On Jun 6, 2014, at 1:19 PM, John Robinson <jrswebhome@yahoo.com> wrote:
So if I measure white and it reads slight blue and red in the white my screen is lying?
Yes, actually. Not every display's red lights are the same color; indeed, there's lots of variation amongst them. And not every display's red lights are the same brightness at maximum relative to maximum blue brightness. And, to make matters even worse, a white object in direct sunlight will look much yellower than that same object in open shade. Now take that object indoors under an incandescent lamp and it'll look reddish. The point of color management is to both define a standard set of criteria by which to judge all of those variables and to provide consistent ways of transitioning between the two. At its heart are "device-independent" color spaces. Rather than identify colors by percentages of red, green, and blue, they use some sort of absolute measurement derived from both physics and careful studies of human physiology. Lab is perhaps the best known example, but XYZ is perhaps more important. The job of a color profile is to match up absolute device-independent colors with RGB values. In the case of a display profile, software displays various RGB combinations and reads the absolute color values measured by a colorimeter (a device you place against the screen). The software then is able to build a map between the two; that map is the ICC color profile. ICC-aware software (such as Photoshop) can then use the profile to know that, if it wants to display such-and-such an absolute color, it needs to tell the display to use these-and-that RGB values (which might only vaguely resemble the RGB values you read with the eyedropper). What I just wrote is, of course, a most superficial introduction to the topic. What you should take away from it is that you're complaining about color consistency, and there is a very successful industry standard method for solving that problem, and we're trying to help get you up to speed with that solution. Is the solution perfect? Ha! Far from it. Spend any time here and you'll find plenty of perfectly valid bitching about it at every level. Despite its imperfection, does it really work, and damned well? Yes, especially if you know what you're doing and understand the limitations of the tools available. Is it worth the hassle? Only you can decide. Cheers, b&