[ptx] fulla bugs

[Pablo d'Angelo]

Hi Bruno,

> > On Mon 05-Jun-2006 at 01:40 +0100, Pablo d'Angelo wrote:
>> >>
>> >> Fulla had some serious bugs when combining tca and distortion
>> >> correction coefficients, that could also lead to incorrect red and
>> >> blue coefficients, even if no green coefficients were given.
> >
> > Thanks, though I still get a mismatch between the colours, the red and
> > green channels seem to be remapped differently to blue when all r, g & b
> > parameters are given.

Actually, this is what should happen. My thought was that the user specifies
the geometric distortion with the -g parameter. Then the -r and -b parameter
specify the tca correction parameters, which are applied on top of the
geometric parameters. so:

-r 0:0:0:1 -g 0:-0.1:0:1.1 -b 0:0:0:1

should lead to the same result as specifying -g alone.

This way, the used doesn't have to combine geometrical and tca correction
coefficents for the red and blue channel himself.

> > It also produces no output with 16bit TIFF input, 8bit files are ok.

Have to check that.

> > Vignetting seems to be ok, though which way around do I set the gamma?
> > I'm taking the 16bit linear output from dcraw and applying a gamma of
> > 2.2 like so:
> >
> >  convert -gamma 2.2 dsc_0008.ppm dsc_0008.tif
> >
> > When I correct this TIFF file with fulla, do I use a gamma parameter of
> > 2.2 or 0.45?

You should use gamma 2.2, like in PTStitcher, or just apply the gamma after
fulla has been run.

ciao
  Pablo