LESSION -2 Date 24/4/2021

Topic :-- Fourier Transform of Digital Signals

Frequency Domain Expression of the Discrete Time Non-Periodic Signal

(SPECTRUM OF DISCRETE TIME NON-PERIODIC SIGNAL USING FOURIER TRANSFORM )

o For Non-Periodic Finite Length Discrete Time signals, the spectrum is evaluated by
performing Fourier Transform of the signal.
o By definition of Fourier Transform on a finite length Discrete Time Signal, x (n), we
obtain a function X (ω), of the Angular Frequency, as follows.

n=k
X ( ω )=∑ x ( n ) . e
−jωn

n= p

o By definition of Fourier Transform on a finite length Discrete Time Signal, x (n), we

obtain a function X (ω), of the Angular Frequency ‘ω‘, as follows.

n=k
X ( ω )=∑ x ( n ) . e
−jωn

n= p

The Fourier Transform operation is denoted by the “Fourier Transform Operator” as follows

Ƒ [ x ( n ) ] =X ( ω )

The plot of the function X (ω) as a function of Angular Frequency ω gives us the
Spectrum of the Non-Periodic Discrete Time Signal. By means of an Inverse Transform
Operation on the function X (ω) the Time Domain representation x (n) can be recovered.
The operation of Inverse Fourier Transform is defined as follows.

π
1
x ( n )=
2π ∫ X (ω ). e jωn d ω
−π

The Inverse Fourier Transform operation canalso be denoted by a Transform Operator as

follows
Ƒ
−1
[ X ( ω ) ]=x (n)
 NUMERICALS
Example 1 :- Evaluate the Fourier Transform of a signal x (n) , which is in the form “Sampled
Pulse” as shown in the figure and sketch the spectrum.

Solution :- Using the Fourier transform formula, we have

n=k
Ƒ [ x ( n ) ] =X ( ω ) ¿ ∑ x ( n ) . e
−jωn

n= p

In the given situation, the function x (n) consist of a series of constant sample values
of “A” from the sample number n = 0 to the sample number n = (L-1)
Changing the limits of summation, and substituting the value of x (n) = A for this set
of sample numbers in the Fourier Transform formula we have

n=(L−1)
X (ω) ¿ ∑ A .e− j ω n
n=0

Rearranging the expression , by taking A common , as follows,

n=(L−1)

∑ ( e− j ω )
n
X (ω) ¿ A …………(A)
n=0

We can now use Geometric Sum Formula to evaluate the expression, using

(e ¿¿− j ω)=a ¿

(Recall the Geometric Sum Formula ,as given by the General form , has a solution as
follows--)
 n=(M −1)

P= B ∑ (a)
n
( Where “B” & “M” are constants )
n=0

Putting B = A & M = L , we will have a = 1 when ω=0 & a ≠ 1 when ω ≠ 0

Using this in the Geometric Sum Formula we have the solution of the Fourier
Transform of the signal as --

X (ω) ¿ A L when ω=0

&
( 1−e− j ω L )
X (ω) ¿ A
( 1−e− j ω )
Rearranging as

[ e− j ω L/ 2( e
]
j ω L/2
−e
−j ω L/2
)
X (ω) ¿ A
e− j ω /2 ( e j ω /2−e− j ω /2 )

X (ω) ¿ A e
− j ω (L−1 )/ 2
[ sin ( ω L/2 )
sin ( ω /2 ) ]
The quantity within the square brackets is the “Sample Sinc Function” which is
defined within a Continuous Frequency frequency range from - π to + π as shown
below.

It is a tedious job to calculate the Fourier Transform of complex signals using the
Fourier Transform Formula given above. However, for simple functions, this is possible.
Fourier transform also has a set of properties. Breaking up a complex function in terms of
the simpler functions and using the properties and using the Fourier Transforms of simpler
functions , it is possible to evaluate the Fourier transform of the complex functions also.

Frequency Response of LTI Systems

 If an LTI System is excited by a simple phasor in the form of
x (n) = e j ω n
Or by a simple Sinusoidal signal such as
x (n) = cos (ω n)
We obtain “The Frequency Response” of the LTI System.
It can be shown that the Frequency Response of the LTI System is nothing bu the Fourier
transform of the Impulse Response Function of the LTI System. Consider an LTI System
having an Impulse Response Function h (n) , being applied with a Simple Phasor as the input
excitation.

The output of the LTI System to any excitation is given by the Convolution Sum Formula as
k=∞
y ( n )= ∑ h ( k ) . x (n−k )
k=−∞

Substituting the given phasor as the input with the amplitude “A” taken outside we have
k=∞
y ( n )= A ∑ h ( k ) . e j ω (n−k )
k=−∞

[∑ ]
k=∞
y ( n )= A h ( k ) .e− j ω k e j ω n
k=−∞

In this expression, the term within the square brackets is the Fourier Transform of
the Impulse Response h (n) of the LTI System. (The Sampling Number variable ‘n’ has been
replaced by a dummy variable ‘k’)

k=∞

∑ h ( k ) . e− j ω k = Ƒ [ h ( k ) ]
k=−∞

The Fourier Transform of a function in Time Domain ‘n’ is a function of Angular Frequency
‘ω’
k=∞
Ƒ [h (k )] ∑ h ( k ) . e− j ω k =¿ H (ω)
= k=−∞
Therefore, the output of the LTI System is
 y ( n )= A [ H (ω) ] e
j ωn

Where the term outside the bracket is nothing but the input signal
A . e j ω n = x (n)
In other words, the output of the LTI System to a Simple Phasor of frequency ‘ω’ is the
product of the input excitation with the Fourier Transform of the Impulse Response
Function evaluated at the signal frequency , as
y (n) = H (ω) . x (n)
Example 2 :- Calculate the response of an LTI System whose Impulse response h (n) and the
excitation x (n) to be applied input are as shown below..

