EE 469 Introduction to Digital Media Engineering, Spring 2002
(http://wwwclasses.usc.edu/engr/ees/469/index.html)
Instructor
Antonio Ortega
Signal and Image Processing Institute
Integrated Media Systems Center
University of Southern California
3740 McClintock Ave., EEB 436
Los Angeles, CA 900892564
Tel: (213) 7402320
Fax: (213) 7404651
Email: ortega@sipi.usc.edu
Office Hours: 1:30pm3:30pm Tuesdays, Thursdays
Teaching Assistant/Grader
Lavanya Vasudevan
Tel: (213) 7400022
Email: vasudeva@usc.edu
Office hours: 45pm Mondays, Wednesdays
Schedule

Classes 11am  12:20pm Tuesdays, Thursdays OHE 100

Discussion 1pm  1:50pm Wednesdays OHE 100

Midterm 1 11am  12:15pm, Feb 21, 2002

Midterm 2 11am  12:15pm, Apr 4, 2002

Final 11am  1pm, May 2, 2002
Abstract
The purpose of this course is to present the basic techniques for processing,
storage and delivery of media such as audio, images and video. This requires
an understanding of digital signal processing for one dimensional signals,
as well as its extensions to 2D cases. The emphasis in this class will
be on the theoretical basis for the processing, rather than on the tools
used for processing. The class will include extensive hands on problem
solving using tools such as Matlab. This course is targeted at EE or CS
seniors, and first year graduate students.
This semester EE 469 will be structured around the understanding of
a single video compression system, loosely based on the MPEG, H.26x international
video compression standards. The goal is not to study MPEG as such, but
rather to study the key components and tools that are used in designing
such a system. We will cover both 1D and 2D tools. The key components to
be studied are: sampling, digital filtering, image transforms, quantization,
entropy coding and motion estimation and compensation.
Course organization
The homeworks will include both computer assignments and solution of textbook
problems. The computer assignments will be based on Matlab. The exams
will be cumulative and will be open book/notes. The grade will be based
on homeworks (25%), midterms (25% each) and a final (25%).
Prerequisites
EE 301a, Introduction to Linear Systems or EE 483 Digital Signal Processing,
EE364 or EE 464 Probability Theory or MATH 407, or equivalent courses.
While EE 483 is not required, it is preferable to take EE 469 after EE
483. Please send me email or call me if you have any questions about the
prerequisites.
Textbook and tools
(Note: Class notes will be distributed).

Required Introduction to Data Compression, Khalid Sayood,
Morgan Kaufmann Publishers.

Recommended Digital Signal Processing Using Matlab, Vinay
Ingle and John Proakis, PWS Publishing Company.
Material covered and deadlines (based on Spring '00, will be updated with
every lecture)

Lecture
#1, 01/08/02. Quiz  ps,
pdf.
Introduction: A complete digital video compression system. Compression
Examples.
Note: Remote students, please fax the completed quiz to the Professor.

Lecture
#2, 01/10/02. Examples of energy compaction and perceptual masking,
need for efficient signal representations.

Lecture
#3, 01/15/02. Formalizing signal representations  vector
spaces, bases.

Lecture
#4, 01/17/02. Definition of basis, Orthogonality.

Lecture
#5, 01/22/02. Orthogonal Projection, Matrix Representation.

Lecture
#6, 01/24/02. Parseval's Relation, Biorthogonal bases, GramSchmidt
Orthogonalization.

Lecture
#7, 01/29/02. Basis Functions and Linear Operators, DTFT.

Handout #1, 1/30/02  ps, pdf.

Lecture
#8, 01/31/02. Time and Frequency Localization. DTFT, DFS, DFT.

Lecture
#9, 02/05/02. Block transforms continued.

Lecture
#10, 02/07/02. 2D DCT.

Homework #1 Solutions, 02/08/02  pdf.Matlab
files and Results.

Homework #2, 02/12/02  ps,
pdf.

Lecture
#11, 02/12/02. 2D transforms continued.

Lecture
#12, 02/14/02. DWT as block transform and filterbank, Ztransform.

Lecture
#13, 02/19/02. Perfect reconstruction filterbanks.

Homework #2 Solutions, 02/19/02  pdf.

Midterm 1 Crib Sheet  ps, pdf

Midterm 1 11am  12:15pm, Feb 21, 2002

Lecture
#14, 02/26/02. Z transform, Multirate Identities.

Lecture
#15, 02/28/02. 2 Channel Filterbanks

Lecture
#16, 05/03/02. Treestructured filterbanks. Time frequencytradeoffs.

Lecture
#17, 07/03/02. Implementation issues. Filterbanks. 2D separable filterbanks.
Signal Extensions.

Midterm
1 Solutions, pdf

JPEG Powerpoint slides

Homework #3, 03/20/02  ps, pdf

Lecture
#18, 3/19/02. Quantizer design. Nearest neighbor, Centroid conditions.

Lecture
#19, 3/21/02. Modelling signal distributions. Empirical mean,
variance, energy.

Lecture
#20, 3/26/02. Predictive coding. Decorrelating transforms.

Lecture
#21, 3/28/02. Quantizer design and analysis. Practical
quantizers.

Lecture
#22, 04/02/02. Optimal Quantizers. Bit allocation problem.

Homework
#4 Solutions.

Midterm 2 Crib Sheet  ps, pdf

Midterm 2 11am  12:15pm, Apr 4, 2002

Lecture
#23, 4/09/02. Notion of Entropy and Huffman coding.

Lecture
#24, 4/11/02. Block coding, Run length coding.

Lecture
#25, 4/16/02. Practical entropy coding in JPEG, JPEG 2000.

Homework #5, 04/16/02  ps, pdf.

Lecture
#26, 4/18/02. DWT Based coding continued. Overview of MPEG audio.

Midterm
2 Solutions

Lecture
#27, 4/23/02. MPEG Audio continued. Video Coding introduction.

Lecture
#28, 4/25/02. MPEG Video Coding  Motion estimation and compensation.

Homework
#5 Solutions

Final Crib Sheet  ps, pdf

Final 11am  1pm, May 2, 2002
©19982002 Antonio Ortega. Last modified:
Mon April 29 2002