EEE4001F: Digital Signal Processing

Course presentation

Wednesday periods 3 and 4 (M202), Thursday period 4 (LS1D), Friday period 4 (M320). Some useful course information is available here, and this is the official course handout. A rough schedule for the lectures is here,

Teaching assistance

Teaching assistant: Mayuresh Kulkarni (mayuresh.kulkarni@uct.ac.za).
Tutors: Tracy Semmelink (tracy.semmelink@uct.ac.za), Jatin Harribhai (jatin.harribhai@uct.ac.za).

Lecture notes

Introduction (2up).
Discrete-time signals and systems (2up).
The z-transform (2up).
Sampling of continuous-time signals (2up).
The Discrete Fourier Transform (2up).
Transform analysis of LTI systems (2up).
Filter design (2up).

These notes are licensed under a Creative Commons Attribution-ShareAlike 3.0 Unported License. Latex source for producing them can be downloaded here.

Example Matlab code

Zero padding for linear convolution using the FFT.
Use of windows for Spectrum estimation.
Any other code in the examples directory.

Problem Sets and Matlab Assignments

To be handled through Vula.

DSP Project

You are expected to undertake a substantial project (approximately 10-20 hours) for this course. You may work in groups of two.

Details for this year's project can be found here (2up). Resources are available here.

You may alternatively select a topic from previous years:

Project 1: DTMF coding and decoding (2up).
Project 2: Signal restoration (2up). Sound files available here.
Project 3: Sampling rate conversion (2up).
Project 4: Transform coding of images (2up).
(Resources: project04_jpeg.m, gpeppers.pgm, dct.pdf, compres.pdf.)
Project 5: Echo cancellation (2up).
Project 6: Multirate filtering (2up).
Project 7: Pitch shifting of audio signals (2up).

I don't mind if you prefer to propose a topic of your choice — just discuss it with me so we can be sure it is suitable.

Imaging applications section

Note that not all of these sections are covered every year. If it's not in the lectures you don't need to know it.

Basic notes:

Multidimensional DSP (slides, notes, 2up).
Human visual system (slides, notes, 2up).
Edge detection (slides, notes, 2up).
Image formation (slides, notes, 2up).

Optimisation notes:

Linear regression (slides, notes, 2up).
Robust regression (slides, notes, 2up).
Nonlinear regression (slides, notes, 2up).

Radar section

A link to Prof. Wilkinson's radar/sonar section can be found here.

Class Tests

There will be two class tests during the course. Dates to be determined. Past tests and exams can be found here.

Exam information

The exam consists of two parts:

Core DSP: There will be (usually seven) mandatory questions on the core DSP concepts presented in the first 24 lectures of the course (up to the end of filter design). These will count 70% of the final exam mark.
Applied DSP: There will be an additional two questions, one for each of the application sections presented. These make up the remaining 30% of the exam mark.

Nominally assignments and projects will count 20% towards the final mark, class tests will count 20%, and the remaining 60% will come from the exam. However, these proportions may be adjusted to give you the maximum benefit.

Resources

Some web links.
Past tests and exams for the course.

Last modified: Wed Apr 18 09:15:24 SAST 2012