Announcements:
Last modified at 1:15 pm (EDT) on Friday, 12/12/08.
Final Exam for On-Campus: Wed., Dec. 17, 7-9 pm, in MSEE B012.
Coverage for Final Exam.,
Misc Notes for Final Exam.
Matlab Hmwk #3: P&M Prob. 7.29, 7.30. ** NEW DUE DATE*: December 10.
Matlab 3: Two text Problems from Chap. 7. .
Text Reconstruction Formula for Prob. 7.29. .
Note: Prob. 7.30. f_1 = 1/128 NOT 1/18 -- typo!
Also, plot magnitude of DFT in each case.
Note: Prob. 7.29. For parts (b) and (c), use
reconstruction formula (7.1.13) at bottom of pg. 453.
(b) Use N=21 rather than N=20, and use the
sequence x[n]=a^|n-D|, n=0,1,...20, where D=(N-1)/2=10, and multiply X(w) by exp(-j10w).
(c) Use N=101 rather than N=100, and use the
sequence x[n]=a^|n-D|, n=0,1,...100, where D=(N-1)/2=50, and multiply X(w) by exp(-j50w).
(e) Use the time-domain aliasing formula in (7.1.4) on
page 450. Use three terms:
x[n-N] + x[n] + x[n+N], for n=0,1,...N-1.
Plot what this formula yields on the same graph
as a plot of the IFFT of N samples of the original spectrum
in the interval from 0 to 2pi.
Matlab Hmwk #2. **NEWER Due Date**: Monday, November 17.
Note that there is a scaling by j that results with the
h0[n]={1,j} and h1[n]={1,-j} filter pair. That scalar
problem can be removed by factoring out the half-angle.
Specifically, replace h0[n]={1,j} by h0[n]={exp(-j pi/4) , exp(j pi/4)}
and replace h1[n]={1,-j} by h1[n]={exp(-j pi/4) , -exp(j pi/4)}.
MatlabHmwk2F08.pdf .
PRRC4chan.m, PR4chan.m.
VIP Exam Dates:
Exam 1 held Fri., Sept. 26:
pdf Soln: pdf Histogram: jpg.
Exam 2 held Friday, Oct. 24:
pdf
Soln: pdf Histogram: jpg.
Exam 3 held Monday, Nov. 24:
pdf
Soln: pdf Histogram: jpg.
Course Information and Course Syllabus for Fall 2008:
General Course Info for Fall 2008: pdf
Course Syllabus for Fall 2008: pdf.
TA Info: Jianqi Wang, e-mail: wang128@purdue.edu.
Office: MSEE B022. Office Hours: TBD.
Your primary homework assignment is to solve my old exams (posted
below) without looking at the solutions in a timed setting. You can look
at the solutions before and after attempting the exam, but
it is important to work through them without looking at the solution.
A few students have asked for homework problems out of the textbook,
but I think the best homework, in terms of preparing for the
exam, is to work through my old exams.
Matlab Hmwk #1. Due: Monday, Sept. 22.
pdf
Prob. 2.65 in Text (with changes
discussed in class and described below.)
On-campus students MUST turn in a paper print-out in class.
Note that I have made several changes to the problem
as it is posted in the textbook: pdf and
Prob2_65.txt.
Recommend against
using xcorr to do the cross-correlation -- sorry for using that in
several of my demos -- just use convolution to do correlation as
discussed in class: ryx = conv(y,x(end:-1:1)) and throw away the
first first M-1 values of ryx (where M is the code length) since
those correspond to negative time-shifts and the problem only asks to
plot for positive time-shifts.
Matlab Hmwk #2. **New Due Date**: Friday, November 14.
MatlabHmwk2F08.pdf .
PRRC4chan.m, PR4chan.m,
Matlab Hmwk #3: P&M Prob. 7.29, 7.30. DUE: December 8.
Matlab 3: Two text Problems from Chap. 7. .
Note: Prob. 7.30. f_1 = 1/128 NOT 1/18 -- typo!
