EE483: Introduction to Digital Signal Processing



EE483: Introduction to Digital Signal Processing

Instructor: Homayoun Shahri

EEB102

homayoun@

1. Schedule

Lectures: 7:30am – 8:50am M/W

Discussion: 12:00pm Fridays.

Midterm: 7:30am – 8:50am Wednesday, March 12.

2. Textbooks

Required:

A course in Digital Signal Processing, Boaz Porat, J. Wiley, 1997. This is an excellent

new DSP book - more concise and up-to-date than Oppenheim and Schafer’s classic. Also includes integrated Matlab examples.

The Student Edition of MATLAB, or familiarity and experience with use of MATLAB for

signal analysis and use of graphics. MATLAB is available on the USC computer network

to all USC students through your standard Unix accounts; these systems should have

access to all toolboxes, including the Signal Processing Toolbox, which is useful but not

required for this course.

Suggested:

Discrete Time Signal Processing, A. Oppenheim and R. Schafer, Prentice-Hall, New Jersey, 1989. -- this is the classic textbook for DSP, and a model for most of the other introductory books on the subject.

Introduction to Digital Signal Processing, R. Kuc, McGraw-Hill, 1988. – This book is easy to read and has examples in C.

3. Course Objectives

The objective of this course is to provide a basic introduction to the theory of digital signal processing (DSP). We assume a familiarity with the Fourier and Laplace transforms and concepts such as linearity and shift invariance that are used in the description and analysis of linear analog systems. Much of what we do extends these ideas to the field of discrete time systems. Major parts of the course will concentrate on signal analysis using Fourier transforms, linear system analysis, Filter design and a few more advanced topics. In the first part of the course we will study the discrete Fourier transform and its properties. We will also study the sampling theorem and the relationship between continuous and discrete time transforms. We will see how discrete time, linear shift invariant systems can be characterized using linear difference equations and the impulse response and show how tools such as the z-transform and discrete Fourier transform can be used in the design and analysis of such systems. We will then study the design and implementation of digital filters. We will also try to include some topical material: how do 1-bit A/D converters work?, why is the DCT used in JPEG image compression?, what are adaptive filters? While this course deals largely with the theory of DSP, we will use a powerful software package, MATLAB, to look at applications of this theory, particularly Fourier analysis and digital filter design.

4. Course Evaluation

Homework: approximately weekly (30% of final grade)

Midterm: 7:30am – 8:50am Wednesday, March 12th. (30% of final grade)

Final Examination: (40% of final grade).

All exams are cumulative and closed book - Homeworks must all be turned in by 5:00 pm on the due date - late homeworks will not be graded. Several of the homeworks will involve the use of Matlab - as we discussed above, you can use the Student edition or the standard. Homework assignments will be posted on the website.

USC on-campus students: - please turn in the homeworks to the TA. So that we can generate an electronic mailing list, please send an e-mail message to the TA with the subject “483 on-campus student” including your name and a brief academic background.

USC-DEN and NTU students: please make sure that the homeworks are turned in on the due date. The fax number for homework submissions is (213) 740-9121; the e-mail

address is denhw@usc.edu. If on campus, you can also turn them into directly to the TA.

For information on how we handle homework from remote students, please see http://

den.usc.edu/cont/homework.html. So that we can generate an electronic mailing list,

please send an e-mail message to the TA with the subject “483 TV student” or EE483

NTU student” including your name and a brief academic background. Please note: due

dates are firm since solutions are also posted on our website. It is very important that you

turn your homework in on time.

5. Office Hours, Discussion

My office hours are 9:00-11:00pm on

Mondays, and Wednesdays. You can also email me with questions: E mail: homayoun@.

Your TAs for the course is Dagen Wang. His office is at PHE336, and his office hours are 2:00-5:00pm on Fridays.

TBA. E-mail address: dagenwan@sipi.usc.edu.

6. Suggestions and Policies

My primary interest here is that you learn as much as possible about digital signal processing, that you find the material interesting and that you finish the course and want to know more about this subject. There are a few things that you can do to facilitate this: (i) ask me questions (ii) make use of my office hours (iii) learn to use MATLAB, (vii) remember that exams and grades are a means to an end and not the end in itself.

You can discuss your Matlab assignments with others, but you must write them yourself.

7. Course Outline

Part 1: Signal Analysis

-Introduction and overview

-Continuous time signals, Discrete time signals and Fourier analysis

-Sampling of analog signals: aliasing, sampling theorem and signal reconstruction

-Practical signal reconstruction. Band-pass sampling. Frequency domain sampling

-The discrete Fourier transform: definition and properties

-The discrete cosine transform (DCT) and other unitary transforms

-Fast Fourier transforms

-Relationships between the FT, DFT and DTFT, leakage, resolution

-Practical spectral analysis

Part 2: System Analysis

-LSI systems, causality and BIBO stability, impulse response

-Linear difference equations (LDEs) for LSI systems

-Z transform and its properties, system function

-Frequency response: magnitude and phase

-Linear phase, minimum phase and all-pass systems

-Filter types: FIR, IIR, low pass, high pass, bandpass,

-Differentiators, Hilbert transformers

Part 3: Filter Design and Implementation

-Linear phase FIR filters

-FIR filter design: least squares and window design methods

-FIR filter design: Chebyschev approximation

-Classical IIR designs

-Impulse invariance, bilinear transform and frequency transformation methods.

-IIR design using least squares

-Filter structures

-Quantization effects

Part 4: Other Topics and Applications (time permitting)

-Oversampling and sigma-delta converters.

-Sample-rate conversion

-Introduction to adaptive filtering

................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download