Differences

This shows you the differences between two versions of the page.

--- announcements [2015/10/23 20:37] – magiero
+++ announcements [2015/12/11 02:01] (current) – magiero
@@ Line 4: / Line 4: @@
 Check this page regularly for updates on the course.
+===== Dec. 10, 2015 =====
+Some notes on the final exam.  Similar style to all the other tests I have given you.  Here is the {{formulasheet.pdf | formula sheet}} I'll be giving you.  Exam will cover material from L9 to L15 inclusive (not L16-17).  This overlaps with Assignment 2 Question 18 on.  Of course don't just forget the tools we developed in the first part of the course (e.g. sinc's, power densities, Nyquist sampling, integrating simple things, etc.), but the focus will be on the systems we discussed in the latter half.  Like the midterm the exam questions will go in order from L9 to L15.  I'll split the exam into relatively small questions so if you see lots of questions don't panic, its probably just lots of small parts.  As usual if you don't understand a question just move on to the next or to a section that you feel most confident on.  Remember, understanding the question is part of the test (that is why I don't elaborate on questions, if I don't say anything at least you are in the same boat as your classmates).
+===== Dec. 1, 2015 =====
+Quiz 4 (this Thurs.) will just be concerned with Assignment 3 Questions 1-10 inclusive.  So there will not be any sequence detection, Viterbi stuff on this one.  I'll post the slides for the last couple of lectures tonight.
+===== Nov. 25, 2015 =====
+One thing that I forgot is that the X (convolution) matrix should be normalized give the equations used in the notes.  The reason is that in the MMSE case we have the I/SNR term which effectively normalizes the noise power (variance) to a signal power of 1 (i.e. the identity matrix I).  When that's the case the X should be normalized.  Since X is made out of h you are best to just normalize h before using it to make X.  Normalize the components of h by the square root of the sum of the elements squared [or in Matlab, normalize by sqrt(sum(h.^2))].
+===== Nov. 23, 2015 =====
+Here are the {{quiz3soln.pdf | quiz 3}} solutions.
+===== Nov. 19, 2015 =====
+For tomorrow's quiz, I can't ask you all of Assig. 3 since we haven't finished all the material in it.  Only questions 1-6 (inclusive) are legit for Quiz 3.  As is what we have discussed over in L9 and L10 and L11 up to (and including) 11.4.  Here is the {{quiz3formula.pdf | formula sheet}} I'll be attaching to Quiz 3.
+===== Nov. 16, 2015 =====
+Someone asked me about the quantizer for Lab 3.  Specifically the "formula" for it.  All I can really say is that it is pretty basic.  I implemented it in one line with MATLAB.  Remember all that a quantizer is is just an element that rounds off your input numbers.
+===== Nov. 11, 2015 =====
+Assignment 3 questions are posted.  I have updated the quiz 3 and 4 dates in the wiki (both delayed by 1 week).
+===== Oct. 27, 2015 =====
+Here are the {{midtermsoln.pdf | midterm solutions}}.
 ===== Oct. 23, 2015 =====
-Midterm notes.  Midterm will be similar in style to last year's (see last year's wiki).  I plan to make it roughly out of the same number of points (no guarantees) but I may distribute the questions differently (i.e. you may see more than 4 net questions, but the overall ask should be about the same).  As I write it I'll be referring to material between L1 and L9 up to (and including 9.3).  I'll also be looking at Assig. 1 and 2 and the labs 1 and 2.  I'll also be thinking about questions you asked in class and stuff that I thought we emphasized more deeply (I don't know that off hand now, but assume my memory will be jogged when reviewing the material).  I don't expect you do to massive derivations, but I do expect you to be able to manipulate math a bit.  That is, there may be occasions (not a lot, but some) where I won't be asking you stuff that is exactly on the [[finalinfosheetf14.pdf | formula sheet]] that I will be giving you, but may require a couple more manipulations in order to get the final answer.  In general for the midterm I'm not interested in any epic derivations (as I think you should see from the previous year's test).  Some of you asked about "theory".  Yes, I think I as theory, but it seems some of you have different ideas about what theory is.  For example if I ask you what the blocks of a communication system might look like or be arranged this is "practice" not "theory".  There is no "theoretical" reason why the blocks have to be arranged like that (there are good practical reasons however).  So, yes I often like to ask practice questions (but that does not guarantee that I will do that).  For me theory is are such things as Nyquist sampling, Fourier spectra, PDFs, optimum decisions, etc.  In general I like to keep the questions "fair" (to me that means if we talked about it in class to some reasonable level and/or assigned it and/or dealt with it in labs then it should be fair game).  Some of the questions in the assignments are indeed complex and in theory it would be "fair" to ask them in their entirety, but I often (again no guarantees) prefer to jump into them part way and have you answer them from there.  Another important thing is to know the subject jargon, often knowing what the question actually is asking is the most difficult part because the question uses language particular to the field.  In the midterm if you ask for some clarification on a question I typically just ask you to do the best you can with the question as posed because everything you need to arrive at a reasonable solution has been provided (or can be implied by the suitably prepared student).  Hope that helps.
+Midterm notes.  Midterm will be similar in style to last year's (see last year's wiki).  I plan to make it roughly out of the same number of points (no guarantees) but I may distribute the questions differently (i.e. you may see more than 4 net questions, but the overall ask should be about the same).  As I write it I'll be referring to material between L1 and L9 up to (and including 9.3).  I'll also be looking at Assig. 1 and 2 and the labs 1 and 2.  I'll also be thinking about questions you asked in class and stuff that I thought we emphasized more deeply (I don't know that off hand now, but assume my memory will be jogged when reviewing the material).  I intend to write the midterm question in the order that we covered the material in class (so first page should be L1 and/or L2 stuff and last page should be L8 and/or L9 stuff).
+I don't expect you do to massive derivations, but I do expect you to be able to manipulate math a bit.  That is, there may be occasions (not a lot, but some) where I won't be asking you stuff that is exactly on the {{finalinfosheet4214f14.pdf | formula sheet}} that I will be giving you, but may require a couple more manipulations in order to get the final answer.  In general for the midterm I'm not interested in any epic derivations (as I think you should see from the previous year's test).  Some of you asked about "theory" questions.  Yes, I think I ask theory, but it seems some of you have different ideas about what theory is.  For example if I ask you what the blocks of a communication system might look like or be arranged this is "practice" not "theory".  There is no "theoretical" reason why the blocks have to be arranged like that (there are good practical reasons however).  So, yes I often like to ask practice questions (but that does not guarantee that I will do that).  For me theory is such things as Nyquist sampling, Fourier spectra, PDFs, optimum decision levels, probabilities, etc.  In general I like to keep the questions "fair" (to me that means if we talked about it in class to some reasonable level and/or assigned it and/or dealt with it in labs then it should be fair game).  Some of the questions in the assignments are indeed complex and in theory it would be "fair" to ask them in their entirety, but I often (again no guarantees) prefer to jump into them part way and have you answer them from there.  Another important thing is to know the subject's jargon, often knowing what the question actually is asking is the most difficult part because the question uses language particular to the field.  In fact this is part of the test.  In the midterm if you ask for some clarification on a question I typically just ask you to do the best you can with the question as posed because everything you need to arrive at a reasonable solution has been provided (or can be implied by the suitably prepared student).  Hope that helps.
 ===== Oct. 19, 2015 =====