your mail

[Kai-Uwe Behrmann]

Am 22.10.03, 00:11 +0100 schrieb Bruno Postle:

> On Wed 22-Oct-2003 at 12:22:06AM +0200, Pablo d'Angelo wrote:
> >

> > I'm quite sure that a simple addition, like I did is the wrong
> > way. Probably one should transform the picture to HSV and divide
> > the V channel of though a brightness factor image.

This could been solved by Lab images where one had to change the L
(Luminance) channel only.

> > This brighness factor image would be created from a (smoothed)
> > shot of a subject with equal brighness by calculating: V(x,y) /
> > V_center for all x,y.
>
> The ptools gimp and photoshop plug-ins have a radial luminance
> correction tool, it works by an additive process rather than
> multiplication/division.

As I remember this is the only tool which does an correction for each rgb
channel. So this tool get aware of colored borders in fisheyes. All others
ignore this chromatic abberation error.

> Helmut had some reasoning for this that I can't remember (I can look
> it up).

Would be interessting.

> I works quite well, but there is no single parameter that you can
> use for any one lens.
>
> I wrote a similar 'multiplicative' version that works ok too:
>
>    http://bugbear.blackfish.org.uk/~bruno/imager-test/

Looks very simple code. Is it this easy? :o
Why did we write hugin in C++? ;)

> --
> Bruno
>

regards
Kai-Uwe