Also, plot magnitude of DFT in each case.
Note: Prob. 7.29. For parts (b) and (c), use
reconstruction formula (7.1.13) at bottom of pg. 453.
(b) Use N=21 rather than N=20, and use the
sequence x[n]=a^|n-D|, n=0,1,...21, where D=(N-1)/2=10.
(c) Use N=101 rather than N=100, and use the
sequence x[n]=a^|n-D|, n=0,1,...100, where D=(N-1)/2=50.
(e) Use the time-domain aliasing formula in (7.1.4) on
page 450. Use three terms:
x[n-N] + x[n] + x[n+N], for n=0,1,...N-1.
Plot what this formula yields on the same graph
as a plot of the IFFT of N samples of the original spectrum
in the interval from 0 to 2pi.
TA Info: Jianqi Wang, e-mail: wang128@purdue.edu.
Office: MSEE B022. Office Hours: TBD.
Prof. Zoltowski's Office Hrs: TBD plus send e-mail to make appt.
Some history and miscellaneous course information:
This course emphasizes applications of
Digital Signal Processing (DSP) in compact disc (CD) players, wireless
communictions, radar, and speech processing.
Professor Zoltowski has
taught this course the Fall of every year since 1990. During
the Fall of even years, this course has been offered
over the Indiana cable network IHETS, the National
Technological University (NTU) satellite network,
and to all Purdue Continuing Engineering Education sites since 1990.
A little humor:
Top Ten reasons to take EE538.
|
Module Notes and Demos
-
getspeech.m Needed for course M-files,
this is the platform independent way to read the .wav files here.
- "WAV"-files : [
enter(female)
enter(male)
erase(female)
help(female)
w(female)
zero(female)
one(male)
] The only useful information on the front
page of each module below is the relevant sections in
the P&M Text and the relevant Matlab demo files. The
rest of the front page of each module should be ignored.
The goal is to work through three modules per week, or
1.5 modules per lecture, on average. Additional modules will
be added on material towards the end of the course coverage.
-
Weeks 1 and 2.
- Module 1:
Text Outline Scan (pdf);
Text Chap. 1 Scan (pdf);
Module 1 (pdf);
aliaseg.m,
- Module 2:
Module2.pdf;
Convolution Properties: Brief Notes from Session 2;
notcheg.m
- Module 3:
Session 4 Outline (Roadmap);
Supplemental Notes on Convolution (from class);
Text Chap. 2, Part 1 Scan (pdf);
Text Chap. 2 Part 2 Scan (pdf);
Session5_RoadMap.pdf ;
Module3.pdf ;
Module3_Addendum.pdf ;
- Fig. 2.22 in Ed. 3 is Fig. 2.3.1 in Ed. 4.
- Fig. 2.23 in Ed. 3 is Fig. 2.3.2 in Ed. 4.
- Fig. 2.24 in Ed. 3 is Fig. 2.3.3 in Ed. 4.
- CDMAeg1.m,
cdmaeg.m,
gold1.m,
gold2.m,
gold3.m.
Add'l CDMA and OFDM Matlab Examples:
walsh_cdma_eg.m;
ofdm_eg.m.
-
Week 3.
- Module 4:
pdf
Text Chap. 3, Part 1 Scan (pdf);
Text Chap. 3, Part 2 Scan (pdf);
- Module 5:
pdf,
Chap. 5 Graphical Frequency Response (pdf);
notcheg2.m,
- Module 6:
Session9Roadmap.pdf ;
pdf
Chap. 5 Notch Filters (pdf);
Note on All-Pass Filters;
zpgui.m, zpgui3.m,
-
Week 4.
- Module 7:
CT Fourier Transform: Properties & Pairs;
DT Fourier Transform: Properties & Pairs;
Alternative Derivation of CTFT=DTFT Relationship;
Scan of First Section of Text Chap. 6;
pdf, pdf,
aliaseg2.m,
aliaseg3.m,
- Module 8:
pdf
-
Week 5.
