Warp <war### [at] tagpovrayorg> wrote:
> On 12/02/2010 11:26 AM, clipka wrote:
> > Am 02.12.2010 01:01, schrieb Jaap Frank:
> >
> >> That's odd, because for me the 3.6 side is correct and the 3.7 side is
> >> far too bright.
> >
> > Actually /that/ is odd.
>
>   However, regardless of what the reason might be for RGB mapping to
> physically linear brightness on his monitor (while in most other cases
> the mapping is exponential), this is still the beauty of gamma
> correction (when applied properly): He can still set up his assumed and
> display gammas in POV-Ray so that rgb 0.5 looks on his display as
> physically 50% bright, design his scene that way, distribute the
> resulting PNG, and the what-is-supposed-to-be-50%-gray will look like
> that on someone else's display as well.
>
>   Without any gamma information what looks on his monitor as perfectly
> 50% gray will look significantly darker on someone else's monitor. The
> whole purpose for gamma information is to make the image look the same
> in all systems.
>
>   So while it might be odd (regardless of what the reason for it might
> be), it shouldn't be a problem.

There seems some confusion here about two ways of approaching gamma. Gamma in
the sense of how a monitor must be calibrated to match absolute physical
brightness levels on a display, paper, inks etc, and gamma in the sense of the
*relative* distribution of tonal values from highlights to midtones to shade to
shadow to black, what Photoshop calls "curves" or what POV has called
"assumed_gamma". It seems, too, from my preliminary tests, that POV is
sacrificing one for the other. It's been almost impossible so far, working
within the defaults of 3.7, to get pictures that don't seem washed out and flat,
as the gradient example suggests they would since the fine gradations in the
darker range seem lost. I'm curious to see how the sample scenes will turn out.
It is perfectly possible and even essential in image processing to adjust the
*relative* distribution of tones. usually to add depth to a picture, and still
set gamma correction according to the color profile so that it can be seen
equally well in another medium or display. It doesn't seem to be possible though
to have gamma correction without color profiles.

Post a reply to this message