College of Engineering and Technology

Electronics and Communication Department

EEC3220 Introduction to Communication Systems

LECTURE (10) – DISCRETE FOURIER TRANSFORM –

CONVOLUTION OF DISCRETE SIGNAL

Dr. Dina A. Ragab

1/5/2025 1
Lecture (10) – Discrete Fourier Transform - Convolution
Contacts
Lecturer: Dr. Dina A. Ragab
 Email: [email protected]
Section and Lab: Eng. Hesham Tarek
 Email: [email protected]

Dr. Dina A. Ragab

1/5/2025 2
Lecture (10) – Discrete Fourier Transform - Convolution
Text Book

Charles L. Philips, John Parr, and Eve Riskin “Signals,

Systems and Transforms,” 4th Edition, Prentice Hall,
2008.

Dr. Dina A. Ragab

1/5/2025 3
Lecture (10) – Discrete Fourier Transform - Convolution
Grading System
 7th Week = 30%
 5% : Section (5th)
 5%: Lab (5th)
 20%: Exam (7th)

 12th Week = 20%

 5% : Section (10th)
 15%: Exam (12th)

 10% = Lab Exam (14th)

 40% = Final Exam

Dr. Dina A. Ragab

1/5/2025 4
Lecture (10) – Discrete Fourier Transform - Convolution
ABET Outcomes

Dr. Dina A. Ragab

1/5/2025 5
Lecture (10) – Discrete Fourier Transform - Convolution
Course Outline
Week Description
1 Introduction
2 Types of Signals – Types of Systems
3 Fourier Series
4 Fourier Transform
5 Fourier Transform Properties (1)
6 Fourier Transform Properties (2) – Convolution
7 7th week Exam
8 Fourier Transform of Special Functions
9 Fourier Transform for Periodic Signals
10 Correlation - Sampling
11 Discrete Fourier Transform – Convolution of DFT
12 12th week Exam
13 Hilbert Transform
14 Ideal Filters
15 Revision

Dr. Dina A. Ragab

1/5/2025 6
Lecture (10) – Discrete Fourier Transform - Convolution
 Lecture (12)
Discrete Fourier Transform

Dr. Dina A. Ragab

1/5/2025 7
Lecture (10) – Discrete Fourier Transform - Convolution
What is DFT?
 The Discrete Fourier Transform (DFT) allows the computation of spectra from
discrete-time data.

 The DFT is a discrete version of the Fourier Transform.

 DFT is extremely important in the area of frequency (spectrum) analysis.

 Without a discrete-time to discrete-frequency transform we would not be

able to compute the Fourier transform with a microprocessor or DSP based
system.

Dr. Dina A. Ragab

1/5/2025 8
Lecture (10) – Discrete Fourier Transform - Convolution
• Continuous-Time Signals
 Most signals in the real world are g(t)
continuous time, as the scale is
infinitesimally fine.
 Denote by g(t), where the time
t
interval may be bounded (finite) or
infinite

• Discrete-Time Signals
 Some real world and many digital signals
are discrete time, as they are sampled g[n]
 Contains information about the signal
only at discrete points in time
 Denote by g[n], where n is an integer
value that varies discretely n

Dr. Dina A. Ragab

1/5/2025 9
Lecture (10) – Discrete Fourier Transform - Convolution
 Discrete Fourier Transform
• Let g[n] be an N-point signal.
• The N-point DFT of g[n], denoted G(k) = DFT{g[n]}, is defined as:
N 1 2 k n
j
G ( k )   g [ n] e N
, 0  k  N 1
n 0

Note that:
 G(k) is an N-point sequence, computed from g(n), another N-
point sequence
 k and n are dimensionless variables
 k is the frequency index, n is the time index
Dr. Dina A. Ragab
1/5/2025 10
Lecture (10) – Discrete Fourier Transform - Convolution
Inverse Discrete Fourier Transform (IDFT)
• Let G(k) be an N-point DFT sequence.
• The N-point IDFT of G(k), denoted g[n] = IDFT{X(k)}, is defined as:

N 1 2 k n
1
 G (k ) e
j
g [ n]  N
, 0  n  N 1
N k 0

Dr. Dina A. Ragab

1/5/2025 11
Lecture (10) – Discrete Fourier Transform - Convolution
Example (1):
Calculate DFT of 4 element sequence, g(n)={1,1,0,0}
Solution:
𝑁−1
For k=0: −𝑗
2𝜋𝑘𝑛
𝐺 𝑘 = ෍𝑔 𝑛 𝑒 𝑁
𝑛=0
2𝜋 0 × 0 2𝜋 0 ×1 2𝜋 0 × 2 2𝜋 0 ×3
−𝑗 −𝑗 −𝑗 −𝑗
𝐺 0 =𝑔 0𝑒 4 +𝑔 1 𝑒 4 +𝑔 2 𝑒 4 +𝑔 3 𝑒 4

