Academic Year :

Spring 2013

Course Instructor :

Prof. Animesh Kumar

Motivation:

Its one of the coming-of-age courses for an electrical engineer. Whether you are among those who cry about having to do MA courses rather than some real-engineering-stuff or a proponent of ‘everything is maths’ philosophy, this course will surely appeal you. Understanding why a perfect low pass filter isn’t practically possible, how should signals be modified to enable their separation even after mixing, how to sample a signal to enable good reconstruction, how to compress data; applications of this course span the working of almost every modern device. With a beautiful blend of practical examples and abstract theory, the course will give an insight into topics often dealt in signal processing (filters, modulation,quantization etc). This course serves as the stepping stone for any further course in Signal Processing.

Prerequisites:

Professor introduces every concept from scratch. However having a good hang of fourier transforms, laplace transforms helps.

Course-Content:

1) Introduction:

a) Definition of Signal and it’s properties.

b) Definition and Identifying Linear Time Independent Systems(LTI).

c) BIBO Stability.

d) Impulse Response

e) Eigen-Functions of LTI system

2) Fourier Transform

3) Modulation(examples)

4)Laplace Transform

5)Filters, Direct and Parallel form of implementation of systems(examples)

6)Sampling and Reconstruction (Large topic)

7) Quantization; Lossy Compression

8) Discrete-time fourier transform(subject to time-availability and students’ interest).

Lectures:

Attending the class regularly is very useful in this course. Professor prefers to use board a lot and hence making notes is a good idea. No slides are used.Since the classes are of 1.5 hrs, you’ll be given a break of 5 minutes after 45 minutes before resuming. About 7% marks are for class participation and an interesting rule is followed: Surprise name-calling attendances are taken and you lose points if you’re absent. You gain points if you are active in class and ask questions that professor thinks is good. Also you lose points for disturbing the class (talking, coming late after the break etc). Every time you’re caught you lose marks in fibonacci series. 1st time 1, next 1, then 2..so you disturb the class 12 times and you fail!

Assignments:

About 13 % marks are for some 7-8 assignments, given at regular intervals, each consisting of 3 questions of which 2 have to be solved and submitted a week after being given. Penalty for copying is -5 marks. Solutions have to be perfect; grading is digital (0,1 or 2 marks.). Questions are usually easy and can be solved within 1-2 hours if well-versed with the concepts.

Exams:

Two Mid-sems are conducted of about 20 marks each. Papers are of 1 hour and usually require whole of the time. Questions are easy applications of the concepts and vary from true/false type to design problems. 1 page of formula sheet is allowed.The Endsem is for ~40 marks and is not very different from the mid-sems. No quizzes are taken for the course.

Difficulty:

Easy

Grading:

Please Check ASC and decide for yourself. Don’t know of any procedure followed while grading.

Reference:

“Signals and Systems” by Alan V. Oppenheim and others & MIT OCW for Signals and Systems by the author himself.

Reviewed by Chirag C.Shetty