Mike Chaney's Tech Corner

Hi Terry-

To continue.  I agree with you about gamut.  However, the starting color space is relevant.  Your Perceptual vs. Relative Colorimetric example is made (I am guessing) using Adobe RGB.  Using the premise that you are starting with RAW files.

Instead of re-inventing the wheel, I'll just quote Uwe at Outback Photo:
He also states starting with 16-bit ProPhoto space has a wider initial gamut and is more likened to paper spaces.

"Here we have a situation that we can utilize all the paper colors from ProPhoto RGB. Colors that are outside the paper gamut need to be mapped using a "perceptual" rendering intent. If the colors of your photo mostly are in the the range of the paper you can also use the "relative colorimetric" rendering intent."

The full article is http://www.outbackphoto.com/color_management/cm_06/essay.html (http://www.outbackphoto.com/color_management/cm_06/essay.html).

Also, check this:  http://www.luminous-landscape.com/tutorials/prophoto-rgb.shtml (http://www.luminous-landscape.com/tutorials/prophoto-rgb.shtml).  The same issue.

The problem (?!?!) begins after the camera.  The newer stuff (Nikon, Canon, etc.) have a larger space than Adobe RGB.  (The only camera options for JPGs are Adobe or sRGB though, since they are 8-bit to begin with.)

Since I drag RAW files into ACR, then work them in PS, I watch the out-of-gamut warning on colors that are critical.

I would agree with that; compressed OoG colours would seem better than clipped ones.

Yes, but  .... if the colours are mostly within gamut, why bother with RC intent, Perceptual rendering will not change them.
That's my take on it.
Terry.

Perceptual actually does change the colors in and out of gamut.  Perceptual extent "compresses" the gamut of the image so that more of it will fit.  This has the effect of changing both in gamut and out of gamut colors and generally desaturates a bit.  Relative Colorimetric doesn't do that.  It renders all in-gamut colors correctly while just clipping the out of gamut colors to the closest in gamut color.  This fact (about the non-relative "squeeze" effect of perceptual intent) is something most people don't realize or talk about but it is an important concept to understand.

Mike

and
Thanks Mike, just shows some of us still have a way to go on understanding colour management.
I'll be seeing the guy who supplied those diagrams in September (http://ddisoftware.com/tech/qimage/red-prints-orange/15/)
 I'll speak to him about it - in the nicest possible way of course.  ;)
Terry.

I’m not going to pretend I understand anything about gamut or Perceptual, and Relative Colorimetric, but I’m beginning to learn something I really know nothing about, or at least learn a tad bit about the subject of gamut, Perceptual, and Relative Colorimetric.  In reading your post, I wonder why Perceptual is the (Default) in Qimage, and not Relative Colorimetric?  It almost sounds like it should be the other way around.