Good. The workflow at the institution is absolutely standard. Most of the users of the photography facilities have never seen a print (from an twelve inks inkjet printer) match a display, and I do that with low effort but with the necessary tools. Yes, I know the tools and the glossary. My only question/proposal is why the first answer to the question wasn't "you must buy a colorimeter if you want to know what you are seeing". And yes, I take spectral measurements of the light sources involved in the reproduction of art to build profiles. And many more. Salud 2014-06-06 23:47 GMT+01:00 Ben Goren <ben@trumpetpower.com>:
On Jun 6, 2014, at 2:26 PM, José Ángel Bueno García <jbueno61@gmail.com> wrote:
Are you are saying that if I measure RGB values in PS of achromatic patches of a ColorChecker, the numbers are lying because I haven't profiled my MacBook Air display?. Or that I can't define a white value on a scene with different light sources?.
Both, actually.
The ColorChecker Passport is a superlative target for field use...but even its neutral patches aren't perfectly spectrally flat. Hell, even Spectralon isn't perfectly spectrally flat. That means that a properly white balanced image of a ColorChecker won't have patches with perfectly equal RGB values (even if there are circumstances where rounding reduces them to the same values). (Incidentally, that's also why using an eyedropper to set white balance is always an approximation...though you _can_ build a profile from an image of a ColorChecker and then query the profile to determine exact color balance.)
And then, after questions of the (lack of) neutrality in the neutral patches, you get into a really deep quagmire of questions over color balance and adaptation and illuminants and the like once you consider actual photographs of actual ColorCheckers (original or Passport or otherwise) in actual lighting when viewed on actual displays in actual environments with their actual ambient lighting.
What's the purpose you're trying to achieve? To exactly reproduce the light reflected off the original ColorChecker? Can't do that, not with an RGB display. To exactly create light that creates an equivalent tristimulus response? Within gamut and brightness limits, that may be possible, but, unless ambient light conditions (including monitor brightness and white point) exactly match the original, it's going to look weird. To make the onscreen photograph and the ColorChecker look the same when you hold the ColorChecker up to the screen? That's the main goal of color management, and it's possible to do that quite well. To do the same, but capture the "mood" of the original lighting (such as a landscape at sunset)? That's also possible, and is what most photographers strive for.
But a big part of the problem in this discussion, I think, is a misunderstanding that there's an actual color we can call "white" -- or, indeed, any actual color at all. Instead, what we have are light sources with their spectral power distributions; reflective surfaces that reflect different wavelengths in different proportions; and tristimulus values that are very much dependent on ambient observing conditions.
Fortunately, in practice, most of all that chaos resolves itself and things "just work," at least if you have a decent color-managed workflow. But you can very quickly drive yourself crazy chasing after ever better "matches" between things that, even in principle, can never be the same.
My general advice: profile everything to standards, and deviate from that only if you have good and specific reason to do so. If it's "close enough," leave it at that. If not, first question again if it's "close enough" -- and do so by reminding yourself of the environment in which the images will be viewed. _National Geographic,_ for example, has exemplary color...but their images look like dark mud in the typical living room in the evening when lit with an incandescent bulb, and they look equally bad in a doctor's office lit with cheap fluorescent bulbs. Their industry-best standards are wasted on the overwhelming majority of their readers. Unless your own audience is more critical than that of _National Geographic,_ obsessing over minor variations that you have to seek out is serious overkill. If the print looks "just fine" in the light through a window, you're already at the point of diminishing returns.
There's one notable exception, and that's when you know _exactly_ what the viewing environment is going to be, and you can measure the light with a spectrophotometer and build your print profile using that measurement. In other words, when making prints specifically for an exhibition in a particular gallery. In that case, you really can dial in things as much as your heart might desire and reasonably expect to see all that extra work come to fruition. But if it's your generic magazine or Web site or even art fair print...if your workflow is color managed and it looks "good enough," it _is_ good enough.
Cheers,
b&