Md.

Anawar Ibrahim (Ananta)

Reg: 3098; EEE-03

➢ Signal: A signal is a physical quantity that carries information about a

phenomenon.
A signal consists of a function of one or two variables which may vary
with time.
Signals can be classified into different categories based on their
characteristics:

1. Continuous time signal and Discrete time signal

2. Periodic signal and Aperiodic signal
3. Even signal and Odd signal
4. Deterministic signal and Random signal
5. Energy signal and Power signal

➢ Digital Signal Processing: Digital Signal Processing (DSP) is a field

that deals with the processing, analysis and manipulation of digital
signals.
Digital Signal Processing works with the help of discrete time signals
𝒙[𝒏]

Fig: Digital Signal Processing Overview

1
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

➢ Discrete-time signals (DT signals):

DT signals are that type of signals which exist only at specific points of
time intervals, typically intervals are equally spaced.
They are typically represented by sequences of numbers, where each
number represents the value of the signal at a particular time instant.
DT signals are often used in digital signal processing (DSP)
applications.

➢ Various Forms of Discrete Time Signals:

1. Sequence Form:
𝐱(𝐧) = {−𝟐, 𝟑, 𝟎, −𝟏, 𝟐, 𝟑, 𝟏}

2. Tabular Form:
Position, −3 −2 −1 0 1 2 3
n
𝐱(𝐧) −2 3 0 −1 2 3 1

3. Graphical Form:

➢ Unit Impulse Signal:

2
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

➢ Unit Step Signal:

➢ Unit Ramp Signal:

➢ Time Shifting of Discrete Time Signal:

3
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

➢ Time Scaling of Discrete Time Signal:

➢ Time Reflection of Discrete Time Signal:

➢ Even and Odd Discrete Time Signal:

From the definition of odd and even signal,
𝐱 𝐞 (𝐧) = 𝐱𝐞 (−𝐧) … … . (𝒊)
𝐱𝐨 (𝐧) = −𝐱𝐨 (−𝐧) … … . (𝒊𝒊)
Now,
𝐱(𝐧) = 𝐱𝐞 (𝐧) + 𝐱𝐨 (𝐧) … … . (𝒊𝒊𝒊)

𝐱(𝐧)+𝐱(−𝐧) 𝐱(𝐧)−𝐱(−𝐧)
Even signal, 𝐱𝐞 (𝐧) = ; Odd signal, 𝐱 𝐨 (𝐧) =
𝟐 𝟐

4
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

➢ LTI System: LTI means Linear Time Invariant.

LTI System has two properties of a system. They are:
1. System has to be Linear
2. System has to be Time Invariant

▪ Linear System:
Linear system has two properties:
1. Additive Property
𝒙𝟏 [𝒏] + 𝒙𝟐 [𝒏] → 𝒚𝟏 [𝒏] + 𝒚𝟐 [𝒏]

2. Homogeneity Property
𝒂 𝒙𝟏 [𝒏] + 𝒃 𝒙𝟐 [𝒏] → 𝒂 𝒚𝟏 [𝒏] + 𝒃 𝒚𝟐 [𝒏]

➢ Is 𝒚[𝒏] = 𝒙[𝒏] − 𝟑𝒙[𝒏 − 𝟏] linear or non-linear?

 Given that,
𝒚[𝒏] = 𝒙[𝒏] − 𝟑𝒙[𝒏 − 𝟏]
Here,
𝑦1 [𝑛] = 𝑥1 [𝑛] − 3𝑥1 [𝑛 − 1]

𝑦2 [𝑛] = 𝑥2 [𝑛] − 3𝑥2 [𝑛 − 1]

Now,
𝒙𝟑 [𝒏] = 𝒂𝒙𝟏 [𝒏] + 𝒃𝒙𝟐 [𝒏]

𝑦3 [𝑛] = 𝑥3 [𝑛] − 3𝑥3 [𝑛 − 1]

𝑦3 [𝑛] = [𝑎𝑥1 (𝑛) + 𝑏𝑥2 (𝑛)] − 3[𝑎𝑥1 (𝑛 − 1) + 𝑏𝑥2 (𝑛 − 1)]

𝑦3 [𝑛] = [𝑎𝑦1 (𝑛) + 𝑏𝑦2 (𝑛)] − 3[𝑎𝑦1 (𝑛 − 1) + 𝑏𝑦2 (𝑛 − 1)]

𝒚𝟑 [𝒏] = 𝒚𝟑 [𝒏] − 𝟑𝒚𝟑 [𝒏 − 𝟏]

This System is a linear system

5
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

➢ Is 𝒚[𝒏] = 𝟑𝒙[𝒏 − 𝟏] linear or non-linear?

 Given that,
𝒚[𝒏] = 𝟑𝒙[𝒏 − 𝟏]
Here,
𝑦1 [𝑛] = 3𝑥1 [𝑛 − 1]

𝑦2 [𝑛] = 3𝑥2 [𝑛 − 1]
Now,
𝒙𝟑 [𝒏] = 𝒂𝒙𝟏 [𝒏] + 𝒃𝒙𝟐 [𝒏]

𝑦3 [𝑛] = 3𝑥3 [𝑛 − 1]

𝑦3 [𝑛] = 3[𝑎𝑥1 (𝑛 − 1) + 𝑏𝑥2 (𝑛 − 1)]

𝑦3 [𝑛] = 3[𝑎𝑦1 (𝑛 − 1) + 𝑏𝑦2 (𝑛 − 1)]

𝒚𝟑 [𝒏] = 𝟑𝒚𝟑 [𝒏 − 𝟏]

This System is a linear system

▪ Time-invariant system: A system is said to be time-invariant if the

input signal is delayed or advanced by any factor that leads the
results in the same delay or advancement in the output signal by the
same time factor. In other words, in this system output signal’s time
delay or advancement will be same as the input signal’s time delay
or advancement.

6
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

𝒙[𝒏 − 𝒏𝒐 ] → 𝒚[𝒏 − 𝒏𝒐 ]

𝒙[𝒏 + 𝒏𝒐 ] → 𝒚[𝒏 + 𝒏𝒐 ]

➢ Is 𝒚[𝒏] = 𝒙[𝒏] − 𝟑𝒙[𝒏 − 𝟏] time variant or invariant?

 Given that,
𝒚[𝒏] = 𝒙[𝒏] − 𝟑𝒙[𝒏 − 𝟏]
Here,
𝑦1 [𝑛] = 𝑥1 [𝑛] − 3𝑥1 [𝑛 − 1]

𝑦2 [𝑛] = 𝑥2 [𝑛] − 3𝑥2 [𝑛 − 1]

Now,
𝒙𝟐 [𝒏] = 𝒙𝟏 [𝒏 − 𝒏𝒐 ]

𝑦2 [𝑛] = 𝑥1 (𝑛 − 𝑛𝑜 ) − 3𝑥1 [(𝑛 − 𝑛𝑜 ) − 1]

𝒚𝟏 [𝒏 − 𝒏𝒐 ] = 𝒙𝟏 [𝒏 − 𝒏𝒐 ] − 𝟑𝒙𝟏 [(𝒏 − 𝒏𝒐 ) − 𝟏]

This System is a Time invariant system

➢ Is 𝒚[𝒏] = 𝒙[𝒏𝟐 ] time variant or invariant?

 Given that,
𝒚[𝒏] = 𝒙[𝒏𝟐 ]
Here,
𝑦1 [𝑛] = 𝑥1 [𝑛2 ]

𝑦2 [𝑛] = 𝑥2 [𝑛2 ]
Now,
𝒙𝟐 [𝒏] = 𝒙𝟏 [𝒏 − 𝒏𝒐 ]
7
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

𝑦2 [𝑛] = 𝑥1 [(𝑛 − 𝑛𝑜 )2 ]

𝑦2 [𝑛] = 𝑥1 [𝑛2 − 2𝑛𝑛𝑜 + 𝑛𝑜2 ]

𝒚𝟏 [𝒏 − 𝒏𝒐 ] = 𝒙𝟏 [𝒏𝟐 − 𝟐𝒏𝒏𝒐 + 𝒏𝟐𝒐 ]

This System is a Time variant system

𝜹[𝒏 + 𝟐] − 𝜹[𝒏 + 𝟏] + 𝜹[𝒏] + 𝟐𝜹[𝒏 − 𝟐]

8
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

Here,
𝒙[𝒏] = Input
𝒚[𝒏] = Output
𝒉[𝒏] = 𝜹[𝒏] = Impulse Response
We know,
𝒚[𝒏]
𝒉[ 𝒏] =
𝒙[𝒏]
 𝒚[𝒏] = 𝒙[𝒏] ∗ 𝒉[𝒏]

∴ 𝒚[𝒏] = 𝒙[𝒏] ∗ 𝜹[𝒏]

We know that, impulse signal will be present only at 𝒏 = 𝟎. So, we can
write 𝜹[𝒏] as 𝜹[𝟎]
𝒚[𝒏] = 𝒙[𝒏] ∗ 𝜹[𝟎]
∴ 𝒚[𝒏] = 𝒙[𝒏] ∗ 𝜹[𝒏 − 𝒌]
Here,
𝒌 = Shifting Value
To make 𝜹[𝟎]; 𝒌 = 𝒏
For LTI System,
∞

∴ 𝒚[𝒏] = ∑ 𝒙[𝒌] ∗ 𝜹[𝒏 − 𝒌]

𝒌=−∞

9
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

For this,
∴ 𝒚[𝒏] = 𝒙(−𝟐)𝜹(𝒏 + 𝟐) + 𝒙(−𝟏)𝜹(𝒏 + 𝟏) + 𝒙(𝟎)𝜹(𝒏)
+ 𝒙(𝟏)𝜹(𝒏 − 𝟏) + 𝒙(𝟐)𝜹(𝒏 − 𝟐)

➢ Z-Transform:
The Z-transform is a mathematical tool used for analyzing discrete-time
signals. The Z-transform converts a discrete-time signal into a complex
frequency domain.
For a discrete-time signal 𝑥(𝑛); The Z-transform will be,
∞

𝑿(𝒛) = ∑ 𝒙(𝒏). 𝒛−𝒏

𝒏=−∞

Where:
𝒛 = A complex number, often expressed in polar form as 𝒛 = 𝒓𝒆𝒋𝝎
𝒓 = Magnitude
𝝎 = The angle (𝑓𝑟𝑒𝑞𝑢𝑒𝑛𝑐𝑦 𝑐𝑜𝑚𝑝𝑜𝑛𝑒𝑛𝑡)
𝒏 = Integer representing the time index

10
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

➢ Properties of Z-Transform:
1. Linearity
2. Time shifting
3. Time Scaling
4. Time reversal
5. Multiplication
6. Conjugation and Conjugate Symmetry
7. Parseval’s Relation
8. Difference in Time
9. Differentiation in Frequency
10. Convolution

➢ Region of Convergence (ROC):

The ROC is the set of values of 𝑧 in the complex plane for which the Z-
transform converges to a finite value. The ROC is crucial because it
determines the stability and causality of the system.

➢ Properties of ROC:
1. ROC Cannot Include Any Poles:
The ROC of a Z-transform does not include any poles (values of z where
𝑿(𝒛) becomes infinite). This is because, at the poles, the Z-transform
does not converge.
2. ROC is a Connected Region:
The ROC is always a connected region in the complex z-plane. It cannot
be split into separate, disjoint regions.

11
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

3. ROC for Different Types of Sequences:

i. Right-Sided Sequences (Causal Sequences): For Right-Sided
Sequences, the ROC is the region outside a circle of radius 𝑟.
(|𝑧| > 𝑟).
ii. Left-Sided Sequences (Anti-Causal Sequences): For Left-Sided
Sequences the ROC is the region inside a circle of radius 𝑟.
(|𝑧| < 𝑟).
iii. Two-Sided Sequences (Non-Causal Sequences): For two sided
sequences, the ROC is a ring-shaped region between two circles
of radius 𝑟1 and 𝑟2 . (𝑟1 < |𝑧| < 𝑟2 ).

• Find the Z-transform and ROC of 𝒙(𝒏) = 𝒖(𝒏)

 Given that,
𝒙(𝒏) = 𝒖(𝒏)
We know that,
∞

𝑿(𝑧) = ∑ 𝑥 (𝑛). 𝑧 −𝑛
𝑛=−∞
∞

𝑋(𝑧) = ∑ 𝑢(𝑛). 𝑧 −𝑛
𝑛=0
∞

𝑋(𝑧) = ∑ 1. 𝑧 −𝑛
𝑛=0
∞

𝑋(𝑧) = ∑(𝑧 −1 )𝑛
𝑛=0

1 ∞
𝑋 (𝑧 ) = ∑ 𝒂𝒏 =
𝟏
1 − 𝑧 −1 𝟏−𝒂
𝒏=𝟎
1
𝑋 (𝑧 ) =
1
1−
𝑧

12
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

1
𝑋 (𝑧 ) =
𝑧−1
𝑧
𝒛
∴ 𝑿(𝒛) =
𝒛−𝟏
Here,
𝒓 = 𝟏; 𝒛 > 𝟏; Outside of the circle

• Find the Z-transform and ROC of 𝒙(𝒏) = 𝒂𝒏 𝒖(𝒏)

 Given that,
𝒙(𝒏) = 𝒂𝒏 𝒖(𝒏)
We know that,
∞

𝑿(𝑧) = ∑ 𝑥 (𝑛). 𝑧 −𝑛
𝑛=−∞
∞

𝑋(𝑧) = ∑ 𝑎𝑛 𝑢(𝑛). 𝑧 −𝑛
𝑛=0
∞

𝑋(𝑧) = ∑ 𝑎𝑛 . 𝑧 −𝑛
𝑛=0
∞

𝑋(𝑧) = ∑(𝑎. 𝑧 −1 )𝑛
𝑛=0

1
𝑋 (𝑧 ) = ∞
𝟏
1 − 𝑎. 𝑧 −1 ∑ 𝒂𝒏 =
𝟏−𝒂
1 𝒏=𝟎
𝑋 (𝑧 ) = 𝑎
1−
𝑧

13
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

1
𝑋 (𝑧 ) = 𝑧 − 𝑎
𝑧
𝒛
∴ 𝑿(𝒛) =
𝒛−𝒂
Here,
𝒓 = 𝒂; 𝒛 > 𝒂; Outside of the circle

• Find the Z-transform and ROC of 𝒙(𝒏) = −𝒂𝒏 𝒖(−𝒏 − 𝟏)

 Given that,
𝒙(𝒏) = −𝒂𝒏 𝒖(−𝒏 − 𝟏)
We know that,
∞

𝑿(𝑧) = ∑ 𝑥 (𝑛). 𝑧 −𝑛
𝑛=−∞
−1

𝑋(𝑧) = ∑ −𝑎𝑛 𝑢(−𝑛 − 1). 𝑧 −𝑛

𝑛=−∞
−1

𝑋(𝑧) = ∑ −𝑎𝑛 . 𝑧 −𝑛
𝑛=−∞
∞

𝑋(𝑧) = − ∑ 𝑎−𝑛 . 𝑧 𝑛
𝑛=1
∞

𝑋(𝑧) = − ∑(a−1 . 𝑧)𝑛

𝑛=1
∞
(a−1 . 𝑧)1 𝒂𝒌
𝑋 (𝑧 ) = − ∑ 𝒂𝒏 =
1 − 𝑎−1 . z 𝟏−𝒂
𝒏=𝐤

14
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03
𝑧
𝑋 (𝑧 ) = − 𝑎 𝑧
1−
𝑎
𝑧
𝑎
𝑋 (𝑧 ) = − 𝑎 − 𝑧
𝑎
−𝒛
∴ 𝑿(𝒛) =
𝒂−𝒛
Here,
𝒓 = 𝒂; 𝒛 < 𝒂; Inside of the circle

• Find the Z-transform and ROC of 𝒙(𝒏) = 𝒂𝒏 𝒖(𝒏) − 𝒃𝒏 𝒖(𝒏)

 Given that,
𝒙(𝒏) = 𝒂𝒏 𝒖(𝒏) − 𝒃𝒏 𝒖(𝒏)
We know that,
∞

𝑿(𝑧) = ∑ 𝑥 (𝑛). 𝑧 −𝑛
𝑛=−∞
∞

𝑋(𝑧) = ∑[𝑎𝑛 𝑢(𝑛) − 𝑏𝑛 𝑢(𝑛)]. 𝑧 −𝑛

𝑛=0
∞ ∞

𝑋(𝑧) = ∑ 𝑎𝑛 𝑢(𝑛). 𝑧 −𝑛 − ∑ 𝑏𝑛 𝑢(𝑛). 𝑧 −𝑛

𝑛=0 𝑛=0
∞ ∞

𝑋(𝑧) = ∑ 𝑎𝑛 . 𝑧 −𝑛 − ∑ 𝑏𝑛 . 𝑧 −𝑛
𝑛=0 𝑛=0
∞ ∞ ∞
−1 )𝑛 −1 )𝑛 𝟏
𝑋(𝑧) = ∑(𝑎. 𝑧 − ∑ (𝑏. 𝑧 ∑ 𝒂𝒏 =
𝟏−𝒂
𝒏=𝟎
𝑛=0 𝑛=0

15
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

1 1
𝑋 (𝑧 ) = −
1 − 𝑎. 𝑧 −1 1 − 𝑏. 𝑧 −1
1 1
𝑋 (𝑧 ) = 𝑎 −
1− 𝑏
𝑧 1 −
z
1 1
𝑋 (𝑧 ) = 𝑧 − 𝑎 −
𝑧−𝑏
𝑧 z
𝒛 𝒛
∴ 𝑿(𝒛) = −
𝒛−𝒂 𝒛−𝒃
Here,
𝒓𝟏 = 𝒂; 𝒓𝟐 = 𝒃; 𝒛 > 𝒂, 𝒛 > 𝒃; Outside both of the circles

• Find the Z-transform and ROC of 𝒙(𝒏) = −𝒂𝒏 𝒖(−𝒏 − 𝟏) −

𝒃𝒏 𝒖(−𝒏 − 𝟏)
 Given that,
𝒙(𝒏) = −𝒂𝒏 𝒖(−𝒏 − 𝟏) − 𝒃𝒏 𝒖(−𝒏 − 𝟏)
We know that,
∞

𝑿(𝑧) = ∑ 𝑥 (𝑛). 𝑧 −𝑛
𝑛=−∞
−1

𝑋(𝑧) = ∑ [−𝑎𝑛 𝑢(−𝑛 − 1)−𝑏𝑛 𝑢(−𝑛 − 1)]. 𝑧 −𝑛

𝑛=−∞
−1 −1

𝑋(𝑧) = − ∑ 𝑎𝑛 𝑢(−𝑛 − 1). 𝑧 −𝑛 − ∑ 𝑏𝑛 𝑢(−𝑛 − 1). 𝑧 −𝑛

𝑛=−∞ 𝑛=−∞
−1 −1

𝑋(𝑧) = − ∑ 𝑎𝑛 . 𝑧 −𝑛 − ∑ 𝑏𝑛 . 𝑧 −𝑛
𝑛=−∞ 𝑛=−∞

16
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

∞ ∞

𝑋(𝑧) = − ∑ 𝑎−𝑛 . 𝑧 𝑛 − ∑ 𝑏 −𝑛 . 𝑧 𝑛
𝑛=1 𝑛=1
∞ ∞

𝑋(𝑧) = − ∑(a−1 . 𝑧)𝑛 − ∑(b−1 . 𝑧)𝑛

𝑛=1 𝑛=1
∞
(a−1 . 𝑧)1 (b−1 . 𝑧)1 ∑𝒂 =
𝒂𝒌
𝒏
𝑋 (𝑧 ) = − − 𝟏−𝒂
1 − 𝑎−1 . z 1 − 𝑏−1 . z 𝒏=𝐤

𝑧 𝑧
𝑋 (𝑧 ) = − 𝑎 𝑧 − 𝑏 𝑧
1− 1−
𝑎 𝑏
𝑧 𝑧
𝑋 (𝑧 ) = − 𝑎 − 𝑎 − 𝑏
𝑧 𝑎−𝑧
𝑎 𝑏
−𝒛 −𝒛
∴ 𝑿(𝒛) = +
𝒂−𝒛 𝒃−𝒛
Here,
𝒓𝟏 = 𝒂; 𝒓𝟐 = 𝒃; 𝒂 > 𝒛, 𝒃 > 𝒛; Inside both of the circles

17
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

• Find the Z-transform of 𝒙(𝒏) = 𝐬𝐢𝐧 𝝎𝒏 𝒖(𝒏)

 Given that,
𝒙(𝒏) = 𝐬𝐢𝐧 𝝎𝒏 𝒖(𝒏)
We know that,
∞

𝑿(𝑧) = ∑ 𝑥 (𝑛). 𝑧 −𝑛
𝑛=−∞
𝒆𝒋𝝎 − 𝒆−𝒋𝝎
∞ 𝐬𝐢𝐧 𝝎 =
𝟐𝒋
𝑋(𝑧) = ∑ sin 𝜔𝑛 𝑢(𝑛). 𝑧 −𝑛
𝒆𝒋𝝎 + 𝒆−𝒋𝝎
𝑛=0 𝐜𝐨𝐬 𝝎 =
𝟐
∞

