Discrete Time Signal Processing (3 cr.)
Senate on January 29, 2003 approved a resolution on academic integrity, which requires that the following reminder to students be printed on every course outline:
McGILL UNIVERSITY VALUES ACADEMIC INTEGRITY. THEREFORE ALL STUDENTS MUST UNDERSTAND THE MEANING AND CONSEQUENCES OF CHEATING, PLAGIARISM AND OTHER ACADEMIC OFFENCES UNDER THE CODE OF STUDENT CONDUCT AND DISCIPLINARY PROCEDURES (see http://www.mcgill.ca/integrity for more information).
General Information:
Instructor:
· Prof. Benoît Champagne
· Office: McConnell Engineering Building, Room 756
· Tel: (514) 398-5701
· Email:
· Office Hours:
o Monday & Wednesday: 1:30pm to 4:30pm
o Otherwise by appointment
Lectures (CRN 2324):
· Monday, Wednesday and Friday: 11h30am to 12h30pm
· Location: Trottier Building, Room ENGTR 2100
· Lectures will start on Wednesday, September 3, 2008
· NOTE: According to the official McGill schedule, an extra Monday class will be given on Tuesday, Dec. 2, to replace lectures that would have been held on Monday, Oct. 13 (Thanksgiving Day).
Tutorials (CRN 2325):
· Friday: 3h30pm to 4h30pm
· Room ENGTR 2100
· Tutorials will start on Friday, September 12, 2008
Teaching Assistants:
Name / Office / Email / ResponsibilityNeda Ehtiati / MC751 / / Tutorials
Hafsa Qureshi / MC736 / / Problem Sets
Fang Shang / MC751 / / Grading
TA’s Office hours:
· Ms. Hafsa Qureshi will be available to answer student questions, especially as they relate to the problem sets: Tuesday, 9h00am to 10h00am, Room MC736 (otherwise by appointment).
Course Description:
Prerequisite:
· ECSE 304: Signals and Systems II, or
· ECSE 306: Fundamentals of Signals and Systems
Objectives:
· Digital signal processing (DSP) is prevalent in a wide variety of applications in electrical and computer engineering, including: consumer electronics, web-based multi-media processing, advanced wireline and wireless digital communications, sonar and radar processing, remote sensing and biomedical engineering, to name a few.
· The main objective of the course is to equip the students with the fundamental principles, methods and algorithms that are commonly used by practicing engineers and scientists in the analysis, development and design of modern DSP systems.
· The emphasis will bear on one-dimensional signals, although certain generalizations may be discussed.
List of Topics:
Introduction(1 hour)
· Definition of digital signal processing (DSP), overview of a DSP system, the discrete-time signal processing (DTSP) paradigm.
Part I: Basic concepts (~10 hour)
· Discrete-time characterization of signals and systems: linear time-invariant (LTI) systems, convolution, finite (FIR) and infinite (IIR) impulse responses, difference equations.
· Discrete-time Fourier transform (DTFT): definition, types of convergence, properties, frequency analysis of LTI systems, ideal frequency selective filters, phase and group delays.
· z-transform (ZT): definition, properties, rational forms, unilateral ZT and difference equations, z-domain analysis of LTI systems, pole-zero characterization , inverse system, all-pass and minimum-phase decomposition, linear phase property.
· Discrete Fourier transform (DFT): definition, properties, relationship between DFT and DTFT, circular versus linear convolution.
Part II: Design and implementation issues for DSP systems (~16 hours)
· Digital processing of analog signals: uniform sampling, sampling theorem, reconstruction formula, basic A/D and D/A functions, digital filtering of analog signals.
· Filter structures: signal flow graph representation, basic structures for FIR and IIR systems (direct forms, parallel, cascade, etc.), transposition theorem, lattice filter structures.
· Filter design techniques: numerical approximation problem, transformation techniques for the design of IIR filters, FIR filter design by windowing and frequency sampling, introduction to computer optimization techniques.
· Finite-precision effects: review of binary arithmetic, effects of coefficient quantization on frequency response, round-off noise in digital filtering, SQNR computation, other effects.
· Practical computation of the DFT: fast Fourier transform (FFT) algorithms, radix-2 algorithms based on decimation-in-time and in-frequency, generalization to mixed radix, FFT-based linear convolution (overlap-add and overlap-save methods).
· Introduction to programmable digital signal processors (PDSP): motivation, basic characterization and structural features, benchmarking of PDSP.
Part III: Advanced topics and applications (~10 hours)
· Frequency analysis of signals: DFT-based analysis, trade-off between temporal and frequency resolution, effects of windowing and spectral sampling, time-dependent Fourier analysis (spectrograms).
· Multirate systems: sampling rate conversion, fundamental identities, polyphase decomposition, introduction to subband filtering, oversampling in A/D and D/A conversion.
· Adaptive filtering (if time permits): motivation, reference signal, cost function, recursive optimization, the LMS algorithm, selected applications.