- Module 9:
Quick Review of Ideal D/A Conversion;
pdf
upsamplex2eg1.m,
- Module 10:
pdf
upsample2eg2.m,
ZOHeg2.m,
- Module 11:
pdf,
subbandeg3.m,
Supplemental Notes on Hilbert Transform;
Supplemental Notes on VSB Modulation;
Final Words on Efficient Upsampling;
-
- Module 12:
pdf
quantizeb2.m.
-
Week 6.
- Review for Exam 2.
- Solve Exam 2 Fall 2005.
- Solve Exam 2 Fall 2007.
- Solve Exam 2 Fall 2004.
- Solve Exam 1 Fall 2002. Problem 3.
-
Week 7.
- Module 13:
pdf ;
Elliptic Filters,
- Module 14:
pdf
- Module 15:
pdf
buttereg.m,
chebyeg1.m,
chebyeg2.m,
ellipeg.m,
-
Week 8.
- Module 16:
pdf , Text: Equi-ripple FIR Filter Design
FIRlowpasseg.m,
FIRbandpasseg.m,
- Module 17:
pdf
deriveg.m,
deriveg2.m,
hilberteg.m,
-
Week 9.
-
Chap.7 Frequency Domain Sampling Notes;
- Module 18:
pdf
- Module 19:
pdf
DivideConquer.m,
sineDFTeg1.m,
sineDFTeg2.m,
sineDFTeg3.m,
- Module 20:
pdf
timealias.m,
-
Week 10.
-
Text notes on efficient DFT/FFT based computation;
Efficient Computation of DFT of two
real-valued sequences plus efficient computation
of 2N-pt DFT of real-valued sequence.
EFF_FFT_Real.m
Illustration of overlap-save method using FFT's:
OverlapSaveEff.m ;
Text notes on Properties of the DFT.;
Notes on DFT Based Processing.;
Basic DFT Pair.;
Time Domain Aliasing of Multi-Pole Causal Signals.;
-
Review for Exam 3. Work Problem 4 from Final
Exam of Fall 98. PerfReconstruct.m
-
Spectrogram Examples for Session 31: voweleg.m ;
voweleg2.m ;
vowelwin.m .
0af1s1t0.wav ;
0ef1s1t0.wav
-
Module 21: pdf
windowseg.m,
trunceffects.m,
windowseg2.m,
windowedsines.m,
-
Week 11.
- Module 22:
pdf
- Module 23:
Condensed Overview/Derivation of AR and ARMA spectral estimation methods:
pdf.
Illustrative Example of Linear MMSE estimation:
pdf.
pdf
SOSextrap.m,
SOSviaAR.m,
ARspecest.m.
-
Week 12.
- Module 24:
pdf
SOSestviaAR.m,
ARMAviaAR.m.
- Module 25:
pdf
Derivation of Levinson-Durbin Algorithm:
pdf. Minus
sign on RHS of last eqn at bottom of last page should
be plus sign.
-
Week 13.
- Module 26:
pdf
Notes on MA(q) random process:
pdf.
YWvsULS.m,
ARMA2stepest.m,
- Module 27:
pdf
MinVarforSOS.m,
MinVarforARMA.m.
-
Derivation of Levinson-Durbin Algorithm.
pdf. Minus
sign on RHS of last eqn at bottom of last page should
be plus sign.
-
Week 14.
- Session 38. Review for Exam 3.
Solve Exam 3 from Fall 2001.
- Session 39. Review for Exam 3.
Solve Exam 3 from Fall 2000.
- Module 28: Adaptive Filtering.
pdf
- Module 29:
pdf
NoiseCancel.m.
-
Week 15.
Exam 3 during Session 41. Thanksgiving Break.
-
Week 16.
- Module 29:
pdf
AdaptCancel.m.
- Module 30:
pdf
MinVarforSOS.m.
- Module 31:
pdf
CancelTone.m.
