Re: On the use of wide-gamut RGB working spaces [was: Photography editing spaces]
Re: On the use of wide-gamut RGB working spaces [was: Photography editing spaces]
- Subject: Re: On the use of wide-gamut RGB working spaces [was: Photography editing spaces]
- From: "Fleisher, Ken" <email@hidden>
- Date: Tue, 07 Oct 2008 08:36:52 -0400
- Thread-topic: On the use of wide-gamut RGB working spaces [was: Photography editing spaces]
Marco, thank you for your thoughtful response.
> I feel obligated to interject that it's not just "a bit" better -- it's *a
> whole lot* better. Roughly 45% better, since the approximate volume gamut of
> sRGB is 897,000 versus AdobeRGB's 1,299,000.
I agree. A few months ago I had one to test for two weeks and it looks
remarkable. I was just being cautious in my statement.
> Using ProPhoto RGB as the preferred working color space allows not so much
> for more *precision*, as for preserving colors in the file that
> bleeding-edge or future printing and/or display technologies might be able
> to reproduce.
But that's the crux of my argument--that the file either does not actually
have those colors in the first place, or if colors do lie in the extended
gamut, that you cannot see what they look like on your monitor or on a
print, and therefore if you engage in visual editing (important
qualification), how can you know what you are supposedly preserving for the
future?
Images from cameras that are profiled will have gamuts that are already
limited to the gamut of the target used which will mean colors are already
not in the extended gamut. Cameras that are not profiled will have
unpredictable colors in the extended gamut making visual editing necessary,
but impractical (I won't go as far as to say it's impossible, but I doubt
that most users will to anything besides base their decisions on what they
see on their monitor, or on a print).
> But it's easy to see how a fine artist could be interested in producing
> prints that are even more vibrant and vivid than the ones that are possible
> with the current means. So, it all depends on the user and the intended use.
I create digital captures of fine art and am certainly interested in
capturing the full color gamut of the original. However, as I stated, I
don't think the current tools allow me to accurately do this (not counting
ColorSage type tools which I have not had the chance to evaluate), so I use
Adobe RGB (1998) as the working color space. I still have my raw file, so if
in the future something magic happens with new printing technologies and I
need a more extended gamut, I can always re-export to ProPhoto RGB and
re-color correct, but I don't forsee this happening anytime soon.
>> By the process of making sound decisions, the colors in your file will
>> generally wind up being restricted to the color gamut of your monitor anyhow.
>
> That need not be necessarily so. Several inkjet printers already surpass the
> gamut of many middle-of-the-road monitor displays. <snip>
True. If you are basing your color corrections on a print that has a wider
gamut than your monitor, then by all means make sure you working color space
includes the printer gamut, but you still won't need anything like ProPhoto
RGB. This was what I meant to convey when I said:
>> Therefore, I believe that there is no point in using a working color space
>> that is much larger than your monitor¹s color gamut if you are making color
>> editing decisions of your digital camera captures based on the visual image
>> on your screen--or a hardcopy print of it for that matter.
I could have stated that much better. Thank you for helping clarify that
point.
> But please do not jump from such fortuitous coincidences (where the actual
> colors in an image fall short of the available gamut) to the conclusion that
> images in ProPhoto RGB really occupy a smaller gamut anyway, and so can be
> safely chopped to a smaller color space. That would be an overly broad
> generalization, and certainly not all users would be well-served by it.
Every user must always make their own decisions and I agree that it's
"possible" that a ProPhoto RGB workflow for digital cameras can work for
some people. However, I still believe that color will either be already
clipped (calibrated camera workflow) or will be unpredictable (uncalibrated
camera workflow) due to the inability to see what those colors
are--ESPECIALLY if you plan to engage in visual editing, which is an
important point.
> Each image is different. Even if a majority of the respondents should happen
> to agree with your conclusion, that would still not make it *universally*
> valid for *all* types of users.
Yes, but isn't it something like doing color edits (on an extended gamut
image) while viewing a grayscale image on screen? You can make your best
educated guesses, perhaps by viewing RGB or LAB values, and feel that you
have done a great job, and when you finally have a chance to view that image
in color you might actually have a great image. But it's highly likely that
there will be areas of color that make you fall off your seat thinking
"Whoa, that's pretty horrible!" The point being, you can't know what it is
you are editing if there is no way to see it. Just because you have edited
an image in ProPhoto RGB and some color may still lie outside of your visual
editing space (i.e. monitor or print) does not mean those colors are what
you intended them to be. Again, I'm not saying it's impossible, only that
it's very difficult and impractical.
> That is valid for, say, imaging oriented towards scientific applications,
> where dead-accurate source colorimetry may be needed for exact results. But,
> barring that, the rest of us is usually interested in plausible color (i.e.,
> accurate, certainly, though not to the demanding standards of exact
> science), but also *pleasing* color -- much more so than simply
> *dead-accurate* color. Of course, starting from color that is "accurate
> enough" does help a lot. But from that point onward one is free to exercise
> artistic license and stretch the results this and that way, as circumstances
> and artistic inclinations dictate.
Exactly. But how can you exercise artistic license on something that you
can't see? Either the color in the extended gamut is 1) accurate to begin
with, 2) inaccurate, but left as the camera recorded it and as interpreted
by your raw converter, or 3) moved around blindly within the extended gamut
and landing on some unknown color, which is not necessarily any better than
where it started.
> Common sense would seem to advise that a fine art photographer whose work
> contains strongly-saturated colors use a color space that is larger than
> that of any output device available today (inkjet or others), in order to
> preserve elements that may add esthetic value to the images if not
> immediately, then in a reasonably not too distant future -- or simply to
> keep that option open as a cautionary measure, if nothing else. After all,
> once clipped, saturated color and detail cannot be recovered.
Common sense would say that, but practical application does not allow for
it. This is why I'm excited about these new products beginning to emerge
because this will hopefully allow what you describe to actually occur.
> So, to say that the use of ProPhoto RGB makes sense only when accurate
> source colorimetry (scene-referred) can be captured without gamut
> restrictions, with perfect accuracy, and with no need for further editing,
> seems rather limiting to me, because it does not include the needs of fine
> art photography/imaging from Raw captures, which are *not* necessarily in
> agreement with that description.
It's in perfect agreement with that description. Anyone who has done fine
art photography will recognize that with commercial RGB digital cameras,
visual color editing is a necessity, even with a good calibration. The
problems associated with metamerism (camera vs. human) prevent a capture
from correctly imaging the artwork and there are, more often than not,
certain colors that just need to be fixed. This is impossible to do if you
can't see what colors are wrong and can't see what you are fixing them to,
so it's better, in my opinion, to fix them within the more limited gamut
space where you know what you are doing.
> I disagree, and hope that I have already stated my reasons clearly enough.
You were very clear and I thank you for your thoughts on the subject.
--
Kenneth N. Fleisher
Photographer
Imaging & Visual Services
National Gallery of Art
6th Street & Constitution Avenue, N.W.
Washington, DC 20565
Phone: (202) 712-7471
email@hidden
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Colorsync-users mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden