Hello everyone,

Here are a few important announcements regarding homework 7, examples from class and the second Matlab assignment:

• I handed back your graded Homework 7 assignments today. The grade consists of 8 points for the problems, 1 point for having the Matlab plots prepared for class on 11/13 and 1 point for participating in the class discussion on 11/13. Thus the total score was out of 10.


Solutions to Problem 7.4 (not turned in) and the Overlap-Add/Overlap-Save example from lecture on 11/22 will be available in my office late tomorrow and in class next week. I will be happy to give you a copy of the solutions if you can show me that you have worked both of these problems. I think they are useful examples, and I'd rather you try them yourselves before looking over my solutions.


• Solutions to Matlab project 2 will be available outside my office door after 5pm tomorrow (Thursday).

Finally, some hints about IIR filter design using Matlab:

• Matlab programs such as butter and ellip require that you define passband and stopband ripple in terms of dB. The convention for doing this is

  Rp = passband ripple = -20*log10(1-delta1)

  Rs = stopband ripple = -20*log10(delta2)

  where delta1 and delta2 are the pass and stopband tolerances we defined in lecture on Monday.



• When using the elliptic filter, you will note that Matlab's program returns a filter that has a maximum gain of 1. That is, no ripples go above 1 in the passband. This is different than our DT specifications because we set up our tolerances to allow ripple above and below 1. You can easily scale your specifications, design the filter, and then re-scale the filter to get a filter with the proper ripple that is 1 +/- delta1. Figuring out the scaling you need to use is pretty easy.
  
  Note: the reason Matlab scales things to a maximum gain of 1 is that analog filters made out of passive components (resistors, capacitors, etc.) cannot amplify the signal. This is not an issue in discrete time.

We will discuss these and other practical issues on Monday, but I wanted you to have this info in case you'd like to finish the Matlab project 3 over the weekend.

Good luck working on it. Please email me if you have questions.

-Kathleen Wage (kwage@gmu.edu)