2𝜋 0 × 0 2𝜋 0 ×1 2𝜋 0 × 2 2𝜋 0 ×3
−𝑗 −𝑗 −𝑗 −𝑗
𝐺 0 =𝟏×𝑒 4 +𝟏 ×𝑒 4 +𝟎 ×𝑒 4 +𝟎 ×𝑒 4

𝐺 0 =2

Dr. Dina A. Ragab

1/5/2025 12
Lecture (10) – Discrete Fourier Transform - Convolution
For k=1: 𝑁−1
2𝜋𝑘𝑛
−𝑗 𝑁
𝐺 𝑘 = ෍𝑔 𝑛 𝑒
𝑛=0

2𝜋 1 × 0 2𝜋 1 ×1 2𝜋 1 × 2 2𝜋 1 ×3
−𝑗 −𝑗 −𝑗 −𝑗
𝐺 1 =𝑔 0𝑒 4 +𝑔 1 𝑒 4 +𝑔 2 𝑒 4 +𝑔 3 𝑒 4

2𝜋 1 × 0 2𝜋 1×1 2𝜋 1 × 2 2𝜋 1 ×3
−𝑗 −𝑗 −𝑗 −𝑗
𝐺 1 =𝟏×𝑒 4 +𝟏 ×𝑒 4 +𝟎 ×𝑒 4 +𝟎 ×𝑒 4

𝜋 𝜋
𝐺 1 = 1 + cos − 𝑗𝑠𝑖𝑛 =1 −𝑗
2 2

𝐺 1 =1 −𝑗

Dr. Dina A. Ragab

1/5/2025 13
Lecture (10) – Discrete Fourier Transform - Convolution
For k=2: 𝑁−1
2𝜋𝑘𝑛
−𝑗 𝑁
𝐺 𝑘 = ෍𝑔 𝑛 𝑒
𝑛=0

2𝜋 2 × 0 2𝜋 2 ×1 2𝜋 2 × 2 2𝜋 2 ×3
−𝑗 −𝑗 −𝑗 −𝑗
𝐺 2 =𝑔 0𝑒 4 +𝑔 1 𝑒 4 +𝑔 2 𝑒 4 +𝑔 3 𝑒 4

2𝜋 2 × 0 2𝜋 2×1 2𝜋 2 × 2 2𝜋 2 ×3
−𝑗 −𝑗 −𝑗 −𝑗
𝐺 2 =𝟏×𝑒 4 +𝟏 ×𝑒 4 +𝟎 ×𝑒 4 +𝟎 ×𝑒 4

𝐺 2 = 1 + cos 𝜋 − 𝑗𝑠𝑖𝑛 𝜋 =1 −1=0

𝐺 2 =0

Dr. Dina A. Ragab

1/5/2025 14
Lecture (10) – Discrete Fourier Transform - Convolution
 𝑁−1
For k=3: 2𝜋𝑘𝑛
−𝑗 𝑁
𝐺 𝑘 = ෍𝑔 𝑛 𝑒
𝑛=0
2𝜋 3 × 0 2𝜋 3 ×1 2𝜋 3 × 2 2𝜋 3 ×3
−𝑗 −𝑗 −𝑗 −𝑗
𝐺 3 =𝑔 0𝑒 4 +𝑔 1 𝑒 4 +𝑔 2 𝑒 4 +𝑔 3 𝑒 4

2𝜋 3 × 0 2𝜋 3×1 2𝜋 3 × 2 2𝜋 3 ×3
−𝑗 −𝑗 −𝑗 −𝑗
𝐺 3 =𝟏×𝑒 4 +𝟏 ×𝑒 4 +𝟎 ×𝑒 4 +𝟎 ×𝑒 4

3𝜋 3𝜋
𝐺 3 =1+ cos − 𝑗𝑠𝑖𝑛 =1+𝑗
2 2

𝐺 3 =1+𝑗
Can you now apply the IDFT to
∴ DFT({1,1,0,0}) = {2,1-j,0,1+j} above sequence and get back
the original one?

Dr. Dina A. Ragab

1/5/2025 15
Lecture (10) – Discrete Fourier Transform - Convolution
 Thank you
Good Luck

Dr. Dina A. Ragab

1/5/2025 16
Lecture (10) – Discrete Fourier Transform - Convolution