[ptx] why f(r) mixed symmetry?

[Pablo d'Angelo]

Hi Klaus,

> A question probably best to ask Helmut Dersch if he were still active...
> 
> Why are parameters a, c in the lens formula f(r)= r*(1+a*r+b*r^2+c*r^3) ?
> With the rotational symmetry of almost all lenses and an analytical ansatz
> shouldn't the function f be point symmetric, i.e. -f(-r)=f(r) and thus
> only containing odd powers of r?

r is always positive, so its ok to use a mixed polynomial.

> How difficult would it be to add a parameter d*r^5 into the
> hugin/nona/Panotools framework? (Yes I know "d" is already used up)

Hmm, that would need smaller changes a quite a lot of places.

> [and I see fudge parameters useful when photos are not parallaxe free]

I think in this case it is not a good idea to use parameters that affect
the whole image, and thus also the parts without parallax errors.
I believe it's much better to use a local algorithm for these cases,
like morphing (unfortunately not supported by hugin and nona right now,
I don't have enough time to work on it...)

ciao
  Pablo