[SciPy-Dev] trying to understand how zi is used in lfilter

[Robert Dodier]

Hi,

I'm answering my own question here -- I think I got it sorted out.

zi as computed by lfilter_zi is supposed to be internal state of the filter, which are represented by the variables d[0] through d[N - 1] in the comment header for lfilter. So the way that zi comes into play is by assigning zi[k] to d[k] at the outset of lfilter. From there on out, the d values are updated from x and y values.

Maybe this information will be helpful to someone someday.

best,
Robert Dodier