𝑋(𝑧) = ∑ sin 𝜔𝑛 . 𝑧 −𝑛
𝑛=0
∞
𝑒 𝑗𝜔𝑛 − 𝑒 −𝑗𝜔𝑛
𝑋 (𝑧 ) = ∑ [ ] . 𝑧 −𝑛
2𝑗
𝑛=0
∞
1
𝑋(𝑧) = [∑[𝑒 𝑗𝜔𝑛 − 𝑒 −𝑗𝜔𝑛 ]. 𝑧 −𝑛 ]
2𝑗
𝑛=0

∞ ∞
1
𝑋(𝑧) = [∑ 𝑒 𝑗𝜔𝑛 𝑧 −𝑛 − ∑ 𝑒 −𝑗𝜔𝑛 . 𝑧 −𝑛 ]
2𝑗
𝑛=0 𝑛=0

∞ ∞
1
𝑋(𝑧) = [∑(𝑒 𝑗𝜔 𝑧 −1 )𝑛 − ∑(𝑒 −𝑗𝜔 . 𝑧 −1 )𝑛 ]
2𝑗
𝑛=0 𝑛=0

∞
1 1 1 𝟏
𝑋 (𝑧 ) = [ − ] ∑ 𝒂𝒏 =
2𝑗 1 − 𝑒 𝑗𝜔 𝑧 −1 1 − 𝑒 −𝑗𝜔 𝑧 −1 𝟏−𝒂
𝒏=𝟎

18
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

1 1 − 𝑒 −𝑗𝜔 𝑧 −1 − 1 + 𝑒 𝑗𝜔 𝑧 −1
𝑋 (𝑧 ) = [ ]
2𝑗 (1 − 𝑒 𝑗𝜔 𝑧 −1 )(1 − 𝑒 −𝑗𝜔 𝑧 −1 )

1 𝑒 𝑗𝜔 𝑧 −1 − 𝑒 −𝑗𝜔 𝑧 −1
𝑋 (𝑧 ) = [ ]
2𝑗 1 − 𝑒 𝑗𝜔 𝑧 −1 − 𝑒 −𝑗𝜔 𝑧 −1 + 𝑧 −2

1 (𝑒 𝑗𝜔 − 𝑒 −𝑗𝜔 )𝑧 −1
𝑋 (𝑧 ) = [ ]
2𝑗 1 − (𝑒 𝑗𝜔 + 𝑒 −𝑗𝜔 )𝑧 −1 + 𝑧 −2

𝑒 𝑗𝜔 − 𝑒 −𝑗𝜔 −1
1 2𝑗 [ 𝑧 ]
2𝑗
𝑋 (𝑧 ) = [ ]
2𝑗 𝑒 𝑗𝜔 + 𝑒 −𝑗𝜔 −1 −2
1 − 2( 𝑧 )+𝑧
2

𝐬𝐢𝐧 𝝎 𝒛−𝟏
∴ 𝑿(𝒛) =
𝟏 − 𝟐 𝐜𝐨𝐬 𝝎 𝒛−𝟏 + 𝒛−𝟐

• Find the Z-transform of 𝒙(𝒏) = 𝐜𝐨𝐬 𝝎𝒏 𝒖(𝒏)

 Given that,
𝒙(𝒏) = 𝐜𝐨𝐬 𝝎𝒏 𝒖(𝒏)
We know that,
∞

𝑿(𝑧) = ∑ 𝑥 (𝑛). 𝑧 −𝑛 𝐬𝐢𝐧 𝝎 =

𝒆𝒋𝝎 − 𝒆−𝒋𝝎
𝟐𝒋
𝑛=−∞
∞ 𝒆𝒋𝝎 + 𝒆−𝒋𝝎
𝐜𝐨𝐬 𝝎 =
𝟐
𝑋(𝑧) = ∑ cos 𝜔𝑛 𝑢(𝑛). 𝑧 −𝑛
𝑛=0

19
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03
∞

𝑋(𝑧) = ∑ cos 𝜔𝑛 . 𝑧 −𝑛
𝑛=0
∞
𝑒 𝑗𝜔𝑛 + 𝑒 −𝑗𝜔𝑛
𝑋 (𝑧 ) = ∑ [ ] . 𝑧 −𝑛
2
𝑛=0
∞
1
𝑋(𝑧) = [∑[𝑒 𝑗𝜔𝑛 + 𝑒 −𝑗𝜔𝑛 ]. 𝑧 −𝑛 ]
2
𝑛=0

∞ ∞
1
𝑋(𝑧) = [∑ 𝑒 𝑗𝜔𝑛 𝑧 −𝑛 + ∑ 𝑒 −𝑗𝜔𝑛 . 𝑧 −𝑛 ]
2
𝑛=0 𝑛=0

∞ ∞
1
𝑋(𝑧) = [∑(𝑒 𝑗𝜔 𝑧 −1 )𝑛 + ∑(𝑒 −𝑗𝜔 . 𝑧 −1 )𝑛 ]
2
𝑛=0 𝑛=0

∞
1 1 1 𝟏
∑ 𝒂𝒏 =
𝑋 (𝑧 ) = [ + ] 𝟏−𝒂
2 1 − 𝑒 𝑗𝜔 𝑧 −1 1 − 𝑒 −𝑗𝜔 𝑧 −1 𝒏=𝟎

1 1 − 𝑒 −𝑗𝜔 𝑧 −1 + 1 − 𝑒 𝑗𝜔 𝑧 −1
𝑋 (𝑧 ) = [ ]
2 (1 − 𝑒 𝑗𝜔 𝑧 −1 )(1 − 𝑒 −𝑗𝜔 𝑧 −1 )

1 2 − (𝑒 𝑗𝜔 𝑧 −1 + 𝑒 −𝑗𝜔 𝑧 −1 )
𝑋 (𝑧 ) = [ ]
2 1 − 𝑒 𝑗𝜔 𝑧 −1 − 𝑒 −𝑗𝜔 𝑧 −1 + 𝑧 −2

1 2 − (𝑒 𝑗𝜔 + 𝑒 −𝑗𝜔 )𝑧 −1
𝑋 (𝑧 ) = [ ]
2 1 − (𝑒 𝑗𝜔 + 𝑒 −𝑗𝜔 )𝑧 −1 + 𝑧 −2

20
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

𝑒 𝑗𝜔 − 𝑒 −𝑗𝜔 −1
1 2 − 2[ 𝑧 ]
2
𝑋 (𝑧 ) = [ ]
2 𝑒 𝑗𝜔 + 𝑒 −𝑗𝜔 −1 −2
1 − 2( 𝑧 )+𝑧
2
1 2 − 2 cos 𝜔 𝑧 −1
𝑋 (𝑧 ) = [ ]
2 1 − 2(cos 𝜔 𝑧 −1 ) + 𝑧 −2

1 2(1 − cos 𝜔 𝑧 −1 )
𝑋 (𝑧 ) = [ ]
2 1 − 2 cos 𝜔 𝑧 −1 + 𝑧 −2

𝟏 − 𝐜𝐨𝐬 𝝎 𝒛−𝟏
∴ 𝑿(𝒛) =
𝟏 − 𝟐 𝐜𝐨𝐬 𝝎 𝒛−𝟏 + 𝒛−𝟐

➢ Inverse Z Transform:
For Casual System: Right Sequence
∞

∑ 𝑥 (𝑛) 𝑧 −𝑛
0

For Anti Casual System: Left Sequence

−1

∑ 𝑥(−𝑛) 𝑧 𝑛
−∞

21
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

• Find the inverse z-transform using power expansion method for the
given function:
𝟏
𝑿(𝒛) =
𝟏 − 𝟏. 𝟓𝒛−𝟏 + 𝟎. 𝟓𝒛−𝟐
i. ROC |𝒁| > 𝟏
ii. ROC |𝒁| < 𝟏

i. By using power expansion method we get,

Here,
𝑥 (0) = 1
𝑥 (1) = 1.5
𝑥 (2) = 1.75
𝑥 (3) = 1.875
∴ 𝑿(𝒛) = 𝟏 + 𝟏. 𝟓𝒛−𝟏 + 𝟏. 𝟕𝟓𝒛−𝟐 + 𝟏. 𝟖𝟕𝟓𝒛−𝟑 + ⋯ ….

22
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

ii. By using power expansion method we get,

Here,
𝑥 (−1) = 0
𝑥 (−2) = 2
𝑥 (−3) = 6
𝑥 (−4) = 14
∴ 𝑿(𝒛) = 𝟐𝒛𝟐 + 𝟔𝒛𝟑 + 𝟏𝟒𝒛𝟒 + ⋯ ….

23
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

• Find the inverse z-transform using power expansion method for the
given function:
𝟏 + 𝒛−𝟏
𝑿(𝒛) =
𝟏
𝟏 − 𝒛−𝟏
𝟓
𝟏
i. ROC |𝒁| >
𝟓
𝟏
ii. ROC |𝒁| <
𝟓


i. By using power expansion method we get,
1 + 𝑧 −1 1 + 𝑧 −1
𝑋 (𝑧 ) = =
1 −1 1 − 0.2𝑧 −1
1− 𝑧
5

Here,
𝑥 (0) = 1
𝑥 (1) = 1.2
𝑥 (2) = 0.24
𝑥 (3) = 0.048
∴ 𝑿(𝒛) = 𝟏 + 𝟏. 𝟐𝒛−𝟏 + 𝟎. 𝟐𝟒𝒛−𝟐 + 𝟎. 𝟎𝟒𝟖𝒛−𝟑 + ⋯ ….
24
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

ii. By using power expansion method we get,

Here,
𝑥 (−1) = −5
𝑥 (−2) = −30
𝑥 (−3) = 150
∴ 𝑿(𝒛) = −𝟓 − 𝟑𝟎𝒛𝟏 − 𝟏𝟓𝟎𝒛𝟐 + ⋯ ….

• Find the inverse z-transform using partial fraction method for the
given function:
𝟏 + 𝒛−𝟏 + 𝟐𝒛−𝟐
𝑿(𝒛) =
𝟏 𝟏 𝟏
(𝟏 − 𝟐 𝒛−𝟏 ) (𝟏 − 𝟑 𝒛−𝟏 ) (𝟏 − 𝟒 𝒛−𝟏 )
𝟏
i. ROC |𝒁| >
𝟐

 Given that,
𝟏 + 𝒛−𝟏 + 𝟐𝒛−𝟐
𝑿(𝒛) =
𝟏 𝟏 𝟏
(𝟏 − 𝟐 𝒛−𝟏 ) (𝟏 − 𝟑 𝒛−𝟏 ) (𝟏 − 𝟒 𝒛−𝟏 )

25
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

By partial fraction,
𝑨 𝑩 𝑪
𝑿(𝒛) = + +
𝟏 𝟏 𝟏
𝟏 − 𝒛−𝟏 𝟏 − 𝒛−𝟏 𝟏 − 𝒛−𝟏
𝟐 𝟑 𝟒
Here,
For 𝐴, 𝑧 −1 = 2
𝟏 + 𝒛−𝟏 + 𝟐(𝒛−𝟏 )𝟐
𝑨=
𝟏 𝟏
(𝟏 − 𝟑 𝒛−𝟏 ) (𝟏 − 𝟒 𝒛−𝟏 )

1 + 2 + 2 × 22
𝐴=
2 2
(1 − 3) (1 − 4)

∴ 𝑨 = 𝟔𝟔
For 𝐵, 𝑧 −1 = 3
𝟏 + 𝒛−𝟏 + 𝟐(𝒛−𝟏 )𝟐
𝑩=
𝟏 𝟏
(𝟏 − 𝟐 𝒛−𝟏 ) (𝟏 − 𝟒 𝒛−𝟏 )

1 + 3 + 2 × 32
𝐵=
3 3
(1 − 2) (1 − 4)

∴ 𝑩 = −𝟏𝟕𝟔
For 𝐶, 𝑧 −1 = 4
𝟏 + 𝒛−𝟏 + 𝟐(𝒛−𝟏 )𝟐
𝑪=
𝟏 𝟏
(𝟏 − 𝟐 𝒛−𝟏 ) (𝟏 − 𝟑 𝒛−𝟏 )

1 + 4 + 2 × 42
𝐶=
4 4
(1 − ) (1 − )
2 3

∴ 𝑪 = 𝟏𝟏𝟏

26
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

𝟔𝟔 𝟏𝟕𝟔 𝟏𝟏𝟏
𝑿(𝒛) = − +
𝟏 𝟏 𝟏
𝟏 − 𝒛−𝟏 𝟏 − 𝒛−𝟏 𝟏 − 𝒛−𝟏
𝟐 𝟑 𝟒

1 1 1
𝑋(𝑧) = 66 ( ) − 176 ( ) + 111 ( )
1 −1 1 −1 1 −1
1− 𝑧 1− 𝑧 1− 𝑧
2 3 4
𝟏
= 𝒂𝒏 𝒖(𝒏)
𝟏 − 𝒂𝒛−𝟏

𝟏 𝒏 𝟏 𝒏 𝟏 𝒏
∴ 𝒙(𝒏) = 𝟔𝟔 ( ) 𝒖(𝒏) − 𝟏𝟕𝟔 ( ) 𝒖(𝒏) + 𝟏𝟏𝟏 ( ) 𝒖(𝒏)
𝟐 𝟑 𝟒

• Find the inverse z-transform using partial fraction method for the
given function:
𝒛
𝑿(𝒛) =
(𝒛 − 𝟏)(𝒛 − 𝟐)(𝒛 − 𝟑)
i. ROC |𝒁| > 𝟑
ii. ROC 𝟑 > |𝒁| > 𝟐
iii. ROC |𝒁| < 𝟏
 Given that,
𝒛
𝑿(𝒛) =
(𝒛 − 𝟏)(𝒛 − 𝟐)(𝒛 − 𝟑)

27
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

𝑋 (𝑧 ) 1
=
𝑧 (𝑧 − 1)(𝑧 − 2)(𝑧 − 3)
By partial fraction,
𝑿(𝒛) 𝑨 𝑩 𝑪
= + +
𝒛 𝒛−𝟏 𝒛−𝟐 𝒛−𝟑
Here,
For 𝐴, 𝑧 = 1
𝟏 𝟏
𝑨= =
(𝒛 − 𝟐)(𝒛 − 𝟑) (𝟏 − 𝟐)(𝟏 − 𝟑)
1
𝐴=
(−1)(−2)
𝟏
∴𝑨=
𝟐
For 𝐵, 𝑧 = 2
𝟏 𝟏
𝑩= =
(𝒛 − 𝟏)(𝒛 − 𝟑) (𝟐 − 𝟏)(𝟐 − 𝟑)
1
𝐵=
(1)(−1)
∴ 𝑩 = −𝟏
For 𝐶, 𝑧 = 3
𝟏 𝟏
𝑪= =
(𝒛 − 𝟏)(𝒛 − 𝟐) (𝟑 − 𝟏)(𝟑 − 𝟐)
1
𝐶=
(2)(1)
𝟏
∴𝑪=
𝟐
𝟏 𝟏
𝑿(𝒛) 𝟏
= 𝟐 − + 𝟐
𝒛 𝒛−𝟏 𝒛−𝟐 𝒛−𝟑

28
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

1 𝑧 𝑧 1 𝑧
𝑋 (𝑧 ) = ( ) − 1( )+ ( )
2 𝑧−1 𝑧−2 2 𝑧−3

𝒛 𝟏
= = 𝒂𝒏 𝒖(𝒏)
𝒛 − 𝒂 𝟏 − 𝒂𝒛−𝟏

1 1 𝑧 1 𝑧
𝑋 (𝑧 ) = ( ) − 1 ( ) + ( )
2 1 − 𝑧 −1 1 − 2𝑧 −1 2 1 − 3𝑧 −1
1 1
𝑥 (𝑛) = (1)𝑛 𝑢(𝑛) − 1(2)𝑛 𝑢(𝑛) + (3)𝑛 𝑢(𝑛)
2 2
𝟏 𝟏
∴ 𝒙(𝒏) = 𝒖(𝒏) − (𝟐)𝒏 𝒖(𝒏) + (𝟑)𝒏 𝒖(𝒏)
𝟐 𝟐

i. As ROC |𝒁| > 𝟑

𝟏 𝟏
𝒙(𝒏) = 𝒖(𝒏) − (𝟐)𝒏 𝒖(𝒏) + (𝟑)𝒏 𝒖(𝒏)
𝟐 𝟐

ii. As ROC 𝟑 > |𝒁| > 𝟐

𝟏 𝟏
𝒙(𝒏) = 𝒖(𝒏) − (𝟐)𝒏 𝒖(𝒏) − (𝟑)𝒏 𝒖(−𝒏 − 𝟏)
𝟐 𝟐

29
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

iii. As ROC |𝒁| < 𝟏

𝟏 𝟏
𝒙(𝒏) = − 𝒖(−𝒏 − 𝟏) + (𝟐)𝒏 𝒖(−𝒏 − 𝟏) − (𝟑)𝒏 𝒖(−𝒏 − 𝟏)
𝟐 𝟐

➢ Convolution:
Convolution is a mathematical operation used to combine two discrete
signals to produce a third signal. It represents how the shape of one signal
modifies the shape of another.
Given two discrete signals, 𝒙[𝒏] and 𝒉[𝒏], where 𝒙[𝒏] is the input signal
and 𝒉[𝒏] is typically the impulse response of a system, the convolution of
these signals is denoted as,
𝒚[𝒏] = 𝒙[𝒏] ∗ 𝒉[𝒏]
Mathematically, it is expressed as:
∞

𝒚[𝒏] = ∑ 𝒙(𝒌) ∗ 𝒉(𝒏 − 𝒌)

𝒌=−∞

• Steps to Compute Discrete Convolution:

1. Make 𝒙(𝒏) = 𝒙(𝒌) and 𝒉(𝒏) = 𝒉(𝒌)
2. Flip one of the signals (typically 𝒉[𝒏]) around the vertical axis to get
𝒉[−𝒏] = 𝒉[−𝒌].
3. Shift this flipped signal by 𝒏 units to get 𝒉[−𝒌 + 𝒏] = 𝒉[𝒏 − 𝒌].
4. Multiply the overlapping values of 𝒙[𝒌] and 𝒉[𝒏 − 𝒌] for each shift.
5. Sum all the products for each shift to get a single output value for 𝒚[𝒏].
30
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

➢ Linear Convolution:
• Example: Perform the convolution in graphical method of the two
sequences
𝑥 (𝑛) = {1,2,3,4}; ℎ(𝑛) = {1,1,1,1}
 Given that,
𝒙(𝒏) = {𝟏, 𝟐, 𝟑, 𝟒}; 𝒉(𝒏) = {𝟏, 𝟏, 𝟏, 𝟏}
Here,
𝑳 = 𝒆𝒍𝒆𝒎𝒆𝒏𝒕𝒔 𝒐𝒇 𝒙(𝒏) = 𝟒
𝑴 = 𝒆𝒍𝒆𝒎𝒆𝒏𝒕𝒔 𝒐𝒇 𝒉(𝒏) = 𝟒
To find 𝒚(𝒏) we have to do convolution (𝑳 + 𝑴) − 𝟏 = (𝟒 + 𝟒) − 𝟏 = 𝟕
times.
Here,
𝒙(𝒏) starts from 𝒏𝟏 = 𝟎 and Ends at 𝑵𝟏 = 𝟑
𝒉(𝒏) starts from 𝒏𝟐 𝟎 and Ends at 𝑵𝟐 = 𝟑
So,
𝒚(𝒏) starts from (𝒏𝟏 + 𝒏𝟐 ) = (𝟎 + 𝟎) = 𝟎 and Ends at (𝑵𝟏 + 𝑵𝟐 ) =
(𝟑 + 𝟑 ) = 𝟔

Let assume,
𝒙(𝒏) = 𝒙(𝒌)
𝒉(𝒏) = 𝒉(𝒌)
After doing time reversal of the signal 𝒉(𝒌) we get 𝒉(−𝒌)
Then we need to shift the flipped signal by 𝒏 value and get 𝒉(−𝒌 + 𝒏) =
𝒉(𝒏 − 𝒌)

31
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

We know that,
∞

𝒚[𝒏] = ∑ 𝒙(𝒌) ∗ 𝒉(𝒏 − 𝒌)

𝒌=−∞

When,
𝒏=𝟎

∴ 𝒚[𝟎] = 𝐱(𝟎) ∗ 𝒉(𝟎) = 𝟏 ∗ 𝟏 = 𝟏

When,
𝒏=𝟏

∴ 𝒚[𝟏] = 𝐱(𝟎) ∗ 𝐡(𝟎) + 𝐱(𝟏) ∗ 𝐡(𝟏) = (𝟏 ∗ 𝟏) + (𝟐 ∗ 𝟏) = 𝟑

32
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

When,
𝒏=𝟐

∴ 𝒚[𝟐] = 𝐱(𝟎) ∗ 𝐡(𝟎) + 𝐱(𝟏) ∗ 𝐡(𝟏) + 𝐱(𝟐) ∗ 𝐡(𝟐)

= (𝟏 ∗ 𝟏 ) + (𝟐 ∗ 𝟏 ) + (𝟑 ∗ 𝟏 ) = 𝟔
When,
𝒏=𝟑

∴ 𝒚[𝟑] = 𝐱(𝟎) ∗ 𝐡(𝟎) + 𝐱(𝟏) ∗ 𝐡(𝟏) + 𝐱(𝟐) ∗ 𝐡(𝟐) + 𝐱(𝟑) ∗ 𝐡(𝟑)

