Latest revision as of 07:05, 9 April 2013

Lab 2 Discussion, ECE637, Spring 2008

comments:

Rand --matias.zanartu.1, Sun, 27 Jan 2008 16:32:23 -0500 reply Please note that command rand(m,n) generates uniformly distributed random numbers (mxn) between 0 and 1 (not between -0.5 and 0.5). In addition, you cannot specify the mean or the variance directly with this command. However, this information is needed to derive (and plot) the theoretical expression for tex:S_y(e^{-j\mu}, e^{-j\nu}). Use the fact that the sequence was derived from a uniform random variable. The link below (from wikipedia) is useful to get the info you need from the input sequence. Remember that the autocorrelation function evaluated at the origin is equal to the variance of the sequence. I hope this helps. Good luck.

http://en.wikipedia.org/wiki/Uniform_distribution_%28continuous%29

Regarding hamming window --shivani.g.rao.1, Sun, 27 Jan 2008 23:28:43 -0500 reply Does the use of hamming widow decrease the maximum amplitude of an FFT?

Applying hamming window --sungye.kim.1, Mon, 28 Jan 2008 22:52:44 -0500 reply For #1 of lab2, applying hamming window to each 64x64 window means the element multiplication between hamming window and each window. Not matrix multiplication. When I used matrix multiplication, I got weired result. I hope this helps.

Rand --alfa.satyaputra.1, Wed, 30 Jan 2008 12:20:43 -0500 reply Now that you mention it, I'm wondering if its ok to use randn instead when we want to generate the random variable between -0.5 and 0.5...

Rand --rong.zhang.5, Thu, 31 Jan 2008 19:52:38 -0500 reply We need to generate uniformly distributed samples, rand() should be used. In MATLAB, randn() is used to generate normal distributed samples N(0,1).