University of California, Berkeley
Electrical Engineering and Computer Sciences Department
Course Details | Resources | Course Handouts | Lecture Notes | Homework | Discussion | Announcements
 

EE123, Spring 2007
Digital Signal Processing

Tues. and Thurs.: 9:30 am - 11:00 am
61 Evans

Discussion Sections:
Tuesdays, 2:00 - 3:00 pm, 400 Cory
Fridays, 10:00 am - 11:00 am, 299 Cory

Prerequisite:  EE120, graduate standing, or consent of the instructor.

Texts:
″Discrete Time Signal Processing,″ by A.V. Oppenheim and R.W. Schafer with John Buck, Prentice Hall, 1999 (required).

″Wavelets and Filter Banks,″ by G. Strang and T. Nguyen, Wellesley Cambridge Press, (optional)

Prior semester archives:
Archives
Fall 2003 Webcast
Fall 2005 Webcast
Fall 2006 Webcast

Lecturer:
Professor Anant Sahai
267 Cory Hall
sahai@eecs.berkeley.edu

Office Hours:
Tue/Thu, 11:00-12:00 noon in 258 Cory Hall

Teaching Assistant:
Rahul Tandra
264 Cory Hall
Phone:  643-9241
tandra@eecs.berkeley.edu

Office Hours:
Tuesdays, 1:00 PM - 2:00 PM, 479 Cory
Fridays, 2:00 PM - 3:00 PM, 479 Cory

Reader:
Hao Zhang
zhanghao@eecs.berkeley.edu

Office Hours:
Mondays, 6:00 PM - 7:00 PM, 293 Cory

Course Administrative Assistant:
Therese George
(510) 642-2384
therese@eecs.berkeley.edu


Announcements:

  • Welcome to EE123

  • This weeks discussion on Friday (04/13/07) has been cancelled. A make up discussion is scheduled this Tuesday (04/10/2007) from 6:30 - 7:30 PM in 258 CORY.
  • There will not be any discussion sections next week (04/16 - 04/20).
  • The office hours of the Teaching Assistant have been moved to Fridays 2:00 - 3:00 PM.
  • The office hours of the Teaching Assistant have been moved to Tuesdays 1:00 - 2:00 PM.
  • The main lecture room has been moved to 61 Evans Hall starting from Thursday, 25th January 2007.
  • Based on class-input, three make-up lectures have been scheduled in 299 Cory. We are trying to move the main lecture into a room with desks, and will update this page when that has been confirmed.
  • No discussion section the first week of classes
  • Students who do not pick up graded homework in lecture, can pick it up from Therese during regular office hours (8:00-12:00 noon and 1:00-4:30 pm) in 253 Cory Hall.
  • All EE123 students can have "named" accounts on our Instructional computers, which include UNIX, Windows and MacOSX. Matlab runs on them all.. Students can use the computer labs in 199, 105 and 119 Cory. Most students already have computer accounts that work in those labs.

    How to get a "named" account: http://inst.eecs.berkeley.edu/connecting.html#accounts.
    (go to 199 Cory, login as "newacct" with password "newacct")

    We have MATLAB on all instructional Windows and UNIX systems, including the remote-access servers:   http://inst.eecs.berkeley.edu/~inst/iesglabs.html.

    We have the Signal Processing and other toolkits, listed on:   http://inst.eecs.berkeley.edu/cgi-bin/pub.cgi?file=matlab.help.

  • Problem sets handed in late will not be accepted unless consent is obtained from the teaching staff prior to the due date.


Back to top

Course Details:

  • Academic Dishonesty Policy
  • Grading Percentage

    1. 40%: Final Exam (individual effort)
    2. 15% each: 3 Midterms (combination of group/individual effort)
    3. 10% total: Homeworks (group effort)
    4. 5% total: Class Participation
    5. 0-15% total: Extra Credit Projects (optional)
    Undergraduates and graduate students will be treated as distinct populations for grading purposes.
  • Outline of Topics

    1. Overview of DSP and Applications
    2. LTI Systems, Sampling, Z-transform and DTFT
    3. DFT and FFT
    4. Oversampling, Interpolation, A/D, D/A
    5. Filter Design
    6. Finite Register Effects
    7. Multirate Signal Processing and Wavelets
  • Tentative Course Schedule

    This will change in response to class needs and student interest. Midterms are cumulative.

    Lecture Date Reference Topic
    1 Jan. 16 Motivation for DSP: The example of GPS
    2 Jan. 18 O&S 2.0-2.6, 4.0-4.3 Review of LTI systems: convolution, sampling and aliasing
    3 Jan. 23 O&S 2.0-2.6, 4.0-4.3 Review of LTI systems continued
    4 Jan. 25 O&S 2.6-2.9, 3.0-3.4 Review of DTFT, Z-transform, ROC
    5 Jan. 25 7-8:30pm in 299 Cory O&S 2.6-2.9, 3.0-3.4 Review of DTFT, Z-transform, ROC continued
    -- Jan 30 No class --- make up class on Jan 25th evening
    -- Feb. 1 No class --- make up class on Feb 8th evening
    6 Feb. 6 O&S 5.0-5.6 Rational transfer functions
    7 Feb. 8 O&S 5.0-5.6 Rational transfer functions continued
    8 Feb. 8 7-8:30pm in 299 Cory O&S 8.0-8.7, 9.0-9.4 The DFT and FFT
    9 Feb. 13   Review Session
    10 Feb. 15   In Class Midterm 1
    11 Feb. 20 O&S 8.0-8.7, 9.0-9.4 The DFT and FFT continued
    12 Feb. 22 O&S 8.0-8.7, 9.0-9.4 The DFT and FFT continued
    13 Feb. 27 O&S 8.0-8.7, 9.0-9.4 The DFT and FFT continued
    14 March 1 O&S 4.4-4.5 Discrete time processing of continuous time signals, D/A and A/D
    15 March 6 O&S 4.6, 4.9 Oversampling, Upsampling, Downsampling, Interpolation
    16 March 8 O&S 4.6, 4.9 Oversampling, Upsampling, Downsampling, Interpolation
    17 March 13   Review Session
    18 March 15 In Class Midterm 2
    19 March 20 O&S 5.7, 7.2-7.4 FIR filter design
    20 March 22 O&S 5.7, 7.2-7.4 FIR filter design
    -- March 27 No class -- Spring break
    -- March 29 No class -- Spring break
    21 April 3 O&S 7.1, Appendix B IIR filter design
    22 April 5 O&S 7.1, Appendix B IIR filter design
    23 April 10 O&S 4.8, 6.6-6.8 Quantization and its effects
    24 April 12 O&S 4.8, 6.6-6.8 Quantization and its effects
    25 April 12 7-8:30pm in 299 Cory O&S Appendix A Random signals and noise
    26 April 17 Review Session
    27 April 19 In Class Midterm 3
    -- April 24 No class Make-up class on Apr 12th
    28 April 26 Sampling below the Nyquist rate
    29 May 1 Filter banks, multirate processing, and wavelets
    30 May 3 Filter banks, multirate processing, and wavelets
    31 May 8 Applications -- CD, JPEG2000, MP3
    32 May 15 8-11AM in TBA Final Exam

Resources:

 
Fast Convolution
Covers various implementations of linear convolution using the DFT, including Overlap-Add and Overlap-Save.

Upsampling vs. Oversampling for Digital Audio
An article about the benefits of these techniques.

The Scientist and Engineer's Guide to Digital Signal Processing
A great practical introduction to DSP. (Free to download)

Information on Gibbs Phenomenon
Wikipedia article on it.
Articles on Sampling below the Nyquist Rate
Sampling Signals of Finite Rate of Innovation
by Martin Vetterli

M. Vetterli, P. Marziliano, T. Blu, "Sampling Signals with Finite Rate of Innovation,"
IEEE Transactions on Signal Processing, vol. 50, no. 6, pp. 1417-1428, June 2002.

Sampling and Reconstruction of Signals with Finite Rate of Innovation in the Presence of Noise
by Irena Maravic and Martin Vetterli
Applets
Signals, Systems, and Control Demonstrations
A collection of helpful applets from Johns Hopkins University.

Demonstration of Aliasing of a Sinusoidal Signal
Applet that illustrates aliasing visually.

Magnitude/Phase DFT Applet
The "Second Applet" is a helpful tool for getting a better feel for the DFT.

Real/Imaginary DFT Applet
This is great for demonstrating the symmetry properties of the DFT.

Lecture Notes:

Students might find it a useful complement to explore the MIT equivalent of our course: 6.341.

For this course, students must sign up for scribing lecture notes for your fellow students as class participation. The scribed lecture notes will be posted here. You can use the following .tex file and .sty file for scribing lectures.

Homework:

Homework can be done in groups of 1-3 students who can submit a common assignment with all group member's names and SIDs on each page.

Self-Grading Instructions

To encourage students to read the solutions ahead of the night before the midterms, you will have to self-grade.
  1. Hand in a hardcopy of the assignment (one per group) at the beginning of class. Remember to keep a copy to grade.
  2. After solutions are posted, give each numbered problem a score of 0, 0.5, or 1 (for incorrect, at least half right, or completely correct, respectively).
  3. Email your score for each problem along with your total score to the reader for the course, at: TBA.
  4. Scores are due by midnight on Sunday.

Discussion:

  • Discussion 1: The following problems from the textbook were covered in the first discussion section: 2.14, 2.68, 2.69, 2.70
  • Discussion 2: Review of sampling, effect of time shifts, non-uniform sampling, non-ideal reconstruction filters
  • Discussion 3: Z transform review, problems from text: 3.33, 3.44, plotting magnitude and phase response from pole-zero plots
  • Discussion 4 (Tu 02/13): Review problems before Midterm 1, questions from students
  • Discussion 4 (F 02/16): Solutions to Midterm 1 problems
  • Discussion 5: Introduction to DFTs. Computing the DFT of a complex exponential with mismatch between its period and the number of point in the DFT. We also worked out problem 8.35 from the text.

Handouts:


Back to top

 Last updated 01/15/07