= (𝟏 ∗ 𝟏) + (𝟐 ∗ 𝟏) + (𝟑 ∗ 𝟏) + (𝟒 ∗ 𝟏) + 𝟎 = 𝟏𝟎

When,
𝒏=𝟒

∴ 𝒚[𝟒] = 𝐱(𝟏) ∗ 𝐡(𝟏) + 𝐱(𝟐) ∗ 𝐡(𝟐) + 𝐱(𝟑) ∗ 𝐡(𝟑)

= (𝟐 ∗ 𝟏 ) + (𝟑 ∗ 𝟏 ) + (𝟒 ∗ 𝟏 ) + 𝟎 = 𝟗

33
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

When,
𝒏=𝟓

∴ 𝒚[𝟓] = 𝐱(𝟐) ∗ 𝐡(𝟐) + 𝐱(𝟑) ∗ 𝐡(𝟑) = (𝟑 ∗ 𝟏) + (𝟒 ∗ 𝟏) + 𝟎 = 𝟕

When,
𝒏=𝟔

∴ 𝒚[𝟔] = 𝐱(𝟑) ∗ 𝐡(𝟑) = 𝟒 ∗ 𝟏 = 𝟒

∴ 𝒚[𝒏] = 𝒙[𝒏] ∗ 𝒉[𝒏] = {𝟏, 𝟑, 𝟔, 𝟏𝟎, 𝟗, 𝟕, 𝟒}

• Example: Perform the convolution in cross tabular method of the two

sequences
𝑥 (𝑛) = {1,2,3,4}; ℎ(𝑛) = {1,1,1,1}


𝒙(𝒏)
𝑥 (0) ∗ ℎ(0) 𝑥 (1) ∗ ℎ(0) 𝑥 (2) ∗ ℎ(0) 𝑥 (3) ∗ ℎ(0)
𝒉(𝒏)
𝑥 (0) ∗ ℎ(1) 𝑥 (1) ∗ ℎ(1) 𝑥 (2) ∗ ℎ(1) 𝑥 (3) ∗ ℎ(1)
𝑥 (0) ∗ ℎ(2) 𝑥 (1) ∗ ℎ(2) 𝑥 (2) ∗ ℎ(2) 𝑥 (3) ∗ ℎ(2)
𝑥 (0) ∗ ℎ(3) 𝑥 (1) ∗ ℎ(3) 𝑥 (2) ∗ ℎ(3) 𝑥 (3) ∗ ℎ(3)

34
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

Here,
Given that,
𝒙(𝒏) = {𝟏, 𝟐, 𝟑, 𝟒}; 𝒉(𝒏) = {𝟏, 𝟏, 𝟏, 𝟏}

𝑥(0) = 1 ℎ(0) = 1

𝑥(1) = 2 ℎ(1) = 1

𝑥(2) = 3 ℎ(2) = 1

𝑥(3) = 4 ℎ(3) = 1

𝒙(𝒏) 𝒙 (𝟎 ) = 𝟏 𝒙(𝟏) = 𝟐 𝒙(𝟐) = 𝟑 𝒙(𝟑) = 𝟒

𝒉(𝒏)

𝒉(𝟎) = 𝟏 1 2 3 4
𝒉(𝟏) = 𝟏 1 2 3 4
𝒉(𝟐) = 𝟏 1 2 3 4
𝒉(𝟑) = 𝟏 1 2 3 4
Here,
𝒚(𝟎) = 𝟏
𝒚(𝟏) = 𝟏 + 𝟐 = 𝟑
𝒚(𝟐) = 𝟏 + 𝟐 + 𝟑 = 𝟔
𝒚(𝟑) = 𝟏 + 𝟐 + 𝟑 + 𝟒 = 𝟏𝟎
𝒚(𝟒) = 𝟐 + 𝟑 + 𝟒 = 𝟗
𝒚(𝟓) = 𝟑 + 𝟒 = 𝟕
𝒚(𝟔) = 𝟒
∴ 𝒚[𝒏] = {𝟏, 𝟑, 𝟔, 𝟏𝟎, 𝟗, 𝟕, 𝟒}

35
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

• Example: Perform the convolution in matrix method of the two sequences

𝑥 (𝑛) = {1,2,3,4}; ℎ(𝑛) = {1,1,1,1}
 Given that,
𝒙(𝒏) = {𝟏, 𝟐, 𝟑, 𝟒}; 𝒉(𝒏) = {𝟏, 𝟏, 𝟏, 𝟏}
Here,
𝑳 = 𝒆𝒍𝒆𝒎𝒆𝒏𝒕𝒔 𝒐𝒇 𝒙(𝒏) = 𝟒
𝑴 = 𝒆𝒍𝒆𝒎𝒆𝒏𝒕𝒔 𝒐𝒇 𝒉(𝒏) = 𝟒
To find 𝒚(𝒏) we have to convert 𝒙(𝒏) into {(𝑳 + 𝑴) − 𝟏} × 𝑳 =
{(𝟒 + 𝟒) − 𝟏} × 𝟒 = 𝟕 × 𝟒 matrix
And convert 𝒉(𝒏) into 𝑴 × 𝟏 = 𝟒 × 𝟏 matrix
Converting 𝒙(𝒏) and 𝒉(𝒏) into matrix form we get,

Here,
𝒚 = [𝒙][𝒉]

36
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

∴ 𝒚[𝒏] = {𝟏, 𝟑, 𝟔, 𝟏𝟎, 𝟗, 𝟕, 𝟒}

• Example: Perform the convolution in tabular method of the two sequences

𝑥 (𝑛) = {1,2,3,4}; ℎ(𝑛) = {1,1,1,1}
 Given that,
𝒙(𝒏) = {𝟏, 𝟐, 𝟑, 𝟒}; 𝒉(𝒏) = {𝟏, 𝟏, 𝟏, 𝟏}

𝑥(0) = 1 ℎ(0) = 1
𝑥(1) = 2 ℎ(1) = 1
𝑥(2) = 3 ℎ(2) = 1
𝑥(3) = 4 ℎ(3) = 1

37
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

By putting these values we get,

∴ 𝒚[𝒏] = {𝟏, 𝟑, 𝟔, 𝟏𝟎, 𝟗, 𝟕, 𝟒}

38
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

➢ Circular Convolution:
1. Stockman’s Method
2. Matrix Method
3. DFT-IDFT Method

➢ Stockman’s Method:
Let the inner circle,
𝒙𝟏 (𝒏) = {𝒙𝟏 (𝟎), 𝒙𝟏 (𝟏), 𝒙𝟏 (𝟐), 𝒙𝟏 (𝟑)}
This will divide anti clock wise
Let the outer circle,
𝒙𝟐 (𝒏) = {𝒙𝟐 (𝟎), 𝒙𝟐 (𝟏), 𝒙𝟐 (𝟐), 𝒙𝟐 (𝟑)}
This will divide clock wise

To get 𝒚[𝒏], the inner circle 𝒙𝟏 (𝒏) will remain constant but outer circle 𝒙𝟐 (𝒏)
will rotate anti clock wise.
Here,
𝒚[𝒏] = 𝒙𝟏 (𝟎)𝒙𝟐 (𝟎) + 𝒙𝟏 (𝟏)𝒙𝟐 (𝟑) + 𝒙𝟏 (𝟐)𝒙𝟐 (𝟐) + 𝒙𝟏 (𝟑)𝒙𝟐 (𝟏)

39
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

• Example: Perform the circular convolution in stockman’s method of the

two circles,
𝑥1 (𝑛) = {2,1,2,1}; 𝑥2 (𝑛) = {1,2,3,4}
 Given that,
𝒙𝟏 (𝒏) = {𝟐, 𝟏, 𝟐, 𝟏}; 𝒙𝟐 (𝒏) = {𝟏, 𝟐, 𝟑, 𝟒}

𝒚(𝟎) = (𝟐 ∗ 𝟏) + (𝟏 ∗ 𝟒) + (𝟐 ∗ 𝟑) + (𝟏 ∗ 𝟐)
𝑦(0) = 2 + 4 + 6 + 2
∴ 𝒚(𝟎) = 𝟏𝟒

𝒚(𝟏) = (𝟐 ∗ 𝟐) + (𝟏 ∗ 𝟏) + (𝟐 ∗ 𝟒) + (𝟏 ∗ 𝟑)
𝑦(1) = 4 + 1 + 8 + 3
∴ 𝒚(𝟏) = 𝟏𝟔

𝒚(𝟐) = (𝟐 ∗ 𝟑) + (𝟏 ∗ 𝟐) + (𝟐 ∗ 𝟏) + (𝟏 ∗ 𝟒)
𝑦(2) = 6 + 2 + 2 + 4
∴ 𝒚(𝟐) = 𝟏𝟒

40
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

𝒚(𝟑) = (𝟐 ∗ 𝟒) + (𝟏 ∗ 𝟑) + (𝟐 ∗ 𝟐) + (𝟏 ∗ 𝟏)
𝑦(2) = 8 + 3 + 4 + 1
∴ 𝒚(𝟐) = 𝟏𝟔

∴ 𝒚[𝒏] = {𝟏𝟒, 𝟏𝟔, 𝟏𝟒, 𝟏𝟔}

➢ Matrix Method:
• Example: Perform the circular convolution in matrix method of the two
circles,
𝑥1 (𝑛) = {2,1,2,1}; 𝑥2 (𝑛) = {1,2,3,4}
 Given that,
𝑥1 (𝑛) = {2,1,2,1}; 𝑥2 (𝑛) = {1,2,3,4}
Here,

41
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

∴ 𝒚[𝒏] = {𝟏𝟒, 𝟏𝟔, 𝟏𝟒, 𝟏𝟔}

➢ DFT Method: Discrete Fourier Transform

The Discrete Fourier Transform (DFT) is a mathematical transform that
converts a sequence of sampled data points from the time domain into the
frequency domain.
𝒆±𝒋𝜽 = 𝐜𝐨𝐬 𝜽 ± 𝒋 𝐬𝐢𝐧 𝜽
𝑵−𝟏 𝑵−𝟏
𝟐𝝅 𝟐𝝅
𝑿(𝒌) = ∑ 𝒙(𝒏) 𝒆−𝒋𝝎𝒏 = ∑ 𝒙(𝒏) 𝒆−𝒋 𝑵 𝒌𝒏 𝝎=
𝑵
𝒌
𝒏=𝟎 𝒏=𝟎
𝑵−𝟏
𝟐𝝅 𝟐𝝅
𝑿(𝒌) = ∑ 𝒙(𝒏) [𝐜𝐨𝐬 ( 𝒌𝒏) − 𝒋 𝐬𝐢𝐧 ( 𝒌𝒏)]
𝑵 𝑵
𝒏=𝟎

• Compute the N-point DFT of the given signal 𝒙(𝒏) = 𝜹(𝒏)

 Given that,
𝒙(𝒏) = 𝜹(𝒏)
We know that,
𝟏, 𝒏 = 𝟎
For 𝜹(𝒏) this signal 𝜹(𝒏) = {
𝟎, 𝒏 ≠ 𝟎
Now,
𝑵−𝟏
𝟐𝝅
𝑿(𝒌) = ∑ 𝒙(𝒏) 𝒆−𝒋 𝑵 𝒌𝒏
𝒏=𝟎
2𝜋 4𝜋 6𝜋
𝑋(𝑘) = 𝛿 (0)𝑒 −0 + 𝛿 (1)𝑒 −𝑗 𝑁 𝑘 + 𝛿 (2)𝑒 −𝑗 𝑁 𝑘 + 𝛿 (3)𝑒 −𝑗 𝑁 𝑘 + ⋯ ….
𝑋(𝑘) = (1 ∗ 1) + 0 + 0 + 0 + ⋯ … . +0
∴ 𝑿(𝒌) = 𝟏

42
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

• Compute the N-point DFT of the given signal 𝒙(𝒏) = 𝜹(𝒏 − 𝒏𝟎 );

𝟎 < 𝒏𝟎 < 𝑵
 Given that,
𝒙(𝒏) = 𝜹(𝒏 − 𝒏𝟎 )
We know,
𝑵−𝟏
𝟐𝝅
𝑿(𝒌) = ∑ 𝒙(𝒏) 𝒆−𝒋 𝑵 𝒌𝒏
𝒏=𝟎

𝑵−𝟏
𝟐𝝅
𝑿(𝒌) = ∑ 𝜹(𝒏 − 𝒏𝟎 ) 𝒆−𝒋 𝑵 𝒌𝒏
𝒏=𝟎

𝟏, 𝒏 = 𝟎
For 𝜹(𝒏) this signal 𝜹(𝒏) = {
𝟎, 𝒏 ≠ 𝟎
Here, To get 𝜹(𝟎) = 𝟏; 𝒏 = 𝒏𝟎
𝑁−1
2𝜋
𝑋(𝑘) = ∑ 𝛿 (𝑛0 − 𝑛0 ) 𝑒 −𝑗 𝑁 𝑘𝑛
𝑛=0
𝑁−1
2𝜋
𝑋(𝑘) = ∑ 𝛿 (0) 𝑒 −𝑗 𝑁 𝑘𝑛𝑜
𝑛=0
2𝜋
𝑋(𝑘) = 𝛿 (0)𝑒 −𝑗 𝑁 𝑘𝑛𝑜
𝟐𝝅
∴ 𝑿(𝒌) = 𝒆−𝒋 𝑵 𝒌𝒏𝒐

43
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

• Find the DFT of the sequence 𝑥 (𝑛) = 1; 0 ≤ 𝑛 ≤ 2 for 𝑁 = 4 also find

magnitude and Phase.
 Given that,
𝒙(𝒏) = {𝟏, 𝟏, 𝟏, 𝟎}
We know,
𝑵−𝟏
𝟐𝝅
𝑿(𝒌) = ∑ 𝒙(𝒏) 𝒆−𝒋 𝑵 𝒌𝒏
𝒏=𝟎
4−1
2𝜋
𝑋(𝑘) = ∑ 𝑥 (𝑛) 𝑒 −𝑗 𝑁 𝑘𝑛
𝑛=0
3
2𝜋
𝑋(𝑘) = ∑ 𝑥 (𝑛) 𝑒 −𝑗 𝑁 𝑘𝑛
𝑛=0
2𝜋 4𝜋 6𝜋
𝑋(𝑘) = 𝑥 (0)𝑒 0 + 𝑥(1)𝑒 −𝑗 4 𝑘 + 𝑥 (2)𝑒 −𝑗 4 𝑘 + 𝑥 (3)𝑒 −𝑗 4 𝑘
𝜋 3𝜋
𝑋(𝑘) = 𝑥 (0) + 𝑥 (1)𝑒 −𝑗 2 𝑘 + 𝑥 (2)𝑒 −𝑗𝜋𝑘 + 𝑥(3)𝑒 −𝑗 2 𝑘
𝜋
𝑋(𝑘) = 1 + 𝑒 −𝑗 2 𝑘 + 𝑒 −𝑗𝜋𝑘 + 0
Here,
𝒙(𝟎) = 𝟏, 𝒙(𝟏) = 𝟏, 𝒙(𝟐) = 𝟏, 𝒙(𝟑) = 𝟎
𝜋
𝑋(𝑘) = 1 + 𝑒 −𝑗 2 𝑘 + 𝑒 −𝑗𝜋𝑘 + 0
𝝅
∴ 𝑿(𝒌) = 𝟏 + 𝒆−𝒋𝟐 𝒌 + 𝒆−𝒋𝝅𝒌
𝝅 𝝅
∴ 𝑿(𝒌) = 𝟏 + [𝐜𝐨𝐬 ( 𝒌) − 𝒋 𝐬𝐢𝐧 ( 𝒌)] + [𝐜𝐨𝐬(𝝅𝒌) − 𝒋 𝐬𝐢𝐧(𝝅𝒌)]
𝟐 𝟐
For 𝒌 = 𝟎,
∴ 𝑿(𝟎) = 𝟏 + 𝟏 + 𝟏 = 𝟑
For 𝒌 = 𝟏,
𝜋 𝜋
𝑋(1) = 1 + [cos ( ) − 𝑗 sin ( )] + [cos(𝜋) − 𝑗 sin(𝜋)]
2 2

44
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03
𝜋 𝜋
𝑋(1) = 1 + [cos ( ) − 𝑗 sin ( )] + [cos(𝜋) − 𝑗 sin(𝜋)]
2 2
𝑋(1) = 1 + [0 − 𝑗] + [−1 − 0]
∴ 𝑿(𝟏) = 𝟏 − 𝒋 − 𝟏 = −𝒋
For 𝒌 = 𝟐,
𝑋(2) = 1 + [cos(𝜋) − 𝑗 sin(𝜋)] + [cos(2𝜋) − 𝑗 sin(2𝜋)]
𝑋(2) = 1 + [−1 + 0] + [1 − 0]
∴ 𝑿(𝟐) = 𝟏 − 𝟏 + 𝟏 = 𝟏
For 𝒌 = 𝟑,
3𝜋 3𝜋
𝑋(3) = 1 + [cos ( ) − 𝑗 sin ( )] + [cos(3𝜋) − 𝑗 sin(3𝜋)]
2 2
𝑋(3) = 1 + [0 + 𝑗] + [−1 − 0]
∴ 𝑿(𝟑) = 𝟏 + 𝒋 − 𝟏 = 𝒋
∴ 𝑿(𝒌) = {𝟑, −𝒋, 𝟏, 𝒋}
Magnitude,

|𝑿(𝒌)| = {√𝟑𝟐 + 𝟎𝟐 , √𝟎𝟐 + (−𝟏)𝟐 , √𝟏𝟐 + 𝟎𝟐 , √𝟎𝟐 + (𝟏)𝟐 }

∴ |𝑿(𝒌)| = {𝟑, 𝟏, 𝟏, 𝟏}

𝟎 −𝟏 𝟎 𝟏
Phase = {𝐭𝐚𝐧−𝟏 ( ) , 𝐭𝐚𝐧−𝟏 ( ) , 𝐭𝐚𝐧−𝟏 ( ) , 𝐭𝐚𝐧−𝟏 ( )}
𝟑 𝟎 𝟏 𝟎
𝝅 𝝅
∴ Phase = {𝟎, , 𝟎, }
𝟐 𝟐

45
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03
1
• Find the DFT of the sequence 𝑥 (𝑛) = ; −1 ≤ 𝑛 ≤ 1
5

 Given that,
𝟏 𝟏 𝟏
𝒙(𝒏) = { , , }
𝟓 𝟓 𝟓
Here, 𝑵=𝟑
We know,
𝑵−𝟏
𝟐𝝅
𝑿(𝒌) = ∑ 𝒙(𝒏) 𝒆−𝒋 𝑵 𝒌𝒏 𝟐𝝅
𝒏=−𝟏 𝝎= 𝒌
𝑵
𝟑−𝟏
𝟐𝝅
𝑿(𝒌) = ∑ 𝒙(𝒏) 𝒆−𝒋 𝑵 𝒌𝒏
𝒏=−𝟏
𝟐
𝟐𝝅
𝑿(𝒌) = ∑ 𝒙(𝒏) 𝒆−𝒋 𝑵 𝒌𝒏
𝒏=−𝟏
2𝜋 2𝜋 2𝜋
𝑗 𝑘 0 −𝑗 𝑘 −𝑗 𝑘
𝑋(𝑘) = 𝑥(−1)𝑒 𝑁 + 𝑥 (0)𝑒 + 𝑥 (1)𝑒 𝑁 + 𝑥(2)𝑒 𝑁

1 2𝜋 1 1 2𝜋
𝑋(𝑘) = ( ) 𝑒 𝑗 𝑁 𝑘 + ( ) + ( ) 𝑒 −𝑗 𝑁 𝑘 + 0
5 5 5
1 2𝜋
𝑗 𝑘
2𝜋
−𝑗 𝑘 1
𝑋(𝑘) = ( ) [𝑒 𝑁 +𝑒 𝑁 ]+( )
5 5
1 1
𝑋(𝑘) = ( ) [𝑒 𝑗𝜔 + 𝑒 −𝑗𝜔 ] + ( )
5 5
1 𝑒 𝑗𝜔 + 𝑒 −𝑗𝜔 1
𝑋 (𝑘 ) = ( ) 2 [ ]+( )
5 2 5
1 1
𝑋(𝑘) = ( ) 2 cos 𝜔 + ( )
5 5
1
𝑋(𝑘) = [2 cos 𝜔 + 1]
5
𝟏 𝟐𝝅
∴ 𝑿(𝒌) = [𝟏 + 𝟐 𝐜𝐨𝐬 𝒌]
𝟓 𝑵

