Principles of Communication-Part 1

Professor Aditya K. Jagannatham

Department of Electrical Engineering
Indian Institute of Technology Kanpur
Module 1
Lecture No 2
Frequency Domain Representation and Introduction to Discrete Fourier series

Keywords: Frequency Domain Representation of a signal, Discrete Fourier Series

(Refer Slide Time: 0:38)

Hello, welcome to another module in this massive open online course, so today let us start
looking at the frequency domain representation of signals, alright. So today we will start
looking at, what is known as and what you must be familiar with, to some extent, the
frequency domain representation of signals. So we are going to look at the frequency domain
representation, alright and this is one of the fundamental tools in available communication.
This is one of the fundamental tools required or one of the fundamental tools used in
communication and this frequency domain representation of the signal is termed as the
spectrum of the signal.
(Refer Slide Time: 2:04)

So this is termed as the frequency domain signal representation also termed as the spectrum
of the signal. Further this spectrum can be of two types, one is we can have a Fourier series
which is discrete in frequency and the other representation is the Fourier transform which is
continuous in frequency. So we have this two different kinds of spectrum or frequency
domain representations, but for the purpose of this particular course, these two are going to
be sufficient First let us start looking at the Fourier series representation, okay.
(Refer Slide Time: 3:55)

So we start by looking at the Fourier series and the Fourier series is defined for a periodic
signal. So let us consider x(t) to be a periodic signal with period capital T, alright.

(Refer Slide Time: 7:59)


j 2 kF0 t 1
Ce k F0 
T and this is termed as
Now the Fourier series of x(t) is given as k  , where
the fundamental frequency of periodic signal x(t), okay. This quantity C k is the kth Fourier
series co-efficient, okay.
(Refer Slide Time: 8:40)

j
. We know that e  cos   j sin  . So this is
j 2 kF0t
Now look at this quantity, e

e j 2 kF0t  cos(2 kF0t )  j sin(2 kF0t ) . This is basically a complex sinusoid. Remember that

we said cosines and sines are sinusoids, so cos(2 kF0t )  j sin(2 kF0t ) is a complex sinusoid.
So what we are saying is that, this summation is basically nothing but a linear combination of
sinusoids at the fundamental frequency F 0 and multiples kF0 of the fundamental frequency,
alright. Each kF0 is a multiple of fundamental frequency F 0, that is the kth multiple of the
fundamental frequency F0.

So what we are saying is that every periodic signal x(t) with fundamental frequency F 0 can be
expressed as a linear combination of an infinite number of sinusoids at either the fundamental
frequency F0 or an integer multiple of the fundamental frequency F 0 that is at kF0, alright. So
this is the intuitive explanation of the discrete Fourier series expansion of a periodic signal
x(t), okay.

The sinusoid that is at an integer multiple of F0 is termed as the kth harmonic and Ck is
j 2 kF0t
basically the co-efficient of the kth harmonic in the linear combination. So e is the kth
harmonic, which is associated with the kth multiple of the fundamental frequency F 0 and Ck is
the discrete Fourier series co-efficient associated with the k th harmonic in this discrete Fourier
series expansion, okay.
(Refer Slide Time: 12:18)

Now, let us look at how to find out the co-efficient C k in the discrete Fourier series, okay. Let
us try to make it Cl since we are already using k as the index of the discrete Fourier series.

T
2
1  j 2 lF0t

T  x(t )e
T
dt  Cl

Now let us consider this quantity 2 boils down to yield the lth co-
efficient Cl of the discrete Fourier series. Now we are going to substitute x(t) in the above

T
2
1   
T  k Ck e j 2 kF0t  e  j 2 lF0t dt
T  

expression, so this becomes, 2 .
(Refer Slide Time: 14:28)

Now interchange the order of summation and integration which takes the summation outside.

