Advanced Digital Signal Processing
Lecture-1
Presented By: Dr Kiran Khurshid
1
Course Information
Course Number and Title: EC-866 Adv Digital Signal Processing
Credits: 3
Instructor(s)-in-charge: Dr Kiran Khurshid
Course
Course type: Lecture
Outline
Required or Elective: Required/Core
Course pre-requisites
Degree and Semester MSCS-1st Semester
Month and Year Fall 2024
Grading: Theory (100%) Lab
(N/A)
Mid term
30%
Exam
Project 10%
Course Quizzes 10%
Assessment
Assignments 10%
Final Exam 40%
Plagiarism Policy
Any work (Assignments, Projects) if
found copied, will have strict penalties
Introduction to Course outline, objectives, teaching plan,
1 1-D and 2-D assessment method, a brief Introduction and
signal processing review to the basics of 1-D and 2-D signals
Linear time-invariant (LTI) systems, convolution
sum, correlation, deconvolution, blind
Fundamentals of
deconvolution, finite (FIR) and infinite (IIR)
2-3 Digital Signal
impulse responses, difference equations,
Course Processing.
Outline discrete-time Fourier transform (DTFT) and its
properties, z-transform (ZT) and its properties
Sampling of continuous-time signal and
sampling rate conversion: the sampling
theorem and some of its variations,
Multi-rate signal reconstruction formulae, application to the
4-5
processing discrete-time processing of continuous-time
signals. Sampling rate conversion in multi-rate
systems, multi-rate signal processing, bandpass
Pole-zero representation for rational systems,
Transform
study of various important systems including
6-7 analysis of LTI
all-pass system, inverse system and minimum-
systems
phase system.
Definition and properties of the discrete
Course Discrete Fourier Fourier series, definition of the DFT and its
8
Outline transform (DFT) properties, application to linear convolution
and circular convolution.
Understanding the computational problem,
moving towards most commonly used Fast
Computation of Fourier Transform (FFT) algorithms (radix-2,
10
the DFT decimation-in-time, decimation-in-frequency,
etc.), possible generalizations and
specializations.
Spectrogram, Understanding the issues with DTF/FFT and
Wavelet how STFT/wavelets resolve them. Introducing
11-12
Transform and the concepts of data driven transform using
EMD EMD as a key algorithm.
Filter design as a numerical approximation
Digital Filter
problem, transformation techniques for the
13-14 design
design of IIR filters, FIR filter design by
techniques
windowing.
Course
Outline Signal Modeling, Least Square Method, Pade
Signal Modeling
15-16 Approximation, Prony’s Method, Finite Data
and Analysis
Record and Stochastic Models
Aim is to study some of the more advance
topics such as:
• Time Series Analysis
17 Advance Topics
• Lattice Filters
• Wiener Filtering
• Adaptive Filters
1. Course book and Related Course Material
Textbooks: 1. Discrete-Time Signal Processing, Alan V. Oppenheim, Ronald W.
Schafer (2009)
2. Digital Signal Processing, Principles, Algorithms, and Applications
3rd Edition. John G. Proakis and Dimitris G. Manolakis
3. Statistical Digital Signal Processing And Modeling, Monson H.
Hayes
Reference Books & 1. Two-Dimensional Signal and Image Processing, Jae S. Lim
Material: 2. Digital Signal Processing- A Practical Approach, Ifeachor Jervis
3. DSP First: A multimedia approach, James H. McClellan, Ronald W.
Schafer, Mark A. Yoder
4. Digital Signal Processing: A Computer-Based Approach, S. K.
Mitra
Introduction to Digital Signal Processing
Signal:
Function that is dependant on some variable.
Signal Types: (Synthetic, Natural)
Discrete Signal:
A discrete signal is one that is defined or measured at distinct points or intervals.
It can have a finite or countably infinite set of possible values.
Digital Signal:
A digital signal is a specific type of discrete signal that is represented using discrete
symbols or levels
Introduction to Digital Signal Processing
Why Signals?
Because signals carry information.
Why Processing?
To obtain them in a more desirable form.
Example: MUX, DeMUX, Noise Filtering (ECG, EMG Signals)
Ways to Process?
Analog, Digital, Mixed
Convenient and accurate?
Vacuum tubes..transistors, Ics…..
Advantages
• Less sensitivity to components tolerance
• Fully integrated. More accurate.
• Same processor can do Time Multiplexing (by utilizing space).
• Different parts can work at different rates. Multirate Sampling.
• Data can be stored easily.
• Easy adjustment of processor characteristics.
Disadvantages
• ???
• Increased Complexity
• A to D conversion- Speed is limited.
• Power dissipations
Evolution of Communication
Fire Wall of China Lighthouse
37
What is Comminucation in Today’s World?
38
Digital Signal
x(n)= { 1, 2, -2, 1, 5}
N= 5
n= 0: N-1
Right Sided Sequence? Left Sided Sequence?
Sequences
Even Odd General
• Even part?
• Odd part?
• Even part?
• Odd part?
Representation of discrete time signals
• Graphical Representation
• Functional Representation
• Tabular Representation
• Sequence Representation
Graphical Representation
Consider a discrete time signal x(n) with the values,
x(−3) = −2,
x(−2) = 3,
x(−1) = 0,
x(0) = −1,
x(1) = 2,
x(2) = 3,
x(3) = 1
This discrete time signal can be represented graphically as shown in the figure below.
Functional Representation
In the functional representation of discrete time signals, the magnitude of the signal is
written against the values of n.
Tabular Representation
In the tabular representation of discrete time signals, the sampling instant n and
the magnitude of the discrete time signal at the corresponding sampling instant
are represented in the form of a table. The discrete time signal x(n) can be
represented in the tabular form as given below.
n -3 -2 -1 0 1 2 3
x(n) -2 3 0 -1 2 3 1
Sequence Representation
• The discrete time signal x(n) can be represented in the
sequence representation as follows:
• x(n)={−2,3,0,−1,2,3,1 }
The arrow mark (↑) denotes the term corresponding to n =
0. When no arrow is indicated in the sequence
representation of a discrete time signal, then the first term
of the sequence corresponds to n = 0.
196380427
