Matlab and filter design
Digital Filtering and Digital Filter Design is a big subject,
a major aspect of digital signal processing.
See my digital signal processing notes.
Matlab tools
Matlab offers several approaches to this, in particular there
is an independent approach in the System Identification Toolbox.
- filterbuilder - does filter design
- fvtool(var); - displays filter response.
I type "filterbuilder" at the matlab command line and this fires up a GUI
that lets me select filter types and enter parameters. When it is all done
it yields a new workspace variable (such as Hpn2) in my workspace.
Using the matlab command "whos Hpn2" shows me this is a matlab object
of type "dfilt.df2sos" (i.e. a digital filter).
LTI models
Matlab has 4 ways to represent LTI models and can convert from one to
another for you. There are TF (transfer function), SS (state space),
ZPK (zero pole gain), and FRD (frequency response data) models.
The tf operator will create a TF object from two arrays (numerator and denominator).
This can then be converted to a state space model
num = [1,0];
den = [1 2 1];
f = tf(num,den);
s = ss(f);
Alternately use:
f = filt(num,den);
Simulink and filters
We are currently using two types of filters in our simulink models:
- Discrete Transfer Functions - defined by numerator and denominator
arrays of values, implemented by SOS algorithms.
- LTI (Linear Time Invariant systems) - defined by matlab objects
of type "tf", inplemented by state space models.
Feedback? Questions?
Drop me a line!
Tom's Computer Info / tom@mmto.org