15 April 2009

The Sunny Side of the Swan

The short version is that I didn't know ufraw could do that.
The slightly longer version is that the sunny side of the swan and the shaded side of the swan are by default much further apart in brightness than you're seeing here, but I have learned a new trick and can now rescue myself from situations where the sunny side feather detail is only visible at -1 stop without resorting to single-exposure HDR techniques.
So with any luck you can see the feather detail on both sides and my current sense of accomplishment is justified.


pir-anha said...

definitely justified. that's lovely feather detail.

so what's the trick?

Graydon said...

ufraw supports a brightness curve and a saturation curve, among many other things.

I had figured out that moving the top of the saturation curve moves the white point. (There's an auto button for setting the black point, which is very very helpful when taking pictures with fixed aperture lenses on bright sunny days!)

What I figured out, fighting with the RAW file with that swan in it, is that the curve doesn't have to increase (linear or S or wiggly shapes, variously) from left to right (which I think is roughly equivalent to brightness) between the left(black point) and right (white point); it can be an inverted parabola. So on the 4x4 grid provided in the curve tool for mental reference/mapping to the histogram, I can auto-set the black point (around 1,0), drag the white point down to (4,3) or so, and have the peak of the parabola around (but not quite at in the y direction) (3,4).

Voila! the top quarter of the image (the sunny side of the swan) gets desaturated enough to recover the feather detail without plunging the shaded side of the swan into dimness and mush.

This is probably well known and obvious but I'm still pleased that I figured it out. :)

And if that didn't make sense, let me know, and I'll do a post with screen captures.