46
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

• Example- 6.7: Derive the DFT of the sample data sequence 𝑥 (𝑛) =
{1,1,2,2,3,3} and compute the corresponding amplitude and phase spectrum.
 Given that,
𝒙(𝒏) = {𝟏, 𝟏, 𝟐, 𝟐, 𝟑, 𝟑}
Here,
𝑵=𝟔
We know,
𝑵−𝟏
𝟐𝝅
𝑿(𝒌) = ∑ 𝒙(𝒏) 𝒆−𝒋 𝑵 𝒌𝒏
𝒏=𝟎
6−1
2𝜋
𝑋(𝑘) = ∑ 𝑥 (𝑛) 𝑒 −𝑗 6 𝑘𝑛
𝑛=0
5
𝜋
𝑋(𝑘) = ∑ 𝑥 (𝑛) 𝑒 −𝑗 3 𝑘𝑛
𝑛=0
𝜋 2𝜋 4𝜋
𝑋(𝑘) = 𝑥(0)𝑒 0 + 𝑥 (1)𝑒 −𝑗 3 𝑘 + 𝑥 (2)𝑒 −𝑗 3 𝑘 + 𝑥 (3)𝑒 −𝑗𝜋𝑘 + 𝑥(4)𝑒 −𝑗 3 𝑘
5𝜋
+ 𝑥 (5)𝑒 −𝑗 3 𝑘
𝝅 𝟐𝝅 𝟒𝝅 𝟓𝝅
∴ 𝑿(𝒌) = 𝟏 + 𝒆−𝒋𝟑 𝒌 + 𝟐𝒆−𝒋 𝟑 𝒌 + 𝟐𝒆−𝒋𝝅𝒌 + 𝟑𝒆−𝒋 𝟑 𝒌 + 𝟑𝒆−𝒋 𝟑 𝒌
𝝅 𝝅 𝟐𝝅 𝟐𝝅
∴ 𝑿(𝒌) = 𝟏 + [𝐜𝐨𝐬 ( 𝒌) − 𝒋 𝐬𝐢𝐧 ( 𝒌)] +𝟐 [𝐜𝐨𝐬 ( 𝒌) − 𝒋 𝐬𝐢𝐧 ( 𝒌)]
𝟑 𝟑 𝟑 𝟑

𝟒𝝅 𝟒𝝅
+𝟐[𝐜𝐨𝐬(𝝅𝒌) − 𝒋 𝐬𝐢𝐧(𝝅𝒌)] + 𝟑 [𝐜𝐨𝐬 ( 𝒌) − 𝒋 𝐬𝐢𝐧 ( 𝒌)]
𝟑 𝟑
𝟓𝝅 𝟓𝝅
+ 𝟑 [𝐜𝐨𝐬 ( 𝒌) − 𝒋 𝐬𝐢𝐧 ( 𝒌)]
𝟑 𝟑
Here,
𝒙(𝟎) = 𝟏, 𝒙(𝟏) = 𝟏, 𝒙(𝟐) = 𝟐, 𝒙(𝟑) = 𝟐, 𝒙(𝟒) = 𝟑, 𝒙(𝟓) = 𝟑

47
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

For 𝒌 = 𝟎,
∴ 𝑿(𝟎) = 𝟏 + 𝟏 + 𝟐 + 𝟐 + 𝟑 + 𝟑 = 𝟏𝟐
For 𝒌 = 𝟏,
𝜋 𝜋 2𝜋 2𝜋
𝑋(1) = 1 + [cos ( ) − 𝑗 sin ( )] + 2 [cos ( ) − 𝑗 sin ( )] +
3 3 3 3
4𝜋 4𝜋
2[cos(𝜋) − 𝑗 sin(𝜋)] + 3 [cos ( ) − 𝑗 sin ( )]
3 3
5𝜋 5𝜋
+ 3 [cos ( ) − 𝑗 sin ( )]
3 3
∴ 𝑿(𝟏) = −𝟏. 𝟓 + 𝒋 𝟐. 𝟓𝟗
For 𝒌 = 𝟐,
2𝜋 2𝜋 4𝜋 4𝜋
𝑋(2) = 1 + [cos ( ) − 𝑗 sin ( )] + 2 [cos ( ) − 𝑗 sin ( )] +
3 3 3 3
8𝜋 8𝜋
2[cos(2𝜋) − 𝑗 sin(2𝜋)] + 3 [cos ( ) − 𝑗 sin ( )]
3 3
10𝜋 10𝜋
+ 3 [cos ( ) − 𝑗 sin ( )]
3 3
∴ 𝑿(𝟐) = −𝟏. 𝟓 + 𝒋 𝟎. 𝟖𝟔
For 𝒌 = 𝟑,
𝑋(3) = 1 + [cos(𝜋) − 𝑗 sin(𝜋)] + 2[cos(2𝜋) − 𝑗 sin(2𝜋)] +
2[cos(3𝜋) − 𝑗 sin(3𝜋)] + 3[cos(4𝜋) − 𝑗 sin(4𝜋)]
+ 3[cos(5𝜋) − 𝑗 sin(5𝜋)]
∴ 𝑿( 𝟑 ) = 𝟎
For 𝒌 = 𝟒,
4𝜋 4𝜋 8𝜋 8𝜋
𝑋(4) = 1 + [cos ( ) − 𝑗 sin ( )] + 2 [cos ( ) − 𝑗 sin ( )] +
3 3 3 3
16𝜋 16𝜋
2[cos(4𝜋) − 𝑗 sin(4𝜋)] + 3 [cos ( ) − 𝑗 sin ( )]
3 3
20𝜋 20𝜋
+ 3 [cos ( ) − 𝑗 sin ( )]
3 3

48
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

∴ 𝑿(𝟒) = −𝟏. 𝟓 − 𝒋 𝟎. 𝟖𝟔
For 𝒌 = 𝟓,
5𝜋 5𝜋 10𝜋 10𝜋
𝑋(5) = 1 + [cos ( ) − 𝑗 sin ( )] + 2 [cos ( ) − 𝑗 sin ( )] +
3 3 3 3
20𝜋 20𝜋
2[cos(5𝜋) − 𝑗 sin(5𝜋)] + 3 [cos ( ) − 𝑗 sin ( )]
3 3
25𝜋 25𝜋
+ 3 [cos ( ) − 𝑗 sin ( )]
3 3
∴ 𝑿(𝟓) = −𝟏. 𝟓 − 𝒋 𝟐. 𝟓𝟗

∴ 𝑿(𝒌) = {𝟏𝟐, −𝟏. 𝟓 + 𝒋 𝟐. 𝟓𝟗, −𝟏. 𝟓 + 𝒋 𝟎. 𝟖𝟔, 𝟎, −𝟏. 𝟓 − 𝒋 𝟎. 𝟖𝟔, −𝟏. 𝟓

− 𝒋 𝟐. 𝟓𝟗 }

Amplitude,
|𝑿(𝒌)|
√𝟏𝟐𝟐 + 𝟎𝟐 , √(−𝟏. 𝟓)𝟐 + (𝟐. 𝟓𝟗)𝟐 , √(−𝟏. 𝟓)𝟐 + (𝟎. 𝟖𝟔)𝟐 , √𝟎𝟐 ,
={ }
𝟐 𝟐 𝟐
√(−𝟏. 𝟓) + (−𝟎. 𝟖𝟔) , √(−𝟏. 𝟓) + (−𝟐. 𝟓𝟗) 𝟐

∴ |𝑿(𝒌)| = {𝟏𝟐, 𝟐. 𝟗𝟗, 𝟏. 𝟕𝟐, 𝟎, 𝟏. 𝟕𝟐, 𝟐. 𝟗𝟗}

Phase =
𝟎 𝟐.𝟓𝟗 𝟎.𝟖𝟔 −𝟎.𝟖𝟔
𝐭𝐚𝐧−𝟏 ( ) , 𝐭𝐚𝐧−𝟏 ( ) , 𝐭𝐚𝐧−𝟏 ( ) , 𝐭𝐚𝐧−𝟏 (𝟎), 𝐭𝐚𝐧−𝟏 ( ),
𝟏𝟐 −𝟏.𝟓 −𝟏.𝟓 −𝟏.𝟓
{ −𝟐.𝟓𝟗
}
𝐭𝐚𝐧−𝟏 ( )
−𝟏.𝟓
𝝅 𝝅 𝝅
∴ Phase = {𝟎, − , 𝟎, , }
𝟑 𝟔 𝟑

49
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

• Example- 6.8: Compute the DFT of the each of the following finite length
sequences considered to be of length 𝑁
a. 𝑥 (𝑛) = 𝛿(𝑛)
b. 𝑥 (𝑛) = 𝛿 (𝑛 − 𝑛0 ); 𝑤ℎ𝑒𝑟𝑒 0 < 𝑛0 < 𝑁

a. Given that,
𝒙(𝒏) = 𝜹(𝒏)
We know that,
𝟏, 𝒏 = 𝟎
For 𝜹(𝒏) this signal 𝜹(𝒏) = {
𝟎, 𝒏 ≠ 𝟎
Now,
𝑵−𝟏
𝟐𝝅
𝑿(𝒌) = ∑ 𝒙(𝒏) 𝒆−𝒋 𝑵 𝒌𝒏
𝒏=𝟎
𝑁−1
2𝜋
𝑋(𝑘) = ∑ 𝛿(𝑛) 𝑒 −𝑗 𝑁 𝑘𝑛
𝑛=0
2𝜋 4𝜋 6𝜋
𝑋(𝑘) = 𝛿 (0)𝑒 −0 + 𝛿 (1)𝑒 −𝑗 𝑁 𝑘 + 𝛿 (2)𝑒 −𝑗 𝑁 𝑘 + 𝛿 (3)𝑒 −𝑗 𝑁 𝑘 + ⋯ ….
𝑋(𝑘) = (1 ∗ 1) + 0 + 0 + 0 + ⋯ … . +0
∴ 𝑿(𝒌) = 𝟏

b. Given that,
𝒙(𝒏) = 𝜹(𝒏 − 𝒏𝟎 )
We know,
𝑵−𝟏
𝟐𝝅
𝑿(𝒌) = ∑ 𝒙(𝒏) 𝒆−𝒋 𝑵 𝒌𝒏
𝒏=𝟎

50
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03
𝑁−1
2𝜋
𝑋(𝑘) = ∑ 𝛿 (𝑛 − 𝑛0 ) 𝑒 −𝑗 𝑁 𝑘𝑛
𝑛=0

𝟏, 𝒏 = 𝟎
For 𝜹(𝒏) this signal 𝜹(𝒏) = {
𝟎, 𝒏 ≠ 𝟎
Here, To get 𝜹(𝟎) = 𝟏; 𝒏 = 𝒏𝟎
𝑁−1
2𝜋
𝑋(𝑘) = ∑ 𝛿 (𝑛0 − 𝑛0 ) 𝑒 −𝑗 𝑁 𝑘𝑛0
𝑛0 =0
𝑁−1
2𝜋
𝑋(𝑘) = ∑ 𝛿 (0) 𝑒 −𝑗 𝑁 𝑘𝑛𝑜
𝑛0 =0
2𝜋
𝑋(𝑘) = 𝛿 (0)𝑒 −𝑗 𝑁 𝑘𝑛𝑜
𝟐𝝅
∴ 𝑿(𝒌) = 𝒆−𝒋 𝑵 𝒌𝒏𝒐

• Example- 6.9: Find the N-point DFT for 𝑥(𝑛) = 𝑎𝑛 for 0 < 𝑎 < 1
 Given that,
𝒙(𝒏) = 𝒂𝒏
We know,
𝑵−𝟏
𝟐𝝅
𝑿(𝒌) = ∑ 𝒙(𝒏) 𝒆−𝒋 𝑵 𝒌𝒏
𝒏=𝟎
𝑁−1
2𝜋
𝑋(𝑘) = ∑ 𝑎𝑛 𝑒 −𝑗 𝑁 𝑘𝑛
𝑛=0
𝑁−1 𝑛
2𝜋
𝑋 (𝑘 ) = ∑ (𝑎𝑒 −𝑗 𝑁 𝑘 )
𝑛=0

51
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

2𝜋 𝑁
−𝑗 𝑘
1 − (𝑎𝑒 𝑁 )
𝑋 (𝑘 ) = 2𝜋
−𝑗 𝑘
1 − 𝑎𝑒 𝑁

1 − 𝑎𝑁 𝑒 −𝑗2𝜋𝑘
𝑋 (𝑘 ) = 2𝜋
1 − 𝑎𝑒 −𝑗 𝑁 𝑘
𝟏 − 𝒂𝑵
∴ 𝑿(𝒌) = 𝟐𝝅
−𝒋 𝒌
𝟏− 𝒂𝒆 𝑵

𝑛𝜋
• Example- 6.10: Find the 4-point DFT of the sequence 𝑥(𝑛) = cos
4

 Given that,
𝑵=𝟒
𝒏𝝅
𝒙(𝒏) = 𝐜𝐨𝐬
𝟒
We know,
𝑵−𝟏
𝟐𝝅
𝑿(𝒌) = ∑ 𝒙(𝒏) 𝒆−𝒋 𝑵 𝒌𝒏
𝒏=𝟎
4−1
𝑛𝜋 −𝑗2𝜋𝑘𝑛
𝑋(𝑘) = ∑ cos 𝑒 4
4
𝑛=0
3
𝑛𝜋 −𝑗𝜋𝑘𝑛
𝑋(𝑘) = ∑ cos 𝑒 2
4
𝑛=0

𝜋 𝜋 𝜋 3𝜋 −𝑗3𝜋𝑘
𝑋(𝑘) = cos 0 𝑒 0 + cos 𝑒 −𝑗 2 𝑘 + cos 𝑒 −𝑗𝜋𝑘 + cos 𝑒 2
4 2 4
𝜋 3𝜋
𝑋(𝑘) = 1 + 0.707 𝑒 −𝑗 2 𝑘 + 0 − 0.707 𝑒 −𝑗 2 𝑘
𝝅 𝟑𝝅
∴ 𝑿(𝒌) = 𝟏 + 𝟎. 𝟕𝟎𝟕 𝒆−𝒋𝟐 𝒌 − 𝟎. 𝟕𝟎𝟕 𝒆−𝒋 𝟐 𝒌

52
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03
𝝅 𝝅
∴ 𝑿(𝒌) = 𝟏 + 𝟎. 𝟕𝟎𝟕 [𝐜𝐨𝐬 ( 𝒌) − 𝒋 𝐬𝐢𝐧 ( 𝒌)]
𝟐 𝟐
𝟑𝝅 𝟑𝝅
− 𝟎. 𝟕𝟎𝟕 [𝐜𝐨𝐬 ( 𝒌) − 𝒋 𝐬𝐢𝐧 ( 𝒌)]
𝟐 𝟐
For 𝒌 = 𝟎,
𝑋(0) = 1 + 0.707 − 0.707
∴ 𝑿( 𝟎 ) = 𝟏
For 𝒌 = 𝟏,
𝜋 𝜋 3𝜋 3𝜋
𝑋(1) = 1 + 0.707 [cos ( ) − 𝑗 sin ( )] − 0.707 [cos ( ) − 𝑗 sin ( )]
2 2 2 2
∴ 𝑿(𝟏) = 𝟏 − 𝒋 𝟏. 𝟒𝟏𝟒
For 𝒌 = 𝟐,
𝑋(2) = 1 + 0.707[cos(𝜋) − 𝑗 sin(𝜋)] − 0.707[cos(3𝜋) − 𝑗 sin(3𝜋)]
∴ 𝑿( 𝟐 ) = 𝟏
For 𝒌 = 𝟑,
3𝜋 3𝜋
𝑋(3) = 1 + 0.707 [cos ( ) − 𝑗 sin ( )]
2 2
9𝜋 9𝜋
− 0.707 [cos ( ) − 𝑗 sin ( )]
2 2
∴ 𝑿(𝟑) = 𝟏 + 𝒋 𝟏. 𝟒𝟏

∴ 𝑿(𝒌) = {𝟏, 𝟏 − 𝒋 𝟏. 𝟒𝟏𝟒, 𝟏, 𝟏 + 𝒋 𝟏. 𝟒𝟏}

53
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

➢ IDFT Method: Inverse Discrete Fourier Transform

The Inverse Discrete Fourier Transform (IDFT) converts frequency-domain
data back into the time domain, reconstructing the original sampled signal.
𝑵−𝟏
𝟏 𝟐𝝅
𝒙(𝒏) = ∑ 𝑿(𝒌) 𝒆𝒋 𝑵 𝒌𝒏
𝑵
𝒌=𝟎

• Find IDFT of 𝑋(𝑘) = {6, −2 + 2𝑗, −2, −2 − 2𝑗}

 Given that,
𝑿(𝒌) = {𝟔, −𝟐 + 𝟐𝒋, −𝟐, −𝟐 − 𝟐𝒋}
Here,
𝑵=𝟒
We know,
𝑵−𝟏
𝟏 𝟐𝝅
𝒙(𝒏) = ∑ 𝑿(𝒌) 𝒆𝒋 𝑵 𝒌𝒏
𝑵
𝒌=𝟎
4−1
1 2𝜋
𝑗 𝑘𝑛
𝑥 (𝑛 ) = ∑ 𝑋 (𝑘 ) 𝑒 4
4
𝑘=0
3
1 𝜋
𝑗 𝑘𝑛
𝑥 (𝑛 ) = ∑ 𝑋 (𝑘 ) 𝑒 2
4
𝑘=0

1 0
𝜋
𝑗 𝑛 𝑗𝜋𝑛
3𝜋
𝑗 𝑛
𝑥(𝑛) = [𝑋(0)𝑒 + 𝑋(1)𝑒 2 + 𝑋(2)𝑒 + 𝑋(3)𝑒 2 ]
4
1 𝜋
𝑗 𝑛 𝑗𝜋𝑛
3𝜋
𝑗 𝑛
( ) ( )
𝑥 𝑛 = [6 + −2 + 2𝑗 𝑒 2 − 2𝑒 + (−2 − 2𝑗)𝑒 2 ]
4
∴ 𝒙(𝒏)
𝝅 𝝅
𝟔 + (−𝟐 + 𝟐𝒋) {𝐜𝐨𝐬 ( 𝒏) + 𝒋 𝐬𝐢𝐧 ( 𝒏)} − 𝟐{𝐜𝐨𝐬(𝝅𝒏) + 𝒋 𝐬𝐢𝐧(𝝅𝒏)} +
𝟏 𝟐 𝟐
= [ 𝟑𝝅 𝟑𝝅 ]
𝟒
(−𝟐 − 𝟐𝒋) {𝐜𝐨𝐬 ( 𝒏) + 𝒋 𝐬𝐢𝐧 ( 𝒏)}
𝟐 𝟐

54
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

when 𝒏 = 𝟎,
1 6 + (−2 + 2𝑗){cos(0) + 𝑗 sin(0)} − 2{cos(0) + 𝑗 sin(0)} +
𝑥 (0) = [ ]
4 (−2 − 2𝑗){cos(0) + 𝑗 sin(0)}
∴ 𝒙(𝟎) = 𝟎
when 𝒏 = 𝟏,
𝜋 𝜋
6 + (−2 + 2𝑗) {cos ( ) + 𝑗 sin ( )} − 2{cos(𝜋) + 𝑗 sin(𝜋)} +
1 2 2
𝑥(1) = [ 3𝜋 3𝜋 ]
4
(−2 − 2𝑗) {cos ( ) + 𝑗 sin ( )}
2 2
∴ 𝒙(𝟏) = 𝟏
when 𝒏 = 𝟐,
1 6 + (−2 + 2𝑗){cos(𝜋) + 𝑗 sin(𝜋)} − 2{cos(2𝜋) + 𝑗 sin(2𝜋)} +
𝑥 (2) = [ ]
4 (−2 − 2𝑗){cos(3𝜋) + 𝑗 sin(3𝜋)}
∴ 𝒙(𝟐) = 𝟐
when 𝒏 = 𝟑,
𝑥(3)
3𝜋 3𝜋
1 6 + ( −2 + 2𝑗 ) {cos ( ) + 𝑗 sin ( )} − 2{cos(3𝜋) + 𝑗 sin(3𝜋)} +
= [ 2 2 ]
4 9𝜋 9𝜋
(−2 − 2𝑗) {cos ( ) + 𝑗 sin ( )}
2 2
∴ 𝒙(𝟑) = 𝟑

∴ 𝒙(𝒏) = {𝟎, 𝟏, 𝟐, 𝟑}

55
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

• Find IDFT of 𝑋(𝑘) = {3, −𝑗, 1, 𝑗}

 Given that,
𝑿(𝒌) = {𝟑, −𝒋, 𝟏, 𝒋}
Here,
𝑵=𝟒
We know,
𝑵−𝟏
𝟏 𝟐𝝅
𝒙(𝒏) = ∑ 𝑿(𝒌) 𝒆𝒋 𝑵 𝒌𝒏
𝑵
𝒌=𝟎
4−1
1 2𝜋
𝑗 𝑘𝑛
𝑥 (𝑛 ) = ∑ 𝑋 (𝑘 ) 𝑒 4
4
𝑘=0
3
1 𝜋
𝑗 𝑘𝑛
𝑥 (𝑛 ) = ∑ 𝑋 (𝑘 ) 𝑒 2
4
𝑘=0