T T
2 2

1 j 2 ( k l ) F0t 1 j 2 ( k l ) F0t
C
k 
k
T e dt
T e dt
T T
 
This gives 2 . Now let us look at this integral 2 and try to
infer a very interesting property.
(Refer Slide Time: 16:06)

T
2
1
T  1dt
T
j 2 ( k l ) F0t
Now here if k  l , then e  1 . So if k  l this integral becomes 
2 , which is, so
this is 1.

(Refer Slide Time: 17:08)

 T
2
1 j 2 ( k l ) F0t

T eT
dt

Now if k  l , that is k  l  0 then this integral becomes . So k  l  0 which


2

T
1 e j 2 ( k l ) F0t 2

T j 2 (k  l ) F0 
T
means I can now evaluate this integral as 2 , which is equal to

1 1
T j 2 (k  l ) F0 .

T 1
j 2 ( k l ) F0t
F0   F0T  1
Now e is evaluated at 2 using the property, T . So using this property

T
j 2 ( k l ) F0
, F0T  1 , so this becomes e
j ( k l )
this becomes e
2
 e  j ( k l ) . Now look at this, this is a

sinusoid evaluated at phase  (k  l ) and phase  (k  l ) . The difference between the two

phases is   2 ( k  l ) that is integer multiple of 2 , therefore these two complex sinusoids

are equal hence the difference is 0.

(Refer Slide Time: 19:48)

So phase difference equals 2 (k  l ) which means this quantity here is equal to 0, therefore if

T
2
1 j 2 ( k l ) F0t

T eT
dt  0
k  l then 
2 .
(Refer Slide Time: 20:20)

T
2
1 ifk l
1 j 2 ( k  l ) F0t 
 e dt  
T T 0 ifk l

So what we have shown is that this quantity 2  . So this shows
j 2 kF0t j 2 lF0t
nothing but the fact that these different harmonics e ,e , these are orthogonal that is,

the kth harmonic and lth harmonic, when k  l , at different multiples of F0, that is kF0 and lF0

T
2
1 j 2 kF0t j 2 lF0t 

T e e  dt  0
T

are orthogonal. That is why you have 2 .
(Refer Slide Time: 22:21)

We say that x(t) and y(t) are orthogonal if their dot product or the inner product is equal to 0

T
2
1
 x(t ) y (t )dt  0


T T

that is, 2 . Here we have two different harmonics, that is harmonics at kF 0
and harmonics at lF0, the kth harmonic and the lth harmonic which are orthogonal.

(Refer Slide Time: 23:51)

And we will now use this property to evaluate C k as Cl and that can be simply shown as

T T
2 2
1 
1
 x(t )e  j 2 lF0t dt   Ck e
j 2 ( k l ) F0t
dt  0
T T k  T T
 
follows. We have shown that 2 2 . The second

integral evaluates to 1 if k  l and 0 if k  l . So only the term corresponding to C l is going to

 C  (k  l )  C
k l
remain and therefore this is simply equal to k  .

(Refer Slide Time: 26:30)

Here this is a delta function where  (0)  1 and delta of any other integer N is 0 if N is not

T
2
1  j 2 lF0t

T  x(t )e
T
dt  Cl

equal to 0. So we have integral 2 which is the co-efficient of the lth
harmonic in discrete Fourier series representation of x(t) or the lth co-efficient in the discrete
Fourier series representation of x(t).

So what we have done so far is basically we have started to look at the Fourier domain or
Frequency domain representation of the signal x(t), that is the discrete Fourier series
representation of a periodic signal x(t). We said that any periodic signal x(t), with

1
F0 
fundamental frequency F0, where T , T is the period of the signal can be represented as
the summation or the linear combination of an infinite number of complex sinusoids at the
fundamental frequency F0 and multiples of the fundamental frequency that is kF0, that is the
complex sinusoid corresponding to the fundamental frequency F 0 and its various harmonics,
correct.

And we have also shown how to compute the discrete Fourier series co-efficient C k
corresponding to kth harmonic in this discrete Fourier series representation. So we will stop
this module here and continue with other aspects in the subsequent modules, thank you very
much.