CIECAM02-a WORKFLOW

Edit--Preliminary results of the "Perfect Monitor Chase"
Compared to flat screen monitors my Sun workstation monitor is too dark by about .5 ev. So what I thought was perfect exposure and color--what this blog post is about--will--most likely--look overexposed and washed out on your monitor.

Why 'most likely'?  Because flat screens are all over the place. My local library has a wide assortment. On Internet 7 this blog post looked pretty good. On Internet 2 its colors were so garish I cringed. So I leave this edit  by saying the examples I posted might look terrible but I still stand by the workflow instructions.

Plus, if anyone knows an accurate way to set up the gain and bias of the three RGB guns in this beast leave a comment to point me in the right direction.


Now that we have an up to date--as of January 20 2013--64 bit win7 build on line, I was able to sit down and work out a workflow for tone mapping  using  CIECAM02.  Tone mapping is important to me since I take a lot of underexposed images. Sometimes deliberately. With low light  and moving subjects like Charlotte and her friends it's 1/125 second shutter priority and let the ISO fall where it may. I can deal with camera noise but blur is forever.

And sometimes...you know how it is keeping track of all those pesky camera settings. In the photo of Charlotte and her birthday check I faced harsh outdoor lighting coming over Charlotte's shoulder but I can't remember where  the  -.67 EV setting came from.

The original NEF converted with the RT neutral profile.

With a +1.41 EV exposure correction. Any greater exposure correction clipped the R channel at my test point on Charlotte's right leg up by her shorts. With this much EV the writing on the check and much of her right leg is washed out.  With deep shadows on her face this is a candidate for tone mapping.

Using the default setting of RT's original LAB based tone mapping . Tone mapping cleaned up the overexposed check and legs but didn't do much with the shadows on her face.

 The tool tip on tone mapping says to use CIECAM02:
1-check the box to enable the CIECAM02 tool- obvious
2 set the algorithm to Brightness + Colorfulness-less obvious
3-check the box to enable tone mapping using CIECAM02--again obvious

What to do next awaits a long manual entry. Jacques, the author of this masterpiece of post processing has written a long and programmer oriented paper on this tool. In French. His email English is far better than what's left of my high school French. So I depend on Google Translates. I won't say Google Translates totally mangles his French but I found it tough going pulling out the details. So if I've missed an obvious workflow...c'est la vie.

Step one of my version is to use the Brightness curves to brighten the image. I used tone mapping's default strength of  .25 but cut its Edge Stopping slider back to .55. It sharpens detail like Charlotte's hair and the default 1.4 setting seemed overaggressive.

I used the two control cage Brightness curves  for this post. After I'd made the screen clips I decided  it really did not matter what curve type or types I used. This step makes the image as bright as you can without clipping any of the color channels. So also check the 'show CIECAM02 histogram in curve' box. By design or bug it shows up only in the curve1 box. Adjust the curves so the histogram touches the right edge just short of clipping. A clipped channel will also develop a peak topped by a colored box in the main histogram box.

The image you get from curves alone is not likely to be pretty. This one is a bit washed out.

 So now you use the sliders to pretty it up.

Between Brightness, Colorfulness and Contrast you can fine tune to anything you want. For aesthetic reasons I set the yard behind Charlotte's head to almost total black. That made the shadows on her face a little darker than I would have liked. But all in all RT with CIECAM02 turned a disaster of a snap into something worth blogging about.

And there is more. CIECAM02 now works with Sharpening. Microcontrast, Defringe and Contrast by Detail.

With no sharpening

And with unsharp on and set a bit heavy. Don't know what they did but unsharp mask now seems to be less prone to halos and other artifacts

You can find this build of RT at-
http://www.visualbakery.com/RawTherapee/Downloads.aspx

Unfortunately as of  January 22 there is no manual entry for CIECAM02. But it will appear shortly. The online manual is at-
https://docs.google.com/document/d/1DHLb_6xNQsEInxiuU8pz1-sWNinnj09bpBUA4_Vl8w8/edit

For Jacques's paper on CIECAM02-
http://jacques.desmis.perso.neuf.fr/RT/ciecamRT3.html