1 0
𝜋
𝑗 𝑛 𝑗𝜋𝑛
3𝜋
𝑗 𝑛
𝑥(𝑛) = [𝑋(0)𝑒 + 𝑋(1)𝑒 2 + 𝑋(2)𝑒 + 𝑋(3)𝑒 2 ]
4
1 𝜋
𝑗 𝑛 𝑗𝜋𝑛
3𝜋
𝑗 𝑛
( )
𝑥 𝑛 = [3 − 𝑗𝑒 2 + 1𝑒 + 𝑗𝑒 2 ]
4
∴ 𝒙(𝒏)
𝝅 𝝅
𝟏 𝟑 − 𝒋 {𝐜𝐨𝐬 ( 𝒏) + 𝒋 𝐬𝐢𝐧 ( 𝒏)} + 𝟏{𝐜𝐨𝐬(𝝅𝒏) + 𝒋 𝐬𝐢𝐧(𝝅𝒏)} +
𝟐 𝟐
= [ 𝟑𝝅 𝟑𝝅 ]
𝟒
𝒋 {𝐜𝐨𝐬 ( 𝒏) + 𝒋 𝐬𝐢𝐧 ( 𝒏)}
𝟐 𝟐

when 𝒏 = 𝟎,
1 3 − 𝑗{cos(0) + 𝑗 sin(0)} + {cos(0) + 𝑗 sin(0)} +
𝑥(0) = [ ]
4 𝑗{cos(0) + 𝑗 sin(0)}
∴ 𝒙(𝟎) = 𝟏

56
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

when 𝒏 = 𝟏,
𝜋 𝜋
1 3 − 𝑗 {cos ( ) + 𝑗 sin ( )} + {cos(𝜋) + 𝑗 sin(𝜋)} +
2 2
𝑥(1) = [ 3𝜋 3𝜋 ]
4
𝑗 {cos ( ) + 𝑗 sin ( )}
2 2
∴ 𝒙(𝟏) = 𝟏
when 𝒏 = 𝟐,
1 3 − 𝑗{cos(𝜋) + 𝑗 sin(𝜋)} + {cos(2𝜋) + 𝑗 sin(2𝜋)} +
𝑥 (2) = [ ]
4 𝑗{cos(3𝜋) + 𝑗 sin(3𝜋)}
∴ 𝒙(𝟐) = 𝟏
when 𝒏 = 𝟑,
3𝜋 3𝜋
1 3 − 𝑗 {cos ( ) + 𝑗 sin ( )} + {cos(3𝜋) + 𝑗 sin(3𝜋)} +
𝑥(3) = [ 2 2 ]
4 9𝜋 9𝜋
𝑗 {cos ( ) + 𝑗 sin ( )}
2 2
∴ 𝒙(𝟑) = 𝟎

∴ 𝒙(𝒏) = {𝟏, 𝟏, 𝟏, 𝟎}

• Example-6.11: Find IDFT of 𝑋(𝑘) = {1,2,3,4}

 Given that,
𝑿(𝒌) = {𝟏, 𝟐, 𝟑, 𝟒}
Here,
𝑵=𝟒
We know,
𝑵−𝟏
𝟏 𝟐𝝅
𝒙(𝒏) = ∑ 𝑿(𝒌) 𝒆𝒋 𝑵 𝒌𝒏
𝑵
𝒌=𝟎

57
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03
4−1
1 2𝜋
𝑥 (𝑛) = ∑ 𝑋(𝑘) 𝑒 𝑗 4 𝑘𝑛
4
𝑘=0
3
1 𝜋
𝑥 (𝑛) = ∑ 𝑋(𝑘) 𝑒 𝑗 2 𝑘𝑛
4
𝑘=0

1 𝜋 3𝜋
𝑥(𝑛) = [𝑋(0)𝑒 0 + 𝑋(1)𝑒 𝑗 2 𝑛 + 𝑋(2)𝑒 𝑗𝜋𝑛 + 𝑋(3)𝑒 𝑗 2 𝑛 ]
4
𝟏 𝝅 𝟑𝝅
∴ 𝒙(𝒏) = [𝟏 + 𝟐 𝒆𝒋𝟐 𝒏 + 𝟑 𝒆𝒋𝝅𝒏 + 𝟒 𝒆𝒋 𝟐 𝒏 ]
𝟒
when 𝒏 = 𝟎,
𝟏
𝒙 (𝟎 ) = [𝟏 + 𝟐 𝒆𝟎 + 𝟑 𝒆𝟎 + 𝟒 𝒆𝟎 ]
𝟒
1
𝑥 (0) = [1 + 2 + 3 + 4 ]
4
∴ 𝒙(𝟎) = 𝟐. 𝟓
when 𝒏 = 𝟏,
𝟏 𝝅 𝟑𝝅
𝒙(𝟏) = [𝟏 + 𝟐 𝒆𝒋𝟐 + 𝟑 𝒆𝒋𝝅 + 𝟒 𝒆𝒋 𝟐 ]
𝟒
𝜋 𝜋
1 1 + 2 {cos ( ) + 𝑗 sin ( )} + 3{cos(𝜋) + 𝑗 sin(𝜋)} +
2 2
𝑥(1) = [ 3𝜋 3𝜋 ]
4
4 {cos ( ) + 𝑗 sin ( )}
2 2
∴ 𝒙(𝟏) = −𝟎. 𝟓 − 𝒋 𝟎. 𝟓
when 𝒏 = 𝟐,
𝟏
𝒙(𝟐) = [𝟏 + 𝟐 𝒆𝒋𝝅 + 𝟑 𝒆𝒋𝟐𝝅 + 𝟒 𝒆𝒋𝟑𝝅 ]𝒙(𝒏)
𝟒
1 1 + 2{cos(𝜋) + 𝑗 sin(𝜋)} + 3{cos(2𝜋) + 𝑗 sin(2𝜋)} +
𝑥(2) = [ ]
4 4{cos(3𝜋) + 𝑗 sin(3𝜋)}

∴ 𝒙(𝟐) = −𝟎. 𝟓

58
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

when 𝒏 = 𝟑,
𝟏 𝟑𝝅 𝟗𝝅
𝒙(𝟑) = [𝟏 + 𝟐 𝒆𝒋 𝟐 + 𝟑 𝒆𝒋𝟑𝝅 + 𝟒 𝒆𝒋 𝟐 ]
𝟒
3𝜋 3𝜋
1 1 + 2 {cos ( ) + 𝑗 sin ( )} + 3{cos(3𝜋) + 𝑗 sin(3𝜋)} +
𝑥(3) = [ 2 2 ]
4 9𝜋 9𝜋
4 {cos ( ) + 𝑗 sin ( )}
2 2
∴ 𝒙(𝟑) = −𝟎. 𝟓 + 𝒋 𝟎. 𝟓

∴ 𝒙(𝒏) = {𝟐. 𝟓, −𝟎. 𝟓 − 𝒋 𝟎. 𝟓, −𝟎. 𝟓, −𝟎. 𝟓 + 𝒋 𝟎. 𝟓}

➢ DFS: Discrete Fourier Series

The Discrete Fourier Series (DFS) represents a periodic discrete signal as a
sum of sinusoidal components.
For continuous time signal,
∞

𝒙(𝒕) = 𝒂𝟎 + ∑ (𝐜𝐨𝐬 𝒏𝝎𝟎 𝒕 + 𝐬𝐢𝐧 𝒏𝝎𝟎 𝒕)

𝒏=𝟏

For discrete time signal,

∞

𝒙(𝒏) = ∑ (𝒂𝒌 𝒆𝒋𝝎𝟎𝒏 )

𝑲=𝑵
∞
𝟐𝝅
𝒙(𝒏) = ∑ (𝒂𝒌 𝒆𝒋 𝑵 𝒌𝒏 )
𝑲=𝑵

59
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

• Problem- 14.10: Find the DFS co - efficient for

2𝜋 2𝜋 6𝜋 8𝜋 𝜋
𝑥(𝑛) = 1 + sin ( ) 𝑛 + cos ( ) 𝑛 + 2 cos ( ) 𝑛 + 3 sin ( 𝑛 + )
𝑁 𝑁 𝑁 𝑁 3
 Given that,
𝟐𝝅 𝟐𝝅 𝟔𝝅
𝒙(𝒏) = 𝟏 + 𝐬𝐢𝐧 ( ) 𝒏 + 𝐜𝐨𝐬 ( ) 𝒏 + 𝟐 𝐜𝐨𝐬 ( ) 𝒏
𝑵 𝑵 𝑵
𝟖𝝅 𝝅
+ 𝟑 𝐬𝐢𝐧 ( 𝒏 + )
𝑵 𝟑
2𝜋 2𝜋 2𝜋 2𝜋 6𝜋 6𝜋
𝑒 𝑗 𝑁 𝑛 − 𝑒 −𝑗 𝑁 𝑛 𝑒 𝑗 𝑁 𝑛 + 𝑒 −𝑗 𝑁 𝑛 𝑒 𝑗 𝑁 𝑛 + 𝑒 −𝑗 𝑁 𝑛
𝑥 (𝑛 ) = 1 + + + 2[ ]
2𝑗 2 2
8𝜋 𝜋 8𝜋 𝜋
𝑒 𝑗( 𝑁 𝑛+ 3 ) − 𝑒 −𝑗( 𝑁 𝑛+ 3 )
+ 3[ ]
2𝑗

1 𝑗2𝜋𝑛 1 −𝑗2𝜋𝑛 1 𝑗2𝜋𝑛 1 −𝑗2𝜋𝑛 6𝜋 6𝜋

𝑥 (𝑛 ) = 1 + 𝑒 𝑁 − 𝑒 𝑁 + 𝑒 𝑁 + 𝑒 𝑁 + 𝑒 𝑗 𝑁 𝑛 + 𝑒 −𝑗 𝑁 𝑛
2𝑗 2𝑗 2 2
3 𝑗(8𝜋𝑛+𝜋) 8𝜋 𝜋
−𝑗( 𝑛+ )
+ (𝑒 𝑁 3 −𝑒 𝑁 3 )
2𝑗
1 1
2𝜋 2𝜋 1 1 2𝜋 2𝜋
𝑥 (𝑛 ) = 1 + 𝑒 𝑗 𝑁 𝑛 ( + ) + 𝑒 −𝑗 𝑁 𝑛 ( − ) + 𝑒 𝑗3 𝑁 𝑛 + 𝑒 −𝑗3 𝑁 𝑛
2𝑗 2 2 2𝑗
3 𝑗4 𝑛 𝑗
2𝜋 𝜋 3 −𝑗42𝜋𝑛 −𝑗𝜋
+ (𝑒 𝑁 . 𝑒 ) − (𝑒 𝑁 . 𝑒 3 )
3
2𝑗 2𝑗
Re arranging the sequence,
𝟐𝝅 𝝅 𝟐𝝅 𝟐𝝅
𝟑 𝟏 𝟏
∴ 𝒙(𝒏) = − (𝒆−𝒋𝟒 𝑵 𝒏 . 𝒆−𝒋𝟑 ) + 𝒆−𝒋𝟑 𝑵 𝒏 + 𝒆−𝒋 𝑵 𝒏 (𝟐 − 𝟐𝒋) + 𝟏 +
𝟐𝒋

𝟐𝝅 𝟏 𝟏 𝟐𝝅 𝟑 𝟐𝝅 𝝅
𝒆𝒋 𝑵 𝒏 ( + ) + 𝒆𝒋𝟑 𝑵 𝒏 + (𝒆𝒋𝟒 𝑵 𝒏 . 𝒆𝒋 𝟑 )
𝟐𝒋 𝟐 𝟐𝒋
Taking co-efficient,
𝟑 −𝒋𝝅
𝒂−𝟒 = − (𝒆 𝟑 )
𝟐𝒋
𝒂−𝟑 = 𝟏

60
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

𝒂−𝟐 = 𝟎
𝟏 𝟏
𝒂−𝟏 = −
𝟐 𝟐𝒋
𝒂𝟎 = 𝟏
𝟏 𝟏
𝒂𝟏 = +
𝟐 𝟐𝒋
𝒂𝟐 = 𝟎
𝒂𝟑 = 𝟏
𝟑 𝒋𝝅
𝒂𝟒 = (𝒆 𝟑 )
𝟐𝒋

• Problem- 14.10: Find the DFS co - efficient for

𝜋
𝑥 (𝑛) = 1 + sin 2𝜔0 𝑛 + 2 cos 2𝜔0 𝑛 + 3 cos (3𝜔0 𝑛 + )
3
 Given that,
𝝅
𝒙(𝒏) = 𝟏 + 𝐬𝐢𝐧 𝟐𝝎𝟎 𝒏 + 𝟐 𝐜𝐨𝐬 𝟐𝝎𝟎 𝒏 + 𝟑 𝐜𝐨𝐬 (𝟑𝝎𝟎 𝒏 + )
𝟑
𝑒 𝑗2𝜔0 𝑛 − 𝑒 −𝑗2𝜔0𝑛 𝑒 𝑗2𝜔0 𝑛 + 𝑒 −𝑗2𝜔0𝑛
𝑥 (𝑛 ) = 1 + + 2[ ]
2𝑗 2
𝜋 𝜋
𝑗(3𝜔0 𝑛+ ) −𝑗(3𝜔0 𝑛+ )
𝑒 3 + 𝑒 3
+ 3[ ]
2

1 𝑗2𝜔 𝑛
𝑥 (𝑛 ) = 1 +(𝑒 0 − 𝑒 −𝑗2𝜔0𝑛 ) + 𝑒 𝑗2𝜔0 𝑛 + 𝑒 −𝑗2𝜔0𝑛
2𝑗
3 𝜋 𝜋
+ [𝑒 𝑗(3𝜔0𝑛+ 3 ) + 𝑒 −𝑗(3𝜔0 𝑛+ 3 ) ]
2
1 1 3 𝜋
𝑥(𝑛) = 1 + 𝑒 𝑗2𝜔0 𝑛 − 𝑒 −𝑗2𝜔0𝑛 + 𝑒 𝑗2𝜔0𝑛 + 𝑒 −𝑗2𝜔0𝑛 + 𝑒 𝑗(3𝜔0𝑛+ 3 )
2𝑗 2𝑗 2
3 𝜋
+ 𝑒 −𝑗(3𝜔0𝑛+ 3 )
2

61
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

1 1 3 𝜋
𝑥 (𝑛) = 1 + 𝑒 𝑗2𝜔0 𝑛 [1 + ] + 𝑒 −𝑗2𝜔0𝑛 [1 − ] + (𝑒 𝑗3𝜔0 𝑛 . 𝑒 𝑗 3 )
2𝑗 2𝑗 2
3 𝜋
+ (𝑒 −𝑗3𝜔0 𝑛 . 𝑒 −𝑗 3 )
2
Re arranging the sequence,
𝟑 −𝒋𝟑𝝎 𝒏 −𝒋𝝅 𝟏 𝟏
∴ 𝒙 (𝒏) = (𝒆 𝟎 . 𝒆 𝟑 ) + 𝒆−𝒋𝟐𝝎𝟎 𝒏 [𝟏 − ] + 𝟏 + 𝒆𝒋𝟐𝝎𝟎𝒏 [𝟏 + ]
𝟐 𝟐𝒋 𝟐𝒋
𝟑 𝝅
+ (𝒆𝒋𝟑𝝎𝟎 𝒏 . 𝒆𝒋 𝟑 )
𝟐
Taking co-efficient,
𝟑 −𝒋𝝅
𝒂−𝟑 = (𝒆 𝟑 )
𝟐
𝟏
𝒂−𝟐 =𝟏−
𝟐𝒋
𝒂−𝟏 = 𝟎
𝒂𝟎 = 𝟏
𝒂𝟏 = 𝟎
𝟏
𝒂𝟐 = 𝟏 +
𝟐𝒋
𝟑 𝒋𝝅
𝒂𝟑 = (𝒆 𝟑 )
𝟐

➢ DTFT Method: Discrete Time Fourier Transform

The Discrete Time Fourier Transform (DTFT) is a mathematical transform
that converts a discrete-time signal into a continuous function in the
frequency domain.
∞

𝑿(𝒆𝒋𝝎 ) = ∑ 𝒙(𝒏)𝒆−𝒋𝝎𝒏
𝒏=−∞

62
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

• Find DTFT of 𝑥(𝑛) = 𝑎𝑛 𝑢(𝑛) and its magnitude.

 Given that,
𝒙(𝒏) = 𝒂𝒏 𝒖(𝒏)
We know,
∞

𝑿(𝒆𝒋𝝎 ) = ∑ 𝒙(𝒏)𝒆−𝒋𝝎𝒏
𝒏=−∞
∞

𝑿(𝒆𝒋𝝎 ) = ∑[𝑎𝑛 𝑢(𝑛)] 𝑒 −𝑗𝜔𝑛

𝑛=0
∞

𝑿(𝒆𝒋𝝎 ) = ∑ 𝑎𝑛 𝑒 −𝑗𝜔𝑛
𝑛=0
∞

𝑿(𝒆𝒋𝝎 ) = ∑ (𝑎 𝑒 −𝑗𝜔 )𝑛
𝑛=0

1
𝑿(𝒆𝒋𝝎 ) =
1 − 𝑎 𝑒 −𝑗𝜔
1
𝑿(𝒆𝒋𝝎 ) =
1 − 𝑎[cos 𝜔 − 𝑗 sin 𝜔]
𝟏
∴ 𝑿(𝒆𝒋𝝎 ) =
𝟏 − 𝒂 𝐜𝐨𝐬 𝝎 + 𝒋 𝒂 𝐬𝐢𝐧 𝝎
Magnitude,
𝟏
|𝑿(𝒆𝒋𝝎 )| =
√(𝟏 − 𝒂 𝐜𝐨𝐬 𝝎)𝟐 + (𝒂 𝐬𝐢𝐧 𝝎)𝟐

• Find DTFT of 𝑥(𝑛) = 𝑎|𝑛| and its magnitude.

 Given that,
𝒙(𝒏) = 𝒂|𝒏|

63
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

We know,
∞

𝑿(𝒆𝒋𝝎 ) = ∑ 𝒙(𝒏)𝒆−𝒋𝝎𝒏
𝒏=−∞
∞

𝑿(𝒆𝒋𝝎 ) = ∑ 𝑎|𝑛| 𝑒 −𝑗𝜔𝑛

𝑛=−∞
−1 ∞

𝑿(𝒆𝒋𝝎 ) = ∑ 𝑎−𝑛 𝑒 −𝑗𝜔𝑛 + ∑ 𝑎𝑛 𝑒 −𝑗𝜔𝑛

𝑛=−∞ 𝑛=0
−1 ∞

𝑿(𝒆𝒋𝝎 ) = ∑ (𝑎−1 𝑒 −𝑗𝜔 )𝑛 + ∑(𝑎 𝑒 −𝑗𝜔 )𝑛

𝑛=−∞ 𝑛=0
∞ ∞

𝑿(𝒆𝒋𝝎 ) = ∑(𝑎 𝑒 𝑗𝜔 )𝑛 + ∑(𝑎 𝑒 −𝑗𝜔 )𝑛

𝑛=1 𝑛=0

𝒋𝝎 )
𝑎 𝑒 𝑗𝜔 1
𝑿(𝒆 = +
1 − 𝑎 𝑒 𝑗𝜔 1 − 𝑎 𝑒 −𝑗𝜔

𝒋𝝎 )
𝑎 𝑒 𝑗𝜔 (1 − 𝑎 𝑒 −𝑗𝜔 ) + 1 − 𝑎 𝑒 𝑗𝜔
𝑿 (𝒆 =
(1 − 𝑎 𝑒 𝑗𝜔 )(1 − 𝑎 𝑒 −𝑗𝜔 )

𝒋𝝎 )
𝑎 𝑒 𝑗𝜔 − 𝑎2 + 1 − 𝑎 𝑒 𝑗𝜔
𝑿(𝒆 =
1 − 𝑎 𝑒 −𝑗𝜔 − 𝑎 𝑒 𝑗𝜔 + 𝑎2

𝒋𝝎 )
1 − 𝑎2
𝑿 (𝒆 =
1 − 𝑎 [cos 𝜔 − 𝑗 sin 𝜔] − 𝑎 [cos 𝜔 + 𝑗 sin 𝜔] + 𝑎2

𝒋𝝎 )
1 − 𝑎2
𝑿 (𝒆 =
1 − 𝑎 cos 𝜔 + 𝑗 𝑎 sin 𝜔 − 𝑎 cos 𝜔 − 𝑗 𝑎 sin 𝜔 + 𝑎2

𝒋𝝎 )
𝟏 − 𝒂𝟐
∴ 𝑿( 𝒆 =
𝟏 − 𝟐 𝒂 𝐜𝐨𝐬 𝝎 + 𝒂𝟐
Magnitude,

𝒋𝝎 )
√(𝟏 − 𝒂𝟐 )𝟐
|𝑿(𝒆 |=
√(𝟏 − 𝟐 𝒂 𝐜𝐨𝐬 𝝎 + 𝒂𝟐 )𝟐

64
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

➢ FFT: Fast Fourier Transform

DIT- Decimation in Time
Light to Heavy
Calculator Formula for Light to Heavy:
𝟐×𝟏𝟖𝟎𝑴
𝒚=𝟏<− : 𝑨 = 𝑪 + 𝑫𝒚: 𝑩 = 𝑪 − 𝑫𝒚
𝑭

DIF- Decimation in Frequency

Heavy to Light
Calculator Formula for Heavy to Light:
𝟐×𝟏𝟖𝟎𝑴
𝒚=𝟏<− : 𝑨 = 𝑪 + 𝑫: 𝑩 = (𝑪 − 𝑫)𝒚
𝑭

𝑵−𝟏 𝑵−𝟏
𝟐𝝅
𝑿(𝒌) = ∑ 𝒙(𝒏) 𝒆−𝒋 𝑵 𝒌𝒏 = ∑ 𝒙(𝒏) 𝝎𝒌𝒏
𝑵
𝒏=𝟎 𝒏=𝟎

Here,
𝟐𝝅
−𝒋 𝒌
Twiddle Factor, 𝝎𝒌𝑵 =𝒆 𝑵

𝟐𝝅×𝟎
−𝒋
𝝎𝟎𝟐 = 𝒆 𝟐 =𝟏
𝟐𝝅×𝟎
𝝎𝟎𝟒 = 𝒆−𝒋 𝟒 =𝟏
𝟐𝝅×𝟏
𝝎𝟏𝟒 = 𝒆−𝒋 𝟒 = −𝒋
𝟐𝝅×𝟎
𝝎𝟎𝟖 = 𝒆−𝒋 𝟖 =𝟏
𝟐𝝅×𝟏
𝝎𝟏𝟖 = 𝒆−𝒋 𝟖 = 𝟎. 𝟕𝟎𝟕 − 𝒋 𝟎. 𝟕𝟎𝟕
𝟐𝝅×𝟐
−𝒋
𝝎𝟐𝟖 = 𝒆 𝟖 = −𝒋
𝟐𝝅×𝟑
−𝒋
𝝎𝟑𝟖 = 𝒆 𝟖 = −𝟎. 𝟕𝟎𝟕 − 𝒋 𝟎. 𝟕𝟎𝟕

65
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

➢ Limitations of Discrete Fourier Transform (DFT):

1. Computational Complexity: The primary limitation of the DFT is its
computational cost. A direct computation of the DFT requires N data
points, making it inefficient for large datasets.
2. Resolution Issues: DFT assumes that the input signal is periodic and its
frequency resolution depends on the length of the input signal.
3. Spectral Leakage: When a signal is reduced, it may result in spectral
leakage, where energy from one frequency "leaks" into other
frequencies, making the frequency spectrum less accurate.
4. Aliasing: DFT assumes the sampled signal is band-limited. If the
sampling rate is too low (below the Nyquist rate), higher frequencies
may be misrepresented as lower frequencies, a phenomenon called
aliasing.
5. Real-Time Limitations: Due to the computational complexity, DFT can
be slow and impractical for real-time applications where rapid or
continuous Fourier transformations are required.