- Module 32:
pdf
FIRequalizer.m.
Prob529.m.
- Module 33:
pdf
AdaptEqualize.m.
You need Adobe Acrobat Reader 2.1 or later to view PDF files. The
latest version is available freely at
www.adobe.com.
|
Homeworks and Solutions
-
Homework Problems from Proakis Text
- Hmwk #1: Problems: 1.6, 1.7, 1.8, 1.9, 1.11, 2.10, 2.11, 2.13,
2.45, 2.46, 2.61. Solution: pdf
-
- Hmwk #2: Problems: 3.43, 3.49, 3.51, 7.3, 7.4, 7.7.
Solution: pdf
-
- Hmwk #3: Problems: 4.51, 4.32 - change input to x[n]=sin(pi n /4) / (pi n), 4.47, 4.49, 4.50, 4.76(a), 4.93, 4.100. Solution: pdf
-
Matlab Based Homeworks
- Matlab Hmwk #1:
Due date: Monday, Sept. 22.
On-campus students MUST turn in a paper print-out.
The problem is 2.65 in the textbook with some modifications.
Note that I have made several changes to the problem
as it is posted in the textbook: pdf
Prob2_62.txt.
Recommend against
using xcorr to do the cross-correlation -- sorry for using that in
several of my demos -- just use convolution to do correlation as
discussed in class: ryx = conv(y,x(end:-1:1)) and throw away the
first first M-1 values of ryx (where M is the code length) since
those correspond to negative time-shifts and the problem only asks to
plot for positive time-shifts.
-
-
Matlab Hmwk #2. Due Friday, November 7 (tentative, may be due later than that.)
MatlabHmwk2F08.pdf .
-
- Matlab Hmwk #3: P&M Prob. 7.29, 7.30. DUE: Thurs, Nov 30 (Session 28.)
- Note: Prob. 7.30. f_1 = 1/128 NOT 1/18 -- typo!
- Also, plot magnitude of DFT in each case.
- Note: Prob. 7.29. For parts (b) and (c), use
- reconstruction formula (7.1.13) at bottom of pg. 453.
- (b) Use N=21 rather than N=20, and use the
sequence x[n]=a^|n-D|, n=0,1,...21, where D=(N-1)/2=10.
- (c) Use N=101 rather than N=100, and use the
sequence x[n]=a^|n-D|, n=0,1,...100, where D=(N-1)/2=50.
- (e) Use the time-domain aliasing formula in (7.1.4) on
page 450. Use three terms:
x[n-N] + x[n] + x[n+N], for n=0,1,...N-1.
Plot what this formula yields on the same graph
as a plot of the IFFT of N samples of the original spectrum
in the interval from 0 to 2pi.
- DUE: Some time in November.
-
-
Fall 2008 Exams plus Solutions
-
- Exam #3 from Fall 2008:
pdf
Soln: pdf.
Histogram: jpg.
-
- Exam #2 from Fall 2008:
pdf
Soln: pdf.
-
- Exam #1 from Fall 2008:
pdf
Soln: pdf.
-
-
-
Fall 2007 Exams plus Solutions
-
- Exam #3 from Fall 2007:
pdf
Soln: pdf.
-
- Exam #2 from Fall 2007:
pdf
Soln: pdf.
-
- Exam #1 from Fall 2007:
pdf
Soln: pdf.
-
-
Fall 2006 Exams plus Solutions
-
- Final Exam from Fall 2006:
pdf.
-
- Exam #3 from Fall 2006:
pdf
Soln: pdf.
-
- Revised Exam #2 from Fall 2006:
pdf
-
- Exam #1 from Fall 2006:
pdf,
Soln: pdf.
Histogram: pdf.
-
-
Fall 2005 Exams plus Solutions
-
- Final Exam from Fall 2005:
pdf.
Soln: pdf.
Or: pdf.
-
- Exam #3 from Fall 2005:
pdf,
Soln: pdf.
