Dan Margulis Applied Color Theory

Wide Working Space and Monitor Gamut

 
Wide working space and monitor gamut...
    Posted by: "raphael_bustin"
    Date: Sun Jul 23, 2006 7:01 am (PDT)

Forgive this dumb question please.  It's been bugging me.  It's surely been asked before.

What happens when (and this is the usual case) the working space gamut is larger than the monitor gamut?  ISTM, the working gamut has to be mapped down (ie., compressed) to fit into the monitor gamut, or else there will be extreme compression or clipping of tones outside the monitor gamut.

And if that's the case, it seems to me that choosing a large working space has an additional risk/burden -- if tones are compressed to fit into the monitor gamut, then that reduces the apparent contrast on the monitor -- ie., ability to perceive subtle gradations of tone.

In other words: we have "soft proofing" to show the gamut limits going from working space to output space.  So what's the equivalent for the monitor?  What really goes on?

rafe bustin
www.terrapinphoto.com
___________________________________________________________________________

Re: Wide working space and monitor gamut...
    Posted by: "john castronovo"
    Date: Sun Jul 23, 2006 7:21 am (PDT)

One workaround is the Photoshop option (from within the advanced color settings menu) to reduce the monitor saturation by 20 percent. I don't know how they came up with 20 percent. The effect isn't meant to represent a correct display of what is in the file, which we already know isn't possible on a device with limited abilities, but it can show you the differences between saturated colors which may appear to be the same otherwise. I think newer profile methodology may be able to handle this better in the future, but for now the options appear to be compress, clip or buy a better monitor.

john castronovo
___________________________________________________________________________

Re: Wide working space and monitor gamut...
    Posted by: Marco Ugolini
    Date: Sun Jul 23, 2006 10:40 am (PDT)

In a message dated 7/23/06 6:23 AM, raphael_bustin wrote:

Forgive this dumb question please.  It's been bugging me.  It's surely
been asked before.

What happens when (and this is the usual case) the working space gamut
is larger than the monitor gamut?  ISTM, the working gamut has to be
mapped down (ie., compressed) to fit into the monitor gamut, or else
there will be extreme compression or clipping of tones outside the
monitor gamut.

Hi Raphael.

It's a legitimate question. Nothing dumb about it.

As John Castronovo noted in his reply, you can use the "Desaturate Monitor Colors By:" option in Photoshop's Color Settings dialog box to compress the color gamut on-screen (which does not compress the gamut in the file: only the way it is perceived on the monitor). The percentage by which you can desaturate is user-defined; while the default is 20%, you can try higher values too, and see what that does for the way your image shows up on your display.

I am not sure how this "Desaturate Monitor..." mechanism works, meaning whether it works in a manner similar to a perceptual rendering intent, or uses some other technique to bring into visible range color differences that live outside the color gamut replicable on your display. To be honest, I am not even sure that it brings out-of-range colors into some sort of visible range, or instead just desaturates the colors you *already* see, without adding any.

Of course, the "Desaturate Monitor..." control will change the hues slightly too, besides the saturation, whether it works like the perceptual rendering intent or not. But, hopefully (and I am not sure of that either), you will be able to perceive tonal transitions that would otherwise be clipped out of visible range.

On the other hand, it's a fact that many displays, at least at this point in time, have a gamut that simply doesn't cover that of many inkjet devices, specially as these become better with the inevitable technological advances. For example, my Apple Cinema Display's gamut comes short of my Epson 2200's cyan-greens, red-purples, and some yellow-oranges. Even the Eizo CG210 shows similar shortcomings, according to a profile I received from an owner, and the CG220 as well comes remarkably short in the green-cyan range.

So, to make a long story short, the fact is that a knowledgeable user has to make a decision:
- Either squeeze the colors in one's images to a color space that has a gamut similar to that of one's display (thus precluding at the outset any chances to use the wider gamut of the inkjet printer, given that one is using a space smaller than what the output is capable of);
- Or live with the limitations of the display's gamut, use a working space that encompasses the *printer's* gamut, and tweak details that may need to be tweaked but live outside the range visible on the display (a bit of hit-and-miss, but for the sake of a livelier, brighter print).

And if that's the case, it seems to me that choosing a large working
space has an additional risk/burden -- if tones are compressed to fit
into the monitor gamut, then that reduces the apparent contrast on the
monitor -- ie., ability to perceive subtle gradations of tone.

The display's video card usually works in 8 bits, whereas a large-gamut working space is best represented and dealt with in 16 bits (yeah yeah, I hear the usual grumbles already from the traffic cops...so please let's agree to disagree).

Compressing the colors of a large-gamut space is not the best of possible compromises, but until large-gamut displays become a reality affordable to most users, we must work around the hurdles.

In other words: we have "soft proofing" to show the gamut limits going
from working space to output space.

I would say that we have soft proofing to show an approximation (at times more accurate, at times less, depending on the type of image) of the final results. In my opinion, the paramount function of soft proofing is that it saves us *some* time and money (in consumables) that would otherwise be spent producing the print, evaluating it, tweaking the file, then making another print, and so on. It's a sort of educated, cut-the-voodoo divining rod, if you pardon my twisted analogy, but no matter how scientifically-based, there is still an amount of conjecture and guesswork that we as users simply have to learn to handle and live with.

So what's the equivalent for the monitor?  What really goes on?

You mean, how do you tell which colors in the file are being clipped by the display? At least two possible ways to approach that task, in my view:  1) In Photoshop, go to View -> Proof Setup, choose "Custom"; in "Device to Simulate" choose your monitor's profile, and make sure that "Preserve RGB Numbers", "Black Point Compensation" and "Simulate Black Ink" are *unchecked*; select the Relative Colorimetric rendering intent (since I believe that it's the one that the monitor profile uses to send values to the monitor); click "OK"; in the View menu, select Gamut Warning. You will now see the clipped colors covered by a flat warning color (I think Cyan is the default).

A note of caution: the Gamut Warning tool has been known not to work properly, so take the results with a big grain of coarse salt -- you know, the kind one uses to cook pasta, at least we Italians... : o).

Also, it's a bit of a "dumb" tool, in that it doesn't respond at all to changes following the activation of "Desaturate Monitor..." (someone correct me here if I'm mistaken).

2) Another way (far more precise, in my opinion) to compare your file's gamut to that of your monitor's requires that you have a copy of Chromix's ColorThink (it's worth every penny of what you pay for it! -- yes, a shameless endorsement, though I have no financial interest in it, just great admiration and respect for Steve Upton and the folks at Chromix).

Go to the Grapher tool (Graph -> Open 3D Graph...), select your monitor display's profile, and you will see a 3D shape of it in L*a*b* space (D50). Now, make a copy of your image file, save it in TIFF format, *with your working space profile embedded* (that's an essential detail!). Now, select that file for 3D display in the ColorThink Grapher (or drag-and-drop it into the Grapher's main window).

The application will take a