CE804 Digital Signal Processing CE804 Digital Signal Processing
CE804 Digital Signal Processing CE804 Digital Signal Processing
Revision lecture
Remember to spend some time reading both questions in section B before deciding which one to answer - dont waste time by answering both questions in Section B
Revision Lecture CE804 2010/2011 2
Division of questions
All lecture material will be tested (except IIR sections in Lecture 8)
Material to review
The exam questions will consist of a mixture of three types:
Theory example of questions: list the something of something discuss the advantages and disadvantages of something explain the principles behind something. Computation these are computation type of questions (some will require heavy use of calculator), see the worked examples in quiz questions for examples Analysis example of questions: why is some method better than some method The figure show results of some analysis of something, explain the reason behind this,.What would be the effect of doing something for some signal.
Additional instructions
If you make any assumption in answering the questions, state them clearly. Describe each step in your calculation. This will be useful in securing marks even if your final answer is wrong or if for some reason you do not have the time to complete the calculations (or if your calculator fails). Write something even if you are not sure if it is related to the answer dont hand in empty pages. You may not get any marks, but it is worth the try. Dont panic during exam. Make sure you get plenty of sleep the day before. Dont forget your scientific calculator! The use of MATLAB codes will not be tested in the exam.
10
11
12
Study guide (Lecture 3 - Signal conditioning) conditioning From this lecture, you should know
Basic mathematics mean, standard deviation, variance, covariance, correlation, SNR, etc.
Uses of these basic maths in signal analysis example given two signals, obtain the correlation between them
Noise reduction methods like ensemble averaging, median filtering, moving average filtering, principal component analysis know both the theory and how to compute for each method. PCA computation is important but due to time constraints in exam, full computation will not be tested.
13
14
15
16
Study guide (Lecture 4 - Fourier analysis and spectral estimation estimation) From this lecture, you should know how to compute Discrete Fourier Transform PSD estimation using classical methods (periodogram, Welch) Remember all the equations in this lecture DFT related concepts like frequency number, actual frequency, picket fence effect, zero padding, truncation, windowing, detrending, effects of signal length on DFT, etc. Theory behind PSD estimation using periodogram and Welch approaches
17
18
19
20
21
Study guide (Lecture 5 digital filtering) From this lecture, you should know How to perform filtering using
Direct filtering in frequency domain Sum and difference (SD) FIR filtering IIR filtering
Direct filtering and SD filtering will be tested in exams - both theory and computation wise.
Remember the equations for SD filtering
IIR filtering will be tested as theory/analytic questions (IIR types, advantages, disadvantages, phase response, etc.) only as you wont have access to MATLAB in exams (so no computation). The basic concepts in filtering like order selection, 3dB cut-off, and filter specifications (ripples, passband, stopband, etc.) are important.
Revision Lecture CE804 2010/2011 22
23
50
40
40
30
30
20
20
10
10
10
20
30 40 Frequency (Hz)
50
60
10
20
30 40 Frequenc y (Hz)
50
60
24
25
26
27
Convolution
From lecture slides
28
29
Study guide (Lecture 7 transfer functions and digital filter structures) From this lecture, you should know
All pass transfer function How to obtain zero-phase transfer function Linear phase FIR transfer function different types Frequency response using z-transform given the transfer function Comb filters Canonic/noncanonic filter forms Equivalent structures through transpose operation Direct and cascade form FIR filter Polyphase FIR structures Using symmetry/antisymmetry property of linear phase FIR filters
to reduce the number of multipliers to obtain the frequency response
30
Given the zeros, obtain the poles for all pass transfer function
Simple - do inverse!
31
Cascade form
Obtain the form in the required order and draw the block diagram Remember the basic block
32
Polyphase structures
Study guide (Lecture 8 FIR filter design) From this lecture, you should know
FIR filter design steps using the window method Gibbs phenomenon Window functions, effects of main/side lobes Length (order) selection using Kaiser formulae Symmetry/anti-symmetry to obtain linear phase response for FIR filter IIR filter structures will not be tested
34
Importance of symmetry/anti-symmetry in the coefficients for linear phase response (see slide 28, Lecture 8)
35
37