➢ Advantages of Fast Fourier Transform (FFT)

1. Efficiency: The primary advantage of FFT is that it reduces the
computational complexity and making it much faster especially for large
datasets.
2. Suitability for Real-Time Processing: Due to its computational
efficiency, FFT is suitable for real-time signal processing applications
like audio, radar, and communication systems.
3. Widely Used Algorithms: FFT is extensively used in scientific
computing, signal processing and data analysis due to its speed and the
availability of optimized libraries in many programming environments.
4. Memory Efficiency: FFT implementations often use in-place
algorithms, meaning they can transform data without requiring additional
memory, improving memory usage efficiency.

66
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

5. Handling Large Data Sets: FFT can handle large data sets more
effectively than DFT, making it ideal for applications like image
processing, where large amounts of data need to be transformed quickly.

• Find DFT-FFT using DIT of 𝑥(𝑛) = {1,2,3,4,4,3,2,1}

 Given that,
𝒙(𝒏) = {𝟏, 𝟐, 𝟑, 𝟒, 𝟒, 𝟑, 𝟐, 𝟏}
Here,
𝑵=𝟖
𝒙(𝟎) = 𝟏
𝒙(𝟏) = 𝟐
𝒙(𝟐) = 𝟑
𝒙(𝟑) = 𝟒
𝒙(𝟒) = 𝟒
𝒙(𝟓) = 𝟑
𝒙(𝟔) = 𝟐
𝒙(𝟕) = 𝟏

𝑿(𝒌) 𝒙(𝒏)
𝟎 𝟎𝟎𝟎 𝟎𝟎𝟎 𝟎
𝟏 𝟎𝟎𝟏 𝟏𝟎𝟎 𝟒
𝟐 𝟎𝟏𝟎 𝟎𝟏𝟎 𝟐
𝟑 𝟎𝟏𝟏 𝟏𝟏𝟎 𝟔
𝟒 𝟏𝟎𝟎 𝟎𝟎𝟏 𝟏
𝟓 𝟏𝟎𝟏 𝟏𝟎𝟏 𝟓
𝟔 𝟏𝟏𝟎 𝟎𝟏𝟏 𝟑
𝟕 𝟏𝟏𝟏 𝟏𝟏𝟏 𝟕

67
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

Now,
𝑨𝟏 = 𝒙(𝟎) + 𝒙(𝟒) ∗ 𝝎𝟎𝟐 = 𝟏 + 𝟒 = 𝟓
𝑨𝟐 = 𝒙(𝟎) − 𝒙(𝟒) ∗ 𝝎𝟎𝟐 = 𝟏 − 𝟒 = −𝟑
𝑨𝟑 = 𝒙(𝟐) + 𝒙(𝟔) ∗ 𝝎𝟎𝟐 = 𝟑 + 𝟐 = 𝟓
𝑨𝟒 = 𝒙(𝟐) − 𝒙(𝟔) ∗ 𝝎𝟎𝟐 = 𝟑 − 𝟐 = 𝟏
𝑨𝟓 = 𝒙(𝟏) + 𝒙(𝟓) ∗ 𝝎𝟎𝟐 = 𝟐 + 𝟑 = 𝟓
𝑨𝟔 = 𝒙(𝟏) − 𝒙(𝟓) ∗ 𝝎𝟎𝟐 = 𝟐 − 𝟑 = −𝟏
𝑨𝟕 = 𝒙(𝟑) + 𝒙(𝟕) ∗ 𝝎𝟎𝟐 = 𝟒 + 𝟏 = 𝟓
𝑨𝟖 = 𝒙(𝟑) − 𝒙(𝟕) ∗ 𝝎𝟎𝟐 = 𝟒 − 𝟏 = 𝟑

68
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

Now,
𝑩𝟏 = 𝑨𝟏 + 𝑨𝟑 ∗ 𝝎𝟎𝟒 = 𝟓 + 𝟓 = 𝟏𝟎
𝑩𝟐 = 𝑨𝟐 + 𝑨𝟒 ∗ 𝝎𝟏𝟒 = −𝟑 + 𝟏(−𝒋) = −𝟑 − 𝒋
𝑩𝟑 = 𝑨𝟏 − 𝑨𝟑 ∗ 𝝎𝟎𝟒 = 𝟓 − 𝟓 = 𝟎
𝑩𝟒 = 𝑨𝟐 − 𝑨𝟒 ∗ 𝝎𝟏𝟒 = −𝟑 − (−𝒋) = −𝟑 + 𝒋
𝑩𝟓 = 𝑨𝟓 + 𝑨𝟕 ∗ 𝝎𝟎𝟒 = 𝟓 + 𝟓 = 𝟏𝟎
𝑩𝟔 = 𝑨𝟔 + 𝑨𝟖 ∗ 𝝎𝟏𝟒 = −𝟏 + 𝟑(−𝒋) = −𝟏 − 𝒋 𝟑
𝑩𝟕 = 𝑨𝟓 − 𝑨𝟕 ∗ 𝝎𝟎𝟒 = 𝟓 − 𝟓 = 𝟎
𝑩𝟖 = 𝑨𝟔 + 𝑨𝟖 ∗ 𝝎𝟏𝟒 = −𝟏 − 𝟑(−𝒋) = −𝟏 + 𝒋 𝟑

Now,
𝑿(𝟎) = 𝑩𝟏 + 𝑩𝟓 ∗ 𝝎𝟎𝟖 = 𝟏𝟎 + 𝟏𝟎 = 𝟐𝟎
𝑿(𝟏) = 𝑩𝟐 + 𝑩𝟔 ∗ 𝝎𝟏𝟖 = −𝟑 − 𝒋 + (−𝟏 − 𝒋 𝟑)(𝟎. 𝟕𝟎𝟕 − 𝒋 𝟎. 𝟕𝟎𝟕)
= −𝟓. 𝟖 − 𝒋 𝟐. 𝟒
𝑿(𝟐) = 𝑩𝟑 + 𝑩𝟕 ∗ 𝝎𝟐𝟖 = 𝟎 + 𝟎 ∗ (−𝒋) = 𝟎
𝑿(𝟑) = 𝑩𝟒 + 𝑩𝟖 ∗ 𝝎𝟑𝟖 = −𝟑 + 𝒋 + (−𝟏 + 𝒋 𝟑)(−𝟎. 𝟕𝟎𝟕 − 𝒋 𝟎. 𝟕𝟎𝟕)
= −𝟎. 𝟏𝟕 − 𝒋 𝟎. 𝟒𝟏
𝑿(𝟒) = 𝑩𝟏 − 𝑩𝟓 ∗ 𝝎𝟎𝟖 = 𝟏𝟎 − 𝟏𝟎 = 𝟎
𝑿(𝟓) = 𝑩𝟐 − 𝑩𝟔 ∗ 𝝎𝟏𝟖 = −𝟑 − 𝒋 − (−𝟏 − 𝒋 𝟑)(𝟎. 𝟕𝟎𝟕 − 𝒋 𝟎. 𝟕𝟎𝟕)
= −𝟎. 𝟏𝟕 + 𝒋 𝟎. 𝟒𝟏
𝑿(𝟔) = 𝑩𝟑 − 𝑩𝟕 ∗ 𝝎𝟐𝟖 = 𝟎 − 𝟎 ∗ (−𝒋) = 𝟎
𝑿(𝟕) = 𝑩𝟒 − 𝑩𝟖 ∗ 𝝎𝟑𝟖 = −𝟑 + 𝒋 − (−𝟏 + 𝒋 𝟑)(−𝟎. 𝟕𝟎𝟕 − 𝒋 𝟎. 𝟕𝟎𝟕)
= −𝟓. 𝟖 + 𝒋 𝟐. 𝟒
∴ 𝑿(𝒌) = {𝟐𝟎, −𝟓. 𝟖 − 𝒋 𝟐. 𝟒, 𝟎, −𝟎. 𝟏𝟕 − 𝒋 𝟎. 𝟒𝟏, 𝟎, −𝟎. 𝟏𝟕
+ 𝒋 𝟎. 𝟒𝟏, 𝟎, −𝟓. 𝟖 + 𝒋 𝟐. 𝟒}

69
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

• Find DFT-FFT using DIF of 𝑥(𝑛) = {−1,2, −3,4,9, −20,12,6}

 Given that,
𝒙(𝒏) = {−𝟏, 𝟐, −𝟑, 𝟒, 𝟗, −𝟐𝟎, 𝟏𝟐, 𝟔}
Here,
𝑵=𝟖
𝒙(𝟎) = −𝟏
𝒙(𝟏) = 𝟐
𝒙(𝟐) = −𝟑
𝒙(𝟑) = 𝟒
𝒙(𝟒) = 𝟗
𝒙(𝟓) = −𝟐𝟎
𝒙(𝟔) = 𝟏𝟐
𝒙(𝟕) = 𝟔

𝒙(𝒏) 𝑿(𝒌)
𝟎 𝟎𝟎𝟎 𝟎𝟎𝟎 𝟎
𝟏 𝟎𝟎𝟏 𝟏𝟎𝟎 𝟒
𝟐 𝟎𝟏𝟎 𝟎𝟏𝟎 𝟐
𝟑 𝟎𝟏𝟏 𝟏𝟏𝟎 𝟔
𝟒 𝟏𝟎𝟎 𝟎𝟎𝟏 𝟏
𝟓 𝟏𝟎𝟏 𝟏𝟎𝟏 𝟓
𝟔 𝟏𝟏𝟎 𝟎𝟏𝟏 𝟑
𝟕 𝟏𝟏𝟏 𝟏𝟏𝟏 𝟕

70
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

Now,
𝑨𝟏 = 𝟖
𝑨𝟐 = −𝟏𝟖
𝑨𝟑 = 𝟗
𝑨𝟒 = 𝟏𝟎
𝑨𝟓 = −𝟏𝟎
𝑨𝟔 = 𝟏𝟓. 𝟓𝟓 − 𝒋𝟏𝟓. 𝟓𝟓
𝑨𝟕 = 𝒋𝟏𝟓
𝑨𝟖 = 𝟏. 𝟒𝟏 + 𝒋𝟏. 𝟒𝟏

Now,
𝑩𝟏 = 𝟏𝟕
𝑩𝟐 = −𝟖
𝑩𝟑 = −𝟏
𝑩𝟒 = 𝒋𝟐𝟖

71
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

𝑩𝟓 = −𝟏𝟎 + 𝒋𝟏𝟓
𝑩𝟔 = 𝟏𝟔. 𝟗𝟔 − 𝒋𝟏𝟒. 𝟏𝟒
𝑩𝟕 = −𝟏𝟎 − 𝒋𝟏𝟓
𝑩𝟖 = −𝟏𝟔. 𝟗𝟔 − 𝒋𝟏𝟒. 𝟏𝟒

Now,
𝑿 (𝟎 ) = 𝟗
𝑿(𝟒) = 𝟐𝟓
𝑿(𝟐) = −𝟏 + 𝒋𝟐𝟖
𝑿(𝟔) = −𝟏 − 𝒋𝟐𝟖
𝑿(𝟏) = 𝟔. 𝟗𝟔 + 𝒋𝟎. 𝟖𝟔
𝑿(𝟓) = −𝟐𝟔. 𝟗𝟔 + 𝒋𝟐𝟗. 𝟏𝟒
𝑿(𝟑) = −𝟐𝟔. 𝟗𝟔 − 𝒋𝟐𝟗. 𝟏𝟒
𝑿(𝟕) = 𝟔. 𝟗𝟔 − 𝒋𝟎. 𝟖𝟔
∴ 𝑿(𝒌) = {𝟗, 𝟔. 𝟗𝟔 + 𝒋𝟎. 𝟖𝟔, −𝟏 + 𝒋𝟐𝟖, −𝟐𝟔. 𝟗𝟔 − 𝒋𝟐𝟗. 𝟏𝟒, 𝟐𝟓, −𝟐𝟔. 𝟗𝟔
+ 𝒋𝟐𝟗. 𝟏𝟒, −𝟏 − 𝒋𝟐𝟖, 𝟔. 𝟗𝟔 − 𝒋𝟎. 𝟖𝟔}

72
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

➢ IFFT: Inverse Fast Fourier Transform

DIT- Decimation in Time
Heavy to Light
Calculator Formula for Heavy to Light:
𝟐 × 𝟏𝟖𝟎𝑴
𝒚=𝟏< : 𝑨 = 𝑪 + 𝑫: 𝑩 = (𝑪 − 𝑫)𝒚
𝑭
DIF- Decimation in Frequency
Light to Heavy
Calculator Formula for Light to Heavy:
𝟐 × 𝟏𝟖𝟎𝑴
𝒚=𝟏< : 𝑨 = 𝑪 + 𝑫𝒚: 𝑩 = 𝑪 − 𝑫𝒚
𝑭

𝑵−𝟏 𝑵−𝟏
𝟏 𝟐𝝅 𝟏
𝑿(𝒌) = ∑ 𝒙(𝒏) 𝒆𝒋 𝑵 𝒌𝒏 = ∑ 𝒙(𝒏) 𝝎−𝒌𝒏
𝑵
𝑵 𝑵
𝒏=𝟎 𝒏=𝟎

Here,
𝟐𝝅 𝟐𝝅
−𝒋 (−𝒌)
Twiddle Factor, 𝝎−𝒌
𝑵 =𝒆 𝑵 = 𝒆𝒋 𝑵 𝒌
𝟐𝝅×𝟎
𝒋
𝝎−𝟎
𝟐 =𝒆 𝟐 =𝟏
𝟐𝝅×𝟎
𝒋
𝝎−𝟎
𝟒 =𝒆 𝟒 =𝟏
𝟐𝝅×𝟏
𝒋
𝝎−𝟏
𝟒 =𝒆 𝟒 =𝒋
𝟐𝝅×𝟎
𝒋
𝝎−𝟎
𝟖 = 𝒆 𝟖 =𝟏
𝟐𝝅×𝟏
𝒋
𝝎−𝟏
𝟖 = 𝒆 𝟖 = 𝟎. 𝟕𝟎𝟕 + 𝒋 𝟎. 𝟕𝟎𝟕
𝟐𝝅×𝟐
𝒋
𝝎−𝟐
𝟖 =𝒆 𝟖 =𝒋
𝟐𝝅×𝟑
𝒋
𝝎−𝟑
𝟖 =𝒆
𝟖 = −𝟎. 𝟕𝟎𝟕 + 𝒋 𝟎. 𝟕𝟎𝟕

73
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

• Find IDFT-FFT using DIT of 𝑋(𝑘) = {20, −5.8 − 𝑗 2.4,0, −0.17 −

𝑗 0.41,0, −0.17 + 𝑗 0.41,0, −5.8 + 𝑗 2.4}
 Given that,
𝑿(𝒌) = {𝟐𝟎, −𝟓. 𝟖 − 𝒋 𝟐. 𝟒, 𝟎, −𝟎. 𝟏𝟕 − 𝒋 𝟎. 𝟒𝟏, 𝟎, −𝟎. 𝟏𝟕
+ 𝒋 𝟎. 𝟒𝟏, 𝟎, −𝟓. 𝟖 + 𝒋 𝟐. 𝟒}
Here,
𝑵=𝟖
𝑿(𝟎) = 𝟐𝟎
𝑿(𝟏) = −𝟓. 𝟖 − 𝒋 𝟐. 𝟒
𝑿( 𝟐 ) = 𝟎
𝑿(𝟑) = −𝟎. 𝟏𝟕 − 𝒋 𝟎. 𝟒𝟏
𝑿( 𝟒 ) = 𝟎
𝑿(𝟓) = −𝟎. 𝟏𝟕 + 𝒋 𝟎. 𝟒𝟏
𝑿( 𝟔 ) = 𝟎
𝑿(𝟕) = −𝟓. 𝟖 + 𝒋 𝟐. 𝟒

𝑿(𝒌) 𝒙(𝒏)
𝟎 𝟎𝟎𝟎 𝟎𝟎𝟎 𝟎
𝟏 𝟎𝟎𝟏 𝟏𝟎𝟎 𝟒
𝟐 𝟎𝟏𝟎 𝟎𝟏𝟎 𝟐
𝟑 𝟎𝟏𝟏 𝟏𝟏𝟎 𝟔
𝟒 𝟏𝟎𝟎 𝟎𝟎𝟏 𝟏
𝟓 𝟏𝟎𝟏 𝟏𝟎𝟏 𝟓
𝟔 𝟏𝟏𝟎 𝟎𝟏𝟏 𝟑
𝟕 𝟏𝟏𝟏 𝟏𝟏𝟏 𝟕

74
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

Now,
𝑨𝟏 = 𝟐𝟎
𝑨𝟐 = −𝟓. 𝟗𝟕 − 𝒋𝟐
𝑨𝟑 = 𝟎
𝑨𝟒 = −𝟓. 𝟗𝟕 + 𝒋𝟐
𝑨𝟓 = 𝟐𝟎
𝑨𝟔 = −𝟐 − 𝒋𝟓. 𝟗𝟕
𝑨𝟕 = 𝟎
𝑨𝟖 = −𝟐 + 𝒋𝟓. 𝟗𝟕

Now,
𝑩𝟏 = 𝟐𝟎
𝑩𝟐 = −𝟏𝟐
𝑩𝟑 = 𝟐𝟎
𝑩𝟒 = 𝟒

75
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

𝑩𝟓 = 𝟐𝟎
𝑩𝟔 = −𝟒
𝑩𝟕 = 𝟐𝟎
𝑩𝟖 = 𝟏𝟐

Now,
𝒙( 𝟎 ) = 𝟖
𝒙(𝟒) = 𝟑𝟐
𝒙(𝟐) = 𝟐𝟒
𝒙(𝟔) = 𝟏𝟔
𝒙(𝟏) = 𝟏𝟔
𝒙(𝟓) = 𝟐𝟒
𝒙(𝟑) = 𝟑𝟐
𝒙( 𝟕 ) = 𝟖
𝟏
∴ 𝒙 (𝒏) = {𝟖, 𝟏𝟔, 𝟐𝟒, 𝟑𝟐, 𝟑𝟐, 𝟐𝟒, 𝟏𝟔, 𝟖}
𝟖
∴ 𝒙(𝒏) = {𝟏, 𝟐, 𝟑, 𝟒, 𝟒, 𝟑, 𝟐, 𝟏}

76
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

• Find IDFT-FFT using DIF of 𝑋(𝑘) = {38, −5.828 + 𝑗 6.07, 𝑗6, −0.172 +
𝑗 8.07, −10, −0.172 − 𝑗 8.07, −𝑗6, −5.828 − 𝑗 6.07}
 Given that,
𝑿(𝒌) = {𝟑𝟖, −𝟓. 𝟖𝟐𝟖 + 𝒋 𝟔. 𝟎𝟕, 𝒋𝟔, −𝟎. 𝟏𝟕𝟐 + 𝒋 𝟖. 𝟎𝟕, −𝟏𝟎, −𝟎. 𝟏𝟕𝟐
− 𝒋 𝟖. 𝟎𝟕, −𝒋𝟔, −𝟓. 𝟖𝟐𝟖 − 𝒋 𝟔. 𝟎𝟕}

