filter.oct -- how to use it?!?

[E. Joshua Rigler]

I recently made a fool of myself on this list by asking a dumb question
about FFTs, so what harm can it do for me to ask another dumb
question...

How are you supposed to use the Octave filter function?  If I do
something like:


 [yn] = filter (0, [1,-.5,-.1], [0,0,0], [10,10])

...the output is:

  yn = 10  15  8.5

...when *I* expect it to be:

  yn = 6.0  4.0  2.6


Isn't the "initial state" the same as saying [y(n-na) ... y(n-1)]?  I am
assuming that filter operates like:

  y(n) = b(1)*x(n) + b(2)*x(n-1) + ... + b(nb+1)*x(n-nb)
                   - a(2)*y(n-1) - ... - a(na+1)*y(n-na)

(where b = x = 0)

It seems like filter is prepending (na-1) zeros to the inital state,
filtering all these zeros, then adding the value to the first element of
the initial state.  Then bumping up to  y(n+1), filtering again, and
adding to the second element of the initial state, and so on.  Why?  I
just want to take the previous n values in a time series and filter them
to predict the (n+1st) value.  This is what the "Direct Form II
Transposed Structure" should be doing for me isn't it?  Am I just being
really dense?

Be kind ;^).

-EJR



-------------------------------------------------------------
Octave is freely available under the terms of the GNU GPL.

Octave's home on the web:  http://www.octave.org
How to fund new projects:  http://www.octave.org/funding.html
Subscription information:  http://www.octave.org/archive.html
-------------------------------------------------------------