Rich Kozick
Spring, 1997
EE 329: Laboratory 8
FIR Filter Design with the Window Method
You will be given samples of an audio file that is corrupted by
interference.
Your assignment is to:
- Analyze the audio file, and specify
requirements on the frequency response of an FIR digital
filter that will remove the interference.
Your filter should remove the interference as much as possible,
but also preserve as much of the original signal as possible.
(These requirements are conflicting, so you will need to
decide on a compromise.)
- Design FIR digital filters to remove the interference.
Use two approaches to design your digital filters:
- By placing zeros in the z-plane, and
- Using the
"window method"
described in Section 6.5 of the text.
Which design achieves better results?
Download the data file and load it into Matlab using
either of the following methods:
- Download the audio file
interfere.au
into your account, and load it into Matlab with the command:
x = auread('interfere.au');
- Downlaod the text file
interfere.dat
and load it into Matlab with variable name x with the commands:
load interfere.dat
x = interfere;
clear interfere
You can listen to the audio file in Matlab with the command
sound(x).
If you would like to here the original, uncorrupted music,
it can be loaded into Matlab with the command:
xgood = auread('/home/pleiades/ACM/snd/spacemusic.au');
Matlab has built-in functions for generating many of the
common window functions. For a listing of the available
window commands, type
help signal and look at the "Windows" section.
Lab Report:
Please submit a brief report describing your design procedure
and your results by Friday, April 11.
Also, I will ask each of you to demonstrate the effectiveness of
your filter to the class.
As we talk about other types of digital filters, feel free to
try them out on this particular signal processing problem.
Thank you.