Here,
𝑵=𝟖
𝑿(𝟎) = 𝟑𝟖
𝑿(𝟏) = −𝟓. 𝟖𝟐𝟖 + 𝒋 𝟔. 𝟎𝟕
𝑿(𝟐) = 𝒋𝟔
𝑿(𝟑) = −𝟎. 𝟏𝟕𝟐 + 𝒋 𝟖. 𝟎𝟕
𝑿(𝟒) = −𝟏𝟎
𝑿(𝟓) = −𝟎. 𝟏𝟕𝟐 − 𝒋 𝟖. 𝟎𝟕
𝑿(𝟔) = −𝒋𝟔
𝑿(𝟕) = −𝟓. 𝟖𝟐𝟖 − 𝒋 𝟔. 𝟎𝟕

𝒙(𝒏) 𝑿(𝒌)
𝟎 𝟎𝟎𝟎 𝟎𝟎𝟎 𝟎
𝟏 𝟎𝟎𝟏 𝟏𝟎𝟎 𝟒
𝟐 𝟎𝟏𝟎 𝟎𝟏𝟎 𝟐
𝟑 𝟎𝟏𝟏 𝟏𝟏𝟎 𝟔
𝟒 𝟏𝟎𝟎 𝟎𝟎𝟏 𝟏
𝟓 𝟏𝟎𝟏 𝟏𝟎𝟏 𝟓
𝟔 𝟏𝟏𝟎 𝟎𝟏𝟏 𝟑
𝟕 𝟏𝟏𝟏 𝟏𝟏𝟏 𝟕

77
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

Now,
𝑨𝟏 = 𝟐𝟖
𝑨𝟐 = 𝟒𝟖
𝑨𝟑 = 𝟎
𝑨𝟒 = 𝒋𝟏𝟐
𝑨𝟓 = −𝟔 − 𝒋𝟐
𝑨𝟔 = −𝟓. 𝟔𝟓𝟔 + 𝒋𝟏𝟒. 𝟏𝟒
𝑨𝟕 = −𝟔 + 𝒋𝟐
𝑨𝟖 = 𝟓. 𝟔𝟓𝟔 + 𝒋𝟏𝟒. 𝟏𝟒

Now,
𝑩𝟏 = 𝟐𝟖
𝑩𝟐 = 𝟑𝟔
𝑩𝟑 = 𝟐𝟖
𝑩𝟒 = 𝟔𝟎
𝑩𝟓 = −𝟏𝟐

78
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

𝑩𝟔 = −𝟏𝟗. 𝟖 + 𝒋𝟏𝟗. 𝟖
𝑩𝟕 = −𝒋𝟒
𝑩𝟖 = 𝟖. 𝟒𝟖𝟒 + 𝒋𝟖. 𝟒𝟖𝟒

Now,
𝒙(𝟎) = 𝟏𝟔
𝒙( 𝟏 ) = 𝟖
𝒙(𝟐) = 𝟑𝟐
𝒙(𝟑) = 𝟒𝟖
𝒙(𝟒) = 𝟒𝟎
𝒙(𝟓) = 𝟔𝟒
𝒙(𝟔) = 𝟐𝟒
𝒙(𝟕) = 𝟕𝟐
𝟏
∴ 𝒙(𝒏) = {𝟏𝟔, 𝟖, 𝟑𝟐, 𝟒𝟖, 𝟒𝟎, 𝟔𝟒, 𝟐𝟒, 𝟕𝟐}
𝟖
∴ 𝒙(𝒏) = {𝟐, 𝟏, 𝟒, 𝟔, 𝟓, 𝟖, 𝟑, 𝟗}

79
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

➢ Digital Filter: A digital filter is a system that performs mathematical

operations on a sampled discrete time signal to reduce or enhance certain
aspects of that signal.
Digital Filters are used for two general purposes:
1. Separation of signals that have been combined.
2. Reconstruction of signals that have been distorted in some way.

➢ Types of Digital Filter:

1. Finite Impulse Filter (FIR)
2. Infinite Impulse Filter (IIR)

➢ Finite Impulse Filter (FIR): FIR system is one which exhibit zero
response outside a finite duration.
The FIR filter is a non-recursive filter in that the output from the filter is
computed by using current and previous inputs.
➢ Infinite Impulse Filter (IIR): IIR system is one which exhibit an impulse
response of infinite duration.
The IIR filter is a recursive filter in which the output form the filter is
computed by using current and previous inputs as well as past outputs.

➢ Difference between FIR and IIR:

FIR IIR
1. Exhibit a finite impulse 1. Exhibit infinite impulse
2. Non-recursive filter 2. Recursive filter
3. No feedback system 3. Requires feedback

80
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

➢ Difference between Analog and Digital Filter:

Analog Filter Digital Filter
1. Processes continuous-time 1. Processes discrete-time signals
signals
2. Use continuous electronic 2. Use algorithms to process
components (resistors, discrete signals, implemented in
capacitors, inductors) in software or digital hardware.
hardware.
3. Difficult to modify once built; 3. Highly flexible; can be
physical changes to components reprogrammed and adjusted
required. dynamically in real-time.
4. Prone to inaccuracies due to 4. More precise and stable, with
component tolerances, aging, consistent performance over
and environmental factors. time.

➢ Basic Filters:
1. Low-Pass Filter (LPF): Allows signals with frequencies lower than a
specified cutoff frequency to pass while attenuating frequencies higher
than the cutoff.

2. High-Pass Filter (HPF): Allows signals with frequencies higher than a

specified cutoff frequency to pass while attenuating frequencies lower
than the cutoff.

81
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

3. Band-Pass Filter (BPF): Passes frequencies within a certain range

(band) and attenuates frequencies outside this range.

4. Band-Stop (Notch) Filter: Attenuates frequencies within a certain range

while passing frequencies outside this range.

82
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

➢ Algorithm to Design an Ideal Filter:

Step-1: Draw Graph
Step-2: Design Impulse Response
𝟏 𝝅
𝒉𝒅 (𝒏) = ∫ 𝑯𝒅 (𝒆𝒋𝝎 )𝒆𝒋𝝎𝒏 𝒅𝝎
𝟐𝝅 −𝝅
Step-3: The FIR Filter Transfer Function
𝑵−𝟏
𝟐

𝑯(𝒛) = 𝒉(𝟎) + ∑ 𝒉(𝒏) [𝒛𝒏 + 𝒛−𝒏 ]

𝒏=𝟏

Step-4: Obtain the Realizable Filter

𝑵−𝟏
̅ (𝒛) = 𝑯(𝒛). 𝒛−(
𝑯 𝟐
)

Step-5: The Impulse Response

i. For Symmetric & Odd
𝑵−𝟏
𝟐
𝑵−𝟏 𝑵−𝟏
̅ (𝒆𝒋𝝎 ) = 𝒉 (
𝑯 ) + 𝟐 ∑ 𝒉( − 𝒏) 𝐜𝐨𝐬 𝝎𝒏
𝟐 𝟐
𝒏=𝟏

▪ Problem-7.1: Design an ideal low filter pass filter using Fourier method
whose designed frequency response is
𝜋 𝜋
1, ≥𝜔≥−
𝑗𝜔
𝐻𝑑 (𝑒 ) = { 3 3
𝜋
0, 𝜋 ≥ 𝜔 ≥
3
Consider 𝑁 = 9
 Given that,
𝝅 𝝅
𝟏, ≥𝝎≥−
𝒋𝝎
𝑯𝒅 ( 𝒆 ) = { 𝟑 𝟑
𝝅
𝟎, 𝝅 ≥ 𝝎 ≥
𝟑

83
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

Step-1: Graph

Step-2: Design Impulse Response

𝟏 𝝅/𝟑
𝒉𝒅 (𝒏) = ∫ 𝑯𝒅 (𝒆𝒋𝝎 )𝒆𝒋𝝎𝒏 𝒅𝝎
𝟐𝝅 −𝝅/𝟑

1 𝜋/3
ℎ𝑑 ( 𝑛 ) = ∫ (1)𝑒 𝑗𝜔𝑛 𝑑𝜔
2𝜋 −𝜋/3

1 𝜋/3 𝑗𝜔𝑛
ℎ 𝑑 (𝑛 ) = ∫ 𝑒 𝑑𝜔
2𝜋 −𝜋/3
𝜋
𝑗𝜔𝑛 3
1 𝑒
ℎ𝑑 (𝑛 ) = [ ]
2𝜋 𝑗𝑛 −𝜋
3
𝜋
1
ℎ 𝑑 (𝑛 ) = [𝑒 ] 3 𝜋
𝑗𝜔𝑛
𝑗2𝑛𝜋 −
3

1 𝑛𝜋 𝑛𝜋
ℎ𝑑 ( 𝑛 ) = [𝑒 𝑗 3 − 𝑒 −𝑗 3 ]
𝑗2𝑛𝜋
𝑛𝜋 𝑛𝜋
1 𝑒 𝑗 3 − 𝑒 −𝑗 3
ℎ𝑑 (𝑛 ) = 2𝑗 [ ]
𝑗2𝑛𝜋 2𝑗

1 𝑛𝜋
ℎ𝑑 (𝑛 ) = [sin ]
𝑛𝜋 3
𝒏𝝅
𝐬𝐢𝐧
∴ 𝒉𝒅 (𝒏) = 𝟑
𝒏𝝅
84
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

Now,
𝑵−𝟏
( ) |𝒏| ≤
𝒉(𝒏) = {𝒉𝒅 𝒏 , 𝟐
= 𝟒; 𝒂𝒔 𝑵 = 𝟗
𝟎, 𝒐𝒕𝒉𝒆𝒓𝒘𝒊𝒔𝒆
When,
𝒏=𝟎
𝒏𝝅
𝐬𝐢𝐧
𝒉(𝟎) = 𝐥𝐢𝐦 𝟑
𝒏→𝟎 𝒏𝝅
𝑛𝜋
1 sin 3
ℎ(0) = lim
𝑛→0 3 1
𝑛𝜋
3
𝑛𝜋 𝐬𝐢𝐧 𝜽
1 sin
ℎ(0) = lim 𝑛𝜋3 𝐥𝐢𝐦
𝜽→𝟎 𝜽
=𝟏
3 𝑛→0
3
1
ℎ(0) = (1)
3
𝟏
∴ 𝒉(𝟎) = = 𝟎. 𝟑𝟑
𝟑
As the graph is symmetrical so, 𝒉(𝒏) = 𝒉(−𝒏)
𝒏 = 𝟏,
𝝅
𝐬𝐢𝐧
∴ 𝒉( 𝟏 ) = 𝟑 = 𝟎. 𝟐𝟕𝟔 = 𝒉(−𝟏)
𝝅
𝒏 = 𝟐,
𝟐𝝅
𝐬𝐢𝐧
∴ 𝒉( 𝟐 ) = 𝟑 = 𝟎. 𝟏𝟑𝟖 = 𝒉(−𝟐)
𝟐𝝅
𝒏 = 𝟑,
𝐬𝐢𝐧 𝝅
∴ 𝒉 (𝟐 ) = = 𝟎 = 𝒉(−𝟑)
𝟑𝝅

85
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

𝒏 = 𝟒,
𝟒𝝅
𝐬𝐢𝐧
∴ 𝒉( 𝟒 ) = 𝟑 = −𝟎. 𝟎𝟔𝟗 = 𝒉(−𝟒)
𝟒𝝅

Step-3: The FIR Filter Transfer Function

𝑵−𝟏
𝟐

𝑯(𝒛) = 𝒉(𝟎) + ∑ 𝒉(𝒏) [𝒛𝒏 + 𝒛−𝒏 ]

𝒏=𝟏
9−1
2

𝐻 (𝑧) = 0.33 + ∑ ℎ(𝑛) [𝑧 𝑛 + 𝑧 −𝑛 ]

𝑛=1
4

𝐻 (𝑧) = 0.33 + ∑ ℎ(𝑛) [𝑧 𝑛 + 𝑧 −𝑛 ]

𝑛=1

𝐻(𝑧) = 0.33 + ℎ(1)[𝑧 1 + 𝑧 −1 ] + ℎ(2)[𝑧 2 + 𝑧 −2 ] + ℎ(3)[𝑧 3 + 𝑧 −3 ]

+ ℎ(4)[𝑧 4 + 𝑧 −4 ]
𝐻(𝑧) = 0.33 + ℎ(1)𝑧 + ℎ(1)𝑧 −1 + ℎ(2)𝑧 2 + ℎ(2)𝑧 −2 + 0 + ℎ(4)𝑧 4
+ ℎ(4)𝑧 −4
∴ 𝑯(𝒛) = 𝟎. 𝟑𝟑 + 𝟎. 𝟐𝟕𝟔𝒛 + 𝟎. 𝟐𝟕𝟔𝒛−𝟏 + 𝟎. 𝟏𝟑𝟖𝒛𝟐 + 𝟎. 𝟏𝟑𝟖𝒛−𝟐 − 𝟎. 𝟎𝟔𝟗𝒛𝟒
− 𝟎. 𝟎𝟔𝟗𝒛−𝟒

Step-4: Realizable Filter

𝑵−𝟏
̅ (𝒛) = 𝑯(𝒛). 𝒛−(
𝑯 𝟐
)

𝟗−𝟏
̅ (𝒛) = 𝑯(𝒛). 𝒛−(
𝑯 𝟐
)

̅ (𝒛) = 𝑯(𝒛). 𝒛−𝟒

𝑯
̅ (𝒛) = {0.33 + 0.276𝑧 + 0.276𝑧 −1 + 0.138𝑧 2 + 0.138𝑧 −2 − 0.069𝑧 4
𝑯
− 0.069𝑧 −4 }𝑧 −4

86
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

̅ (𝒛) = 0.33𝑧 −4 + 0.276𝑧 −3 + 0.276𝑧 −5 + 0.138𝑧 −2 + 0.138𝑧 −6 − 0.069

𝑯
− 0.069𝑧 −8
̅ (𝒛) = −0.069𝑧 −8 + 0.138𝑧 −6 + 0.276𝑧 −5 + 0.33𝑧 −4 + 0.276𝑧 −3
∴𝑯
+ 0.138𝑧 −2 − 0.069
For Symmetric & Odd
𝒉(𝒏) = 𝒉(𝑵 − 𝟏 − 𝒏)
𝒉(𝟎) = 𝒉(𝟗 − 𝟏 − 𝟎) = 𝒉(𝟖) = −𝟎. 𝟎𝟔𝟗
𝒉(𝟏) = 𝒉(𝟗 − 𝟏 − 𝟏) = 𝒉(𝟕) = 𝟎
𝒉(𝟐) = 𝒉(𝟗 − 𝟏 − 𝟐) = 𝒉(𝟔) = 𝟎. 𝟏𝟑𝟖
𝒉(𝟑) = 𝒉(𝟗 − 𝟏 − 𝟑) = 𝒉(𝟓) = 𝟎. 𝟐𝟕𝟔
𝒉(𝟒) = 𝒉(𝟗 − 𝟏 − 𝟒) = 𝒉(𝟒) = 𝟎. 𝟑𝟑

Step-5: The Impulse Response

𝑵−𝟏
𝟐
𝑵−𝟏 𝑵−𝟏
̅ (𝒆𝒋𝝎 ) = 𝒉 (
𝑯 ) + 𝟐 ∑ 𝒉( − 𝒏) 𝐜𝐨𝐬 𝝎𝒏
𝟐 𝟐
𝒏=𝟏
9−1
2
9−1 9−1
̅ (𝑒 𝑗𝜔 ) = ℎ (
𝐻 )+2∑ℎ( − 𝑛) cos 𝜔𝑛
2 2
𝑛=1
4
̅ (𝑒 𝑗𝜔 ) = ℎ(4) + 2 ∑ ℎ (4 − 𝑛) cos 𝜔 𝑛
𝐻
𝑛=1

̅ (𝑒 𝑗𝜔 ) = ℎ(4) + 2ℎ(4 − 1) cos 𝜔 + 2ℎ(4 − 2) cos 2𝜔 + 2ℎ(4 − 3) cos 3𝜔

𝐻
+ 2ℎ(4 − 4) cos 4𝜔
̅ (𝑒 𝑗𝜔 ) = ℎ(4) + 2ℎ(3) cos 𝜔 + 2ℎ(2) cos 2𝜔 + 2ℎ(1) cos 3𝜔
𝐻
+ 2ℎ(0) cos 4𝜔
̅ (𝑒 𝑗𝜔 ) = 0.33 + (2 × 0.276) cos 𝜔 + (2 × 0.138) cos 2𝜔
𝐻
+ (2 × −0.069) cos 4𝜔
̅ (𝒆𝒋𝝎 ) = 𝟎. 𝟑𝟑 + 𝟎. 𝟓𝟓𝟐 𝐜𝐨𝐬 𝝎 + 𝟎. 𝟐𝟕𝟔 𝐜𝐨𝐬 𝟐𝝎 − 𝟎. 𝟏𝟑𝟖 𝐜𝐨𝐬 𝟒𝝎
∴𝑯
87
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

• Problem- 7.2: Design an ideal high pass filter using Fourier method whose
designed frequency response is
𝜋
1, 𝜋 ≥ 𝜔 ≥ −
𝐻𝑑 (𝑒 𝑗𝜔 ) = { 3
0, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
Consider 𝑁 = 9
 Given that,
𝝅
𝟏, 𝝅≥𝝎≥−
𝑯𝒅 (𝒆𝒋𝝎 ) = { 𝟑
𝟎, 𝒐𝒕𝒉𝒆𝒓𝒘𝒊𝒔𝒆

Step-1: Graph (As the filter is ideal so we assume the system is symmetrical)

Step-2: Design Impulse Response

𝝅
𝟏 −𝟑 𝒋𝝎 𝒋𝝎𝒏
𝟏 𝝅
𝒉𝒅 (𝒏) = ∫ 𝑯𝒅 (𝒆 )𝒆 𝒅𝝎 + ∫ 𝑯𝒅 (𝒆𝒋𝝎 )𝒆𝒋𝝎𝒏 𝒅𝝎
𝟐𝝅 −𝝅 𝟐𝝅 𝝅
𝟑
𝜋
1 −3 𝑗𝜔𝑛
1 𝜋
ℎ𝑑 (𝑛 ) = ∫ (1)𝑒 𝑑𝜔 + ∫ (1)𝑒 𝑗𝜔𝑛 𝑑𝜔
2𝜋 −𝜋 2𝜋 𝜋
3
𝜋
1 − 3 𝑗𝜔𝑛 1 𝜋 𝑗𝜔𝑛
ℎ𝑑 (𝑛 ) = ∫ 𝑒 𝑑𝜔 + ∫ 𝑒 𝑑𝜔
2𝜋 −𝜋 2𝜋 𝜋
3
𝜋
𝑗𝜔𝑛 − 3 𝜋
1 𝑒 1 𝑒 𝑗𝜔𝑛
ℎ𝑑 (𝑛 ) = [ ] + [ ]
2𝜋 𝑗𝑛 −𝜋 2𝜋 𝑗𝑛 𝜋
3

88
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03
𝜋
1 − 1
ℎ𝑑 (𝑛 ) = [𝑒 ]−𝜋3 +
𝑗𝜔𝑛 [𝑒 𝑗𝜔𝑛 ]𝜋𝜋
2𝜋𝑗𝑛 2𝜋𝑗𝑛 3

1 𝑛𝜋 1 𝑗𝑛𝜋
ℎ𝑑 (𝑛 ) = [𝑒 −𝑗 3 − 𝑒 −𝑗𝑛𝜋 ] + [𝑒 𝑗𝑛𝜋 − 𝑒 3 ]
2𝜋𝑗𝑛 2𝜋𝑗𝑛
1 −𝑗
𝑛𝜋
−𝑗𝑛𝜋 𝑗𝑛𝜋
𝑗𝑛𝜋
ℎ𝑑 (𝑛 ) = [𝑒 3 −𝑒 +𝑒 −𝑒 3 ]
2𝜋𝑗𝑛
1 𝑗𝑛𝜋 𝑛𝜋
ℎ𝑑 (𝑛 ) = [𝑒 𝑗𝑛𝜋 − 𝑒 −𝑗𝑛𝜋 − 𝑒 3 + 𝑒 −𝑗 3 ]
2𝜋𝑗𝑛
1 1 𝑗𝑛𝜋 𝑛𝜋
ℎ𝑑 (𝑛 ) = [𝑒 𝑗𝑛𝜋 − 𝑒 −𝑗𝑛𝜋 ] − [𝑒 3 − 𝑒 −𝑗 3 ]
2𝜋𝑗𝑛 2𝜋𝑗𝑛
𝑗𝑛𝜋 𝑛𝜋
2𝑗 𝑒 𝑗𝑛𝜋 − 𝑒 −𝑗𝑛𝜋 2𝑗 𝑒 3 − 𝑒 −𝑗 3
ℎ𝑑 (𝑛 ) = [ ]− [ ]
2𝜋𝑗𝑛 2𝑗 2𝜋𝑗𝑛 2𝑗

