0 .
by multiples of 2𝜋.
2𝜋
Periodic only if 𝜔0 = 𝑚 for some
Periodic for any choice of 𝜔0 . 𝑁
integer 𝑁 > 0 and 𝑚.
Fundamental frequency 𝜔0 . Fundamental frequency 𝜔0 /𝑚.

Fundamental period Fundamental period

𝜔0 = 0: undefined 𝜔0 = 0: undefined
2𝜋 2𝜋
𝜔0 ≠ 0: 𝜔0 ≠ 0: 𝑚
𝜔0 𝜔0

Point to Ponder
The low-frequency (that is, slowly varying) discrete-time exponentials have values of 𝜔0 near
0, 2𝜋, and any other even multiple of 𝜋, while the high frequencies (corresponding to rapid
variations) are located near 𝜔0 = ±𝜋 and other odd multiples of ±𝜋 .

© 1992–2008 R. C. Gonzalez & R. E. Woods

 Digital Image Processing, 3rd ed.
Lecture-17-18 Revision-Chap#01

Gonzalez & Woods

How to find Time period ofwww.ImageProcessingPlace.c
Discrete om
Time Exponential signal?
Chapter 10
Image Segmentation

e
j0 ( n + N )
= e j0 n  Periodicity 
e j0 n .e j0 N = e j0 n ,
e j0 N = 1,
j ( 2 m )
e j0 N = e ,
0 N = 2 m,
 2 
N = m 
 0 

‘m’ is the smallest integer that makes “N” an integer.

© 1992–2008 R. C. Gonzalez & R. E. Woods

 Digital Image Processing, 3rd ed.
Lecture-17-18 Revision-Chap#01

Gonzalez & Woods

How to find period of Continuous Time sinusoidal periodic signal?
www.ImageProcessingPlace.c
om
Chapter 10
➢ Example-01 ➢ Example-03
Image Segmentation x ( t ) = A cos (0t +  )


2

x ( t ) = 3cos  4t +  x (t ) = e
j ( t −1)
T0 =
 3 0
0 = 4 0 = 
2  2
T0 =  T0 = 2
4 2 

➢ Example-02 ➢ Example-04

 
x ( t ) = cos ( 2t ) 
2
x ( t ) = 10sin  20 t +  + 10
 3
1 + cos ( 4t )
x (t ) = 0 = 20
2
2 1
1 1
x ( t ) = + cos ( 4t ) T0 = 
2 2 20 10
0 = 4
2 
T0 = 
4 2

© 1992–2008 R. C. Gonzalez & R. E. Woods

 Digital Image Processing, 3rd ed.
Lecture-17-18 Revision-Chap#01

Gonzalez & Woods

How to find period of Discrete Time sinusoidal periodic signal?
www.ImageProcessingPlace.c
om
Chapter 10
➢ Example-01 Image
➢ Example-03
Segmentation x  n  = A cos (0 n +  )
 3 
x  n  = cos 
 2 
n
 5  x  n = e
j ( 5 n )
N0 = m  

 0
3 0 = 5
0 =
5  2  2
N = m   m  where ‘m’ is the smallest integer
 2   10   5  5
N = m  3   m   that makes N as an integer.
 5   3 N 0 = 2 at m = 5
N 0 = 10 at m = 3
Periodic

Periodic ➢ Example-04 ➢ Example-05

➢ Example-02  7   5 
x  n  = sin  n x  n  = cos  n
 2   4 
x  n  = sin 3n  7 5
0 = 0 =
0 = 3 2 4
 2  4  2  8
 2  N = m  7   m   N = m  5   m  
x  n  is not N = m    2  7  4  5
 3  N 0 = 4 at m = 7 N 0 = 8 at m = 5
Not periodic Periodic

© 1992–2008 R. C. Gonzalez & R. E. Woods