I started with this-a 1949 cover for the Saturday Evening Post painted by Norman Rockwell
To prevent one problem, jet black splotches that appear in highly over exposed areas when tonemapping and CIECAM02 are used together, Jacques, the developer who added CIECAN02 to our toolkit (applause, applause) made changes recently. He compressed the highlights for well exposed images taken with bright scene illumination, the highlighted value of 2000 standard candles per meter squared.
While that fixed one problem it added a second. For less well laminated scenes like indoor scenes with over exposed vistas viewed through windows, or like the low illumination I used to create this image, the histograms never go above 75%. White areas end up too grey.
The fix is blindingly obvious if you understand the intricate working of CIECAM02. You adjust the scene illumination slider to match the image. The histogram moves to the right; you stop when it starts to clip. But the rest of us who think a standard candle is the long thin one you push into a brass candle stick we tend to miss what's needed to be done.
Yesterday afternoon I discovered that undocumented fix. Not by some brilliance on my part. In a not-so-secret forum where RT developers hang out and talk RT programming talk, this was being discussed in issue 1827. So last night, me, a lowly user, dared to tell Jacques, the developer, that his technically correct tool tip needed to be rewritten. In the middle of our night and his morning-the time difference between Wisconsin and France-Jacques wrote back saying of course he would do it
Took less than 8 hours. So if any users of commercial programs such as Lightroom happen to stumble over this blog please put in a request for the perfect fix you've been waiting to use. Then let us know how many weeks/months/years/eons it takes to get an answer back. RAWTherapee might bring on slider shock and have its quirks, but its developers listen. And when it rocks, it Rocks!
EDIT. An Oops. Or semi Oops. What was less than blindingly obvious when I wrote this was that the real cause of the problem was a bug in the CIECAM02 calculations. With build 4.0.10.69 that has been fixed, the problem no longer exist and the histograms no longer hang up at 75 percent no matter where you set the illumination slider.
A two slider adjustment in scene illumination and exposure moved the red channel to the clipping point and lightened the image.
A CIECAM02 contrast boost darkened the edges and clipped the red channel
A wild LAB custom CC curve made big changes to the image and brought all the channels together. Not the usual way to set a white balance.
A final scene illumination of 145 cd/m2 sets the overall image whiteness.
Two curve and slider adjustments didn't change the image that much. Running the chromaticity to the max, however, gave me a wild effect I could use.
The final masterpiece ready to be printed and matted for the masterclass.
And because it is also neat the black and white version created by clicking BW tone mapping in the LAB section.