1 1 𝑛𝜋
ℎ𝑑 (𝑛 ) = [sin 𝑛𝜋] − [sin ]
𝑛𝜋 𝜋𝑛 3
𝟏 𝒏𝝅
∴ 𝒉𝒅 (𝒏) = [𝐬𝐢𝐧 𝒏𝝅 − 𝐬𝐢𝐧 ]
𝒏𝝅 𝟑
Now,
𝑵−𝟏
( ) |𝒏| ≤
𝒉(𝒏) = {𝒉𝒅 𝒏 , 𝟐
= 𝟒; 𝒂𝒔 𝑵 = 𝟗
𝟎, 𝒐𝒕𝒉𝒆𝒓𝒘𝒊𝒔𝒆
When,
𝒏=𝟎
𝟏 𝒏𝝅
𝒉(𝟎) = 𝐥𝐢𝐦 [𝐬𝐢𝐧 𝒏𝝅 − 𝐬𝐢𝐧 ]
𝒏→𝟎 𝒏𝝅 𝟑
𝑛𝜋
sin 𝑛𝜋 sin
ℎ(0) = lim − lim 3
𝑛→0 𝑛𝜋 𝑛→0 𝜋𝑛
𝑛𝜋
sin 𝑛𝜋 1 sin 3
ℎ(0) = lim − lim 𝑛𝜋
𝑛→0 𝑛𝜋 𝑛→0 3
3

89
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

𝑛𝜋
sin 𝑛𝜋 1 sin 𝐬𝐢𝐧 𝜽
ℎ(0) = lim − lim 𝑛𝜋3 𝐥𝐢𝐦 =𝟏
𝑛→0 𝑛𝜋 3 𝑛→0 𝜽→𝟎 𝜽
3
1
ℎ(0) = 1 −
3
∴ 𝒉(𝟎) = 𝟎. 𝟔𝟔𝟕
As the graph is symmetrical so, 𝒉(𝒏) = 𝒉(−𝒏)
𝒏 = 𝟏,
𝟏 𝒏𝝅
∴ 𝒉(𝟏) = [𝐬𝐢𝐧 𝒏𝝅 − 𝐬𝐢𝐧 ] = −𝟎. 𝟐𝟕𝟔 = 𝒉(−𝟏)
𝒏𝝅 𝟑
𝒏 = 𝟐,
𝟏 𝒏𝝅
∴ 𝒉(𝟐) = [𝐬𝐢𝐧 𝒏𝝅 − 𝐬𝐢𝐧 ] = −𝟎. 𝟏𝟑𝟖 = 𝒉(−𝟐)
𝒏𝝅 𝟑
𝒏 = 𝟑,
𝟏 𝒏𝝅
∴ 𝒉( 𝟐 ) = [𝐬𝐢𝐧 𝒏𝝅 − 𝐬𝐢𝐧 ] = 𝟎 = 𝒉(−𝟑)
𝒏𝝅 𝟑
𝒏 = 𝟒,
𝟏 𝒏𝝅
∴ 𝒉(𝟒) = [𝐬𝐢𝐧 𝒏𝝅 − 𝐬𝐢𝐧 ] = 𝟎. 𝟎𝟔𝟗 = 𝒉(−𝟒)
𝒏𝝅 𝟑
Step-3: The FIR Filter Transfer Function
𝑵−𝟏
𝟐

𝑯(𝒛) = 𝒉(𝟎) + ∑ 𝒉(𝒏) [𝒛𝒏 + 𝒛−𝒏 ]

𝒏=𝟏
9−1
2

𝐻 (𝑧) = 0.667 + ∑ ℎ(𝑛) [𝑧 𝑛 + 𝑧 −𝑛 ]

𝑛=1
4

𝐻 (𝑧) = 0.667 + ∑ ℎ(𝑛) [𝑧 𝑛 + 𝑧 −𝑛 ]

𝑛=1

90
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

𝐻(𝑧) = 0.667 + ℎ(1)[𝑧 1 + 𝑧 −1 ] + ℎ(2)[𝑧 2 + 𝑧 −2 ] + ℎ(3)[𝑧 3 + 𝑧 −3 ]

+ ℎ(4)[𝑧 4 + 𝑧 −4 ]
𝐻 (𝑧) = 0.667 + ℎ(1)𝑧 + ℎ(1)𝑧 −1 + ℎ(2)𝑧 2 + ℎ(2)𝑧 −2 + 0 + ℎ(4)𝑧 4
+ ℎ(4)𝑧 −4
∴ 𝑯(𝒛) = 𝟎. 𝟔𝟔𝟕 − 𝟎. 𝟐𝟕𝟔𝒛 − 𝟎. 𝟐𝟕𝟔𝒛−𝟏 − 𝟎. 𝟏𝟑𝟖𝒛𝟐 − 𝟎. 𝟏𝟑𝟖𝒛−𝟐
+ 𝟎. 𝟎𝟔𝟗𝒛𝟒 + 𝟎. 𝟎𝟔𝟗𝒛−𝟒

Step-4: Realizable Filter

𝑵−𝟏
̅ (𝒛) = 𝑯(𝒛). 𝒛−(
𝑯 𝟐
)

𝟗−𝟏
̅ (𝒛) = 𝑯(𝒛). 𝒛−(
𝑯 𝟐
)

̅ (𝒛) = 𝑯(𝒛). 𝒛−𝟒

𝑯
̅ (𝒛) = {0.667 − 0.276𝑧 − 0.276𝑧 −1 − 0.138𝑧 2 − 0.138𝑧 −2 + 0.069𝑧 4
𝑯
+ 0.069𝑧 −4 }𝑧 −4
̅ (𝒛) = 0.667𝑧 −4 − 0.276𝑧 −3 − 0.276𝑧 −5 − 0.138𝑧 −2 − 0.138𝑧 −6 + 0.069
𝑯
+ 0.069𝑧 −8
̅ (𝒛) = 𝟎. 𝟎𝟔𝟗𝒛−𝟖 − 𝟎. 𝟏𝟑𝟖𝒛−𝟔 − 𝟎. 𝟐𝟕𝟔𝒛−𝟓 + 𝟎. 𝟔𝟔𝟕𝒛−𝟒 − 𝟎. 𝟐𝟕𝟔𝒛−𝟑
∴𝑯
− 𝟎. 𝟏𝟑𝟖𝒛−𝟐 + 𝟎. 𝟎𝟔𝟗

For Symmetric & Odd

𝒉(𝒏) = 𝒉(𝑵 − 𝟏 − 𝒏)
𝒉(𝟎) = 𝒉(𝟗 − 𝟏 − 𝟎) = 𝒉(𝟖) = 𝟎. 𝟎𝟔𝟗
𝒉(𝟏) = 𝒉(𝟗 − 𝟏 − 𝟏) = 𝒉(𝟕) = 𝟎
𝒉(𝟐) = 𝒉(𝟗 − 𝟏 − 𝟐) = 𝒉(𝟔) = −𝟎. 𝟏𝟑𝟖
𝒉(𝟑) = 𝒉(𝟗 − 𝟏 − 𝟑) = 𝒉(𝟓) = −𝟎. 𝟐𝟕𝟔
𝒉(𝟒) = 𝒉(𝟗 − 𝟏 − 𝟒) = 𝒉(𝟒) = 𝟎. 𝟔𝟔𝟕

91
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

Step-5: The Impulse Response

𝑵−𝟏
𝟐
𝑵−𝟏 𝑵−𝟏
̅ (𝒆𝒋𝝎 ) = 𝒉 (
𝑯 ) + 𝟐 ∑ 𝒉( − 𝒏) 𝐜𝐨𝐬 𝝎𝒏
𝟐 𝟐
𝒏=𝟏
9−1
2
9−1 9−1
̅ (𝑒 𝑗𝜔 ) = ℎ (
𝐻 )+2∑ℎ( − 𝑛) cos 𝜔𝑛
2 2
𝑛=1
4
̅ (𝑒 𝑗𝜔 ) = ℎ(4) + 2 ∑ ℎ (4 − 𝑛) cos 𝜔 𝑛
𝐻
𝑛=1

̅ (𝑒 𝑗𝜔 ) = ℎ(4) + 2ℎ(4 − 1) cos 𝜔 + 2ℎ(4 − 2) cos 2𝜔 + 2ℎ(4 − 3) cos 3𝜔

𝐻
+ 2ℎ(4 − 4) cos 4𝜔
̅ (𝑒 𝑗𝜔 ) = ℎ(4) + 2ℎ(3) cos 𝜔 + 2ℎ(2) cos 2𝜔 + 2ℎ(1) cos 3𝜔
𝐻
+ 2ℎ(0) cos 4𝜔
̅ (𝑒 𝑗𝜔 ) = 0.667 + (2 × −0.276) cos 𝜔 + (2 × −0.138) cos 2𝜔
𝐻
+ (2 × 0.069) cos 4𝜔
̅ (𝒆𝒋𝝎 ) = 𝟎. 𝟔𝟔𝟕 − 𝟎. 𝟓𝟓𝟐 𝐜𝐨𝐬 𝝎 − 𝟎. 𝟐𝟕𝟔 𝐜𝐨𝐬 𝟐𝝎 + 𝟎. 𝟏𝟑𝟖 𝐜𝐨𝐬 𝟒𝝎
∴𝑯

➢ Rectangular Method:
▪ Problem: Design an ideal low filter pass filter using rectangular window
method whose designed frequency response is
𝜋 𝜋
1, ≥ 𝜔 ≥ −
𝐻𝑑 (𝑒 𝑗𝜔 ) = { 3 3
0, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
Consider 𝑁 = 9
 Given that,
𝝅 𝝅
𝟏, ≥𝝎≥−
𝑯𝒅 (𝒆𝒋𝝎 ) = { 𝟑 𝟑
𝟎, 𝒐𝒕𝒉𝒆𝒓𝒘𝒊𝒔𝒆

92
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

Step-1: Graph

Step-2: Design Impulse Response

𝟏 𝝅/𝟑
𝒉𝒅 (𝒏) = ∫ 𝑯𝒅 (𝒆𝒋𝝎 )𝒆𝒋𝝎𝒏 𝒅𝝎
𝟐𝝅 −𝝅/𝟑

1 𝜋/3
ℎ𝑑 ( 𝑛 ) = ∫ (1)𝑒 𝑗𝜔𝑛 𝑑𝜔
2𝜋 −𝜋/3

1 𝜋/3 𝑗𝜔𝑛
ℎ 𝑑 (𝑛 ) = ∫ 𝑒 𝑑𝜔
2𝜋 −𝜋/3
𝜋
𝑗𝜔𝑛 3
1 𝑒
ℎ𝑑 (𝑛 ) = [ ]
2𝜋 𝑗𝑛 −𝜋
3
𝜋
1
ℎ 𝑑 (𝑛 ) = [𝑒 ] 3 𝜋
𝑗𝜔𝑛
𝑗2𝑛𝜋 −
3

1 𝑛𝜋 𝑛𝜋
ℎ𝑑 ( 𝑛 ) = [𝑒 𝑗 3 − 𝑒 −𝑗 3 ]
𝑗2𝑛𝜋
𝑛𝜋 𝑛𝜋
1 𝑒 𝑗 3 − 𝑒 −𝑗 3
ℎ𝑑 (𝑛 ) = 2𝑗 [ ]
𝑗2𝑛𝜋 2𝑗

1 𝑛𝜋
ℎ𝑑 (𝑛 ) = [sin ]
𝑛𝜋 3
𝒏𝝅
𝐬𝐢𝐧
∴ 𝒉𝒅 (𝒏) = 𝟑
𝒏𝝅
93
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

Now,
𝑵−𝟏
( ) |𝒏| ≤
𝒉(𝒏) = {𝒉𝒅 𝒏 , 𝟐
= 𝟒; 𝒂𝒔 𝑵 = 𝟗
𝟎, 𝒐𝒕𝒉𝒆𝒓𝒘𝒊𝒔𝒆
When,
𝒏=𝟎
𝒏𝝅
𝐬𝐢𝐧
𝒉𝒅 (𝟎) = 𝐥𝐢𝐦 𝟑
𝒏→𝟎 𝒏𝝅
𝑛𝜋
1 sin 3
ℎ𝑑 (0) = lim
𝑛→0 3 1
𝑛𝜋
3
𝑛𝜋 𝐬𝐢𝐧 𝜽
1 sin
ℎ𝑑 (0) = lim 𝑛𝜋3 𝐥𝐢𝐦
𝜽→𝟎 𝜽
=𝟏
3 𝑛→0
3
1
ℎ𝑑 (0) = (1)
3
𝟏
∴ 𝒉𝒅 (𝟎) = = 𝟎. 𝟑𝟑
𝟑
𝒏 = 𝟏,
𝝅
𝐬𝐢𝐧
∴ 𝒉𝒅 ( 𝟏 ) = 𝟑 = 𝟎. 𝟐𝟕𝟔
𝝅
𝒏 = 𝟐,
𝟐𝝅
𝐬𝐢𝐧
∴ 𝒉 𝒅 (𝟐 ) = 𝟑 = 𝟎. 𝟏𝟑𝟖
𝟐𝝅
𝒏 = 𝟑,
𝐬𝐢𝐧 𝝅
∴ 𝒉𝒅 (𝟐) = =𝟎
𝟑𝝅

94
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

𝒏 = 𝟒,
𝟒𝝅
𝐬𝐢𝐧
∴ 𝒉 𝒅 (𝟒 ) = 𝟑 = −𝟎. 𝟎𝟔𝟗
𝟒𝝅
Now,
𝒉(𝒏) = 𝒉𝒅 (𝒏) ∗ 𝝎𝒓 (𝒏)
Here,
𝝎𝒓 (𝒏) = Rectangular window factor
𝑵−𝟏
|𝒏| ≤
𝝎𝒓 (𝒏) = {𝟏, 𝟐
= 𝟒; 𝒂𝒔 𝑵 = 𝟗
𝟎, 𝒐𝒕𝒉𝒆𝒓𝒘𝒊𝒔𝒆
As the graph is symmetrical so, 𝒉(𝒏) = 𝒉(−𝒏)
𝒏 = 𝟎,
∴ 𝒉(𝟎) = 𝒉𝒅 (𝟎) ∗ 𝝎𝒓 (𝟎) = 𝟎. 𝟑𝟑 ∗ 𝟏 = 𝟎. 𝟑𝟑
𝒏 = 𝟏,
∴ 𝒉(𝟏) = 𝒉𝒅 (𝟏) ∗ 𝝎𝒓 (𝟏) = 𝟎. 𝟐𝟕𝟔 ∗ 𝟏 = 𝟎. 𝟐𝟕𝟔 = 𝒉(−𝟏)
𝒏 = 𝟐,
∴ 𝒉(𝟐) = 𝒉𝒅 (𝟐) ∗ 𝝎𝒓 (𝟐) = 𝟎. 𝟏𝟑𝟖 ∗ 𝟏 = 𝟎. 𝟏𝟑𝟖 = 𝒉(−𝟐)
𝒏 = 𝟑,
∴ 𝒉(𝟑) = 𝒉𝒅 (𝟑) ∗ 𝝎𝒓 (𝟑) = 𝟎 ∗ 𝟏 = 𝟎 = 𝒉(−𝟑)
𝒏 = 𝟒,
∴ 𝒉(𝟒) = 𝒉𝒅 (𝟒) ∗ 𝝎𝒓 (𝟒) = −𝟎. 𝟎𝟔𝟗 ∗ 𝟏 = −𝟎. 𝟎𝟔𝟗 = 𝒉(−𝟒)

Step-3: The FIR Filter Transfer Function

𝑵−𝟏
𝟐

𝑯(𝒛) = 𝒉(𝟎) + ∑ 𝒉(𝒏) [𝒛𝒏 + 𝒛−𝒏 ]

𝒏=𝟏

95
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03
9−1
2

𝐻 (𝑧) = 0.33 + ∑ ℎ(𝑛) [𝑧 𝑛 + 𝑧 −𝑛 ]

𝑛=1
4

𝐻 (𝑧) = 0.33 + ∑ ℎ(𝑛) [𝑧 𝑛 + 𝑧 −𝑛 ]

𝑛=1

𝐻(𝑧) = 0.33 + ℎ(1)[𝑧 1 + 𝑧 −1 ] + ℎ(2)[𝑧 2 + 𝑧 −2 ] + ℎ(3)[𝑧 3 + 𝑧 −3 ]

+ ℎ(4)[𝑧 4 + 𝑧 −4 ]
𝐻(𝑧) = 0.33 + ℎ(1)𝑧 + ℎ(1)𝑧 −1 + ℎ(2)𝑧 2 + ℎ(2)𝑧 −2 + 0 + ℎ(4)𝑧 4
+ ℎ(4)𝑧 −4
∴ 𝑯(𝒛) = 𝟎. 𝟑𝟑 + 𝟎. 𝟐𝟕𝟔𝒛 + 𝟎. 𝟐𝟕𝟔𝒛−𝟏 + 𝟎. 𝟏𝟑𝟖𝒛𝟐 + 𝟎. 𝟏𝟑𝟖𝒛−𝟐 − 𝟎. 𝟎𝟔𝟗𝒛𝟒
− 𝟎. 𝟎𝟔𝟗𝒛−𝟒

Step-4: Realizable Filter

𝑵−𝟏
̅ (𝒛) = 𝑯(𝒛). 𝒛−(
𝑯 𝟐
)

𝟗−𝟏
̅ (𝒛) = 𝑯(𝒛). 𝒛−(
𝑯 𝟐
)

̅ (𝒛) = 𝑯(𝒛). 𝒛−𝟒

𝑯
̅ (𝒛) = {0.33 + 0.276𝑧 + 0.276𝑧 −1 + 0.138𝑧 2 + 0.138𝑧 −2 − 0.069𝑧 4
𝑯
− 0.069𝑧 −4 }𝑧 −4
̅ (𝒛) = 0.33𝑧 −4 + 0.276𝑧 −3 + 0.276𝑧 −5 + 0.138𝑧 −2 + 0.138𝑧 −6 − 0.069
𝑯
− 0.069𝑧 −8
̅ (𝒛) = −0.069𝑧 −8 + 0.138𝑧 −6 + 0.276𝑧 −5 + 0.33𝑧 −4 + 0.276𝑧 −3
∴𝑯
+ 0.138𝑧 −2 − 0.069
For Symmetric & Odd
𝒉(𝒏) = 𝒉(𝑵 − 𝟏 − 𝒏)
𝒉(𝟎) = 𝒉(𝟗 − 𝟏 − 𝟎) = 𝒉(𝟖) = −𝟎. 𝟎𝟔𝟗
𝒉(𝟏) = 𝒉(𝟗 − 𝟏 − 𝟏) = 𝒉(𝟕) = 𝟎
𝒉(𝟐) = 𝒉(𝟗 − 𝟏 − 𝟐) = 𝒉(𝟔) = 𝟎. 𝟏𝟑𝟖

96
 Md. Anawar Ibrahim (Ananta)
Reg: 3098; EEE-03

𝒉(𝟑) = 𝒉(𝟗 − 𝟏 − 𝟑) = 𝒉(𝟓) = 𝟎. 𝟐𝟕𝟔

𝒉(𝟒) = 𝒉(𝟗 − 𝟏 − 𝟒) = 𝒉(𝟒) = 𝟎. 𝟑𝟑

Step-5: The Impulse Response

𝑵−𝟏
𝟐
𝑵−𝟏 𝑵−𝟏
̅ (𝒆𝒋𝝎 ) = 𝒉 (
𝑯 ) + 𝟐 ∑ 𝒉( − 𝒏) 𝐜𝐨𝐬 𝝎𝒏
𝟐 𝟐
𝒏=𝟏
9−1
2
9−1 9−1
̅ (𝑒 𝑗𝜔 ) = ℎ (
𝐻 )+2∑ℎ( − 𝑛) cos 𝜔𝑛
2 2
𝑛=1
4
̅ (𝑒 𝑗𝜔 ) = ℎ(4) + 2 ∑ ℎ (4 − 𝑛) cos 𝜔 𝑛
𝐻
𝑛=1

̅ (𝑒 𝑗𝜔 ) = ℎ(4) + 2ℎ(4 − 1) cos 𝜔 + 2ℎ(4 − 2) cos 2𝜔 + 2ℎ(4 − 3) cos 3𝜔

𝐻
+ 2ℎ(4 − 4) cos 4𝜔
̅ (𝑒 𝑗𝜔 ) = ℎ(4) + 2ℎ(3) cos 𝜔 + 2ℎ(2) cos 2𝜔 + 2ℎ(1) cos 3𝜔
𝐻
+ 2ℎ(0) cos 4𝜔
̅ (𝑒 𝑗𝜔 ) = 0.33 + (2 × 0.276) cos 𝜔 + (2 × 0.138) cos 2𝜔
𝐻
+ (2 × −0.069) cos 4𝜔
̅ (𝒆𝒋𝝎 ) = 𝟎. 𝟑𝟑 + 𝟎. 𝟓𝟓𝟐 𝐜𝐨𝐬 𝝎 + 𝟎. 𝟐𝟕𝟔 𝐜𝐨𝐬 𝟐𝝎 − 𝟎. 𝟏𝟑𝟖 𝐜𝐨𝐬 𝟒𝝎
∴𝑯

97