()
n
1
h (n) = u(n) x (n) = A . e j π n/ 2
2
Solution :- By the Frequency Response Formula , we have
y (n) = H (ω) . x (n)
The Frequency Response function H (ω) is to be evaluated at the input signal
frequency. Since the Input signal is a Phasor,
x (n) = A . e j π n/ 2
The Signal Frequency is
ω = π /2
Since the Impulse Response Function appears with the “Unitary Function” “u (n)” ,
the function h (n) exists only for the positive values of n. Hence the Fourier
Transform of h (n) will be given by

()
k=∞
1 n −jωn
H (ω) = Ƒ [ h ( n) ] = ∑
k=0 2
.e

( )
k=∞ n
1 − jω
H (ω) = ∑ 2
e
k=0

Taking ∞ = (∞ -1 ) and
1 −jω
a = 2
e

We have
k=(∞−1)

H (ω) = ∑ (a )
n

k=0

Using Geometric Sum Formula, the Summation on the RHS of the expression is
( 1−a ∞ )
H ( ω )=¿
(1−a )
In this, when we substitute

( ) →0
∞
1 −jω 1
a = 2 e we will get 2
Hence we get the expression for Frequency Response , as
 1
H (ω ) =
(
1 −jω
1− e
2 )
The Frequency Response is to be calculated at the signal frequency
ω = π /2
Substituting this value of ω in the expression above , we get

{ 1 −j(2 )
}
π
H (ω)at ω=π / 2=
1− () 2
e

In this expression
−j ( π2 ) = - j ( the Imaginary Number )
e
1
∴ H (ω)at ω=π / 2 =
{ ( )}
1+ j
1
2
Converting into the Polar Form,
1∠ 0
H (ω)at ω=π / 2=¿ √ 5
∠ 26.6 0
2

2 − j 26.6 0

∴ H ( ω )=¿ e
√5
Since the output of the LTI system will be

y (n) = H (ω) . x (n) ( Where H (ω ) is evaluated at the signal frequency)

We get the output as

y (n) = A
[ 2 − j 26.6
√5
e
0

]
. e j π n/ 2

2
y (n) = √5
A¿

In this , the quantity {2} over {√5 “ is the “Gain” of the LTI System and the angular
quantity 26.60 is the “Phase Shift” imparted by the LTI System on the Input Signal.

Hence the Frequency Response of the LTI System to the applied input phasor is
another phasor at the same frequency, but the amplitude is SCALED BY A GAIN and
the signal UNDERGOES A PHASE SHIFT .
 If a signal at a different frequency is applied, the Frequency Response Function H (ω)
will be evaluated at that frequency. Therefore the amount of gain and phase shift
also will be of different values.

Example 3 :- Consider another signal as input to the same LTI System as in Example – 5.

()
n
1
h (n) = 2
u(n) x (n) = 2 cos 2 π n/3

Solution :-
By the Frequency Response Formula

y (n) = H (ω) . x (n)

THE LTI SYSTEM HAS THE IMPULSE RESPONSE FUNCTION

()
n
1
h (n) = 2 u(n)

THE FREQUENCY RESPONSE FUNCTION H (ω) IS TO BE EVALUATED AT THE INPUT

SIGNAL FREQUENCY.
SINCE THE GIVEN SIGNAL IS

x (n) = 2 cos 2 π n/3

ω = 2 π /3

IN THE PREVIOUS EXAMPLE WE HAD

[( ) ] = ∑ ( )
n k=∞
1 1 n −jωn
H (ω) = Ƒ u(n) .e
2 k=0 2

EVALUATING WITH GEOMETRIC SUM FORMULA WE HAD

1
H (ω) =
(
1 −jω
1− e
2 )
THE FREQUENCY RESPONSE IS TO BE CALCULATED AT THE SIGNAL
FREQUENCY
 ω = 2π /3

{ 1 − j (2 3 )
}
π
H (ω)at ω=2 π / 3=
1− ()2
e

SINCE WE HAVE
( π3 )
= -0.5 – j 0.866
−j 2
e
SUBSTITUTING WE GET

1
∴ H (ω)at ω=2 π / 3 =
{ 2 ) (−0.5 – j0.866 ) }
1−( 1

1
∴ H (ω)at ω=2 π / 3 =
{ 1.25+ j 0.433 }

0.76 ∠−19.1
CONVERTING INTO THE POLAR FORM,
1∠ 0
H (ω)at ω=π / 2=¿ 0 =
0
1.32∠ 19.1

0.76e− j−19.1
0
∴ H ( ω )=¿

SINCE THE OUTPUT OF THE LTI SYSTEM IS

y (n) = H (ω) . x (n) (with H (ω ) evaluated at the signal frequency)

y (n) = [ 0.76 e− j−19.1 ] . 2 cos 2 π n / 3

0

y (n) = (2 x 0.76) cos ( 2 π3 n −19.1 )

0

y (n) = 1.52 cos ( 2 π3 n −19.1 ) 0

GAIN OF THE SYSTEM IS |H ( ω )| , WHICH IS THE MAGNITUDE OF THE

FUNCTION H ( ω ) . HENCE “GAIN” = 0.76.
 PHASE SHIFT IMPARTED BY THE SYSTEM IS THE ANGLE OF THE
FUNCTION H ( ω ) . HENCE “PHASE SHIFT” = −19.10