-
- Exam #2 from Fall 2005:
pdf,
Soln: pdf.
Histogram: pdf.
-
- Exam #1 from Fall 2005:
pdf,
Soln: pdf.
Histogram: pdf.
-
-
Fall 2004 Exams plus Solutions
-
- Final Exam from Fall 2004:
pdf.
-
- Exam #3 from Fall 2004:
pdf,
Soln: pdf.
-
- Exam #2 from Fall 2004:
pdf,
Soln: pdf.
-
- Exam #1 from Fall 2004:
pdf,
Soln: pdf.
-
-
-
Fall 2003 Exams plus Solutions
-
- Final Exam from Fall 2003:
pdf.
-
- Exam #3 from Fall 2003:
pdf,
Soln: pdf.
Histogram & Score Stats: pdf
-
- Exam #2 from Fall 2003:
pdf,
Soln: pdf.
Histogram & Score Stats: pdf
-
- Exam #1 from Fall 2003:
pdf,
Soln: pdf.
Histogram & Score Stats: pdf
-
-
Fall 2002 Exams plus Solutions
-
- Final Exam from Fall 2002: pdf.
-
- Exam #3 from Fall 2002:
pdf,
Soln: pdf.
Alternative Soln to Problem 3 (c):
pdf.
Extra Review Problem from recent QE:
pdf.
Only look at 1st problem; second problem relevant to Final
-
- Exam #2 from Fall 2002:
pdf,
Soln: pdf.
Supplemental Soln: pdf.
-
- Exam #1 from Fall 2002:
pdf.
Solution to Prob. 2: pdf;
Supplemental Notes: pdf;
For Problem 3, see soln to Prob 1 of Exam 2 from Fall 99: pdf.
Histogram of Scores for Exam 1: E1Hist_Off.pdf;
E1Hist_On.pdf
-
-
-
Fall 2001 Exams plus Solutions
-
- Final Exam from Fall 2001:
pdf.
-
- Exam #3 from Fall 2001:
pdf
ps,
Solution: pdf.
-
- Exam #2 from Fall 2001:
pdf
ps,
Solution: pdf
-
- Exam #1 from Fall 2001:
pdf
ps,
Solution: pdf
-
Fall 2000 Exams plus Solutions
-
- Exam #3 from Fall 2000:
pdf
ps,
Solution: pdf
Histogram: pdf
-
- Exam #2 from Fall 2000:
pdf
ps,
Solution: pdf
Histogram: pdf
-
- Exam #1 from Fall 2000:
pdf
ps
Solution: pdf
Histogram: pdf
-
- Final Exam from Fall 2000:
pdf
ps
Solution: in class on Tues., Nov. 27.
-
-
Additional Practice Exams
- Final Exam from Fall 1999:
pdf
ps
Solution: pdf
-
- Final Exam from Fall 1998:
pdf
ps
Solution: pdf
-
- Exam #3 from Fall 1999:
pdf
ps
Solution: pdf
-
- Exam #3 from Fall 1998:
pdf
ps
Solution: pdf
-
- Exam #3 from Fall 1995 (only Probs 3 and 4
are relevant to our Exam 3):
pdf
ps
Solution: pdf
- Exam #3 from Fall 1996:
pdf
ps
Solution: pdf
-
- Exam #2 from Fall 1999:
pdf
ps
Solution: pdf
-
- Exam #2 from Fall 1998:
pdf
ps
Solution: pdf
-
- Exam #3 from Fall 1996:
pdf
ps
Solution: pdf
- Exam #2 from Fall 1995 (only Probs 1 and 3
are relevant to our Exam 2):
pdf
ps
Solution: pdf
- Exam #3 from Fall 1995 (only Probs 3 and 4
are relevant to our Exam 2):
pdf
ps
Solution: pdf
-
- Exam #1 from Fall 1998:
pdf
ps
Solution: pdf
- Exam #1 from Fall 1999:
pdf
ps, Solution: pdf,
pdf
-
|