Course Material:
Web Support:
· The course web page on McGill University’s myCourses (WebCT Vista) will be used extensively to post course related information (e.g. course outline, problem sets, exam solutions, etc.). Make sure you consult it regularly.
Class Notes:
· A set of class notes developed over the last few years by Professors Champagne and Labeau will be used as the main text.
· These notes, which are organized in Chapters that closely follow the topics listed under Course Description, will be made available (in PDF format) on the course Web page.
· A list of the specific sections of the class note covered in class during each lecture will be maintained on the Web page.
Matlab Primer:
· A quick introduction to Matlab will be given during the first tutorial. The corresponding slides will be made available on the course Web page.
· For those who want to know more:
o There are several free Matlab tutorials available on the Web; simply search for Matlab tutorials or Matlab primer with your favorite engine.
o A few examples of such tutorials are given below under Some Interesting Links.
o Several textbooks are also available for learning Matlab, such as the one by R. Pratap under Supplementary Texts below.
Old Exams:
· A limited subset of midterms and final exams previously given in 412 will be made available via the course Web page.
· These should be consulted by the students prior to examinations.
Supplementary Texts: For those who like to learn from a more complete textbook, I would recommend the one by Proakis and Manolakis below. Students who are looking for additional problems and exercises should get hold of the Schaum’s Outline by Hayes.
· M. H. Hayes, Digital Signal Processing, Schaum’s Outlines, 1999.
· A. V. Oppenheim and R. W. Schafer, Digital Signal Processing, Prentice-Hall, 1975.
· A. V. Oppenheim, R. W. Schafer with J. R. Buck, Discrete-Time Signal Processing, 2nd Ed., Prentice Hall, 1999.
· R. Pratap, Getting Started with MATLAB: A Quick Introduction for Scientists and Engineers, Oxford University Press, 2002.
· J. G. Proakis and D. G. Manolakis, Digital Signal Processing: Principles, Algorithms, and Applications, 4th Ed., Prentice Hall, 2007.
Some interesting links:
· Nice stuff on signals and systems: http://www.jhu.edu/~signals/index.html
· Filter design applets: http://www.dsptutor.freeuk.com/
· FFT benchmark: http://www.fftw.org/
· Matlab tutorials:
- http://math.ucsd.edu/~driver/21d-s99/matlab-primer.html
- http://users.ece.gatech.edu/~bonnie/book/TUTORIAL/tutorial.html
· IEEE Signal Processing Society:
- http://www.signalprocessingsociety.org/
Homework and Evaluation:
Problem sets:
· There will be six (6) problem sets (i.e. about one every two weeks), to be posted on the course Web page along with due dates and other relevant info.
· These may include more advanced problems that require the use of numerical computing (i.e. Matalb) and/or the study of selected research papers.
· Students are allowed to do the problem sets either alone or in team of two. A standard cover page format will be provided on the Web.
· An assignment box has been reserved for the course.
· For each problem set, only selected problems will be marked.
· Solutions will be posted on the Web once the due date is past.
· All problem sets will be counted towards the final mark.
· Marked assignments will be available for pick up on the 7th floor of McConnell Engineering (near elevators).
Midterm examinations:
· There will be 2 midterm examinations, 50 minutes each in duration, and held during normal class time:
- Midterm #1: , Wednesday, October 8, 2008
- Midterm #2: , Wednesday, November 12, 2008
· These will be of the CLOSED BOOK type: only the faculty standard calculator and a dictionary will be allowed, NO crib sheet allowed.
· Both Midterms will be counted towards the final mark.
· List of material covered by each midterm will be posted on the Web page.
Final examination:
· There will be a final examination, 3 hours in duration (date and time to be announced by the Faculty)
· This will be of the CLOSED BOOK type: only the faculty standard calculator and a dictionary will be allowed; NO crib sheet allowed.
· The final examination will cover all the material included in the class notes and/or seen in class during the term.
Marking scheme:
Problem sets / 15%Midterm #1 / 17.5%
Midterm #2 / 17.5%
Final examination / 50%
Marking Policy (Assignments and Midterms):
· No assignment will be accepted after the assignments have been collected from the assignment box.
· Any requests for re-evaluation of an assignment must be made within one week of its return by contacting the instructor.
· Marked assignments not picked up within two weeks may be discarded.
· There will not be any make-up examinations for students who miss a Midterm.
· Students who miss such a midterm due to illness should notify the instructor within a week of the examination and provide him with an adequate medical certificate stating the date and nature of the illness.
· Under presentation of a proper certificate, and only in this case, the mark for the missed examination will be computed from that obtained at the final examination.
· Students who miss a midterm for unjustified reasons (e.g.: no medical certificate, going to the exam at the wrong time or on the wrong day, etc.) will automatically get a mark of zero.
· Any requestfor reevaluation of a Midterm must be made within a week of its return by contacting the instructor.
· Marked Midterms that have not been picked up after two weeks of their return may be discarded.
B. champagne Page 3 9/12/2008