Re: [UFRaw-Devel] UFRaw 0.14 release plan

[Bruce G. <lis...@br...>]

On Thu, Aug 21, 2008 at 11:46:20AM -0500, Udi Fuchs wrote:
> Current contrast correction is useful when the luminosity distribution
> is Gaussian with the peak at the middle (value 128).

Or at least when the features that need contrasting are on either side
of the middle.  Two smaller peaks straddling the middlemay also benefit
if they are too close.

> Most of my photos
> seem to have a peak in the shadows and then random distribution all
> over.

Would you describe this as low-key -- most of the scene is in shadows?

> It somehow seems that my photos are much different from yours. I
> wonder if the difference is in the cameras or in our style of
> photography.

It is likely.  A lot of what I shoot is indoor available light, which
tends to have large amounts of middling tones and lower contrast, or
brighter lit outdoors.

FWIW I shoot with a Minolta 7D.

> One thing that might improve the current contrast correction would be
> to calculate the average luminosity and apply the S-curve around this
> value.

That's a possibility to try.  Where in the code would you prefer this
calculation get added?

-- 
Bruce Guenter <br...@un...>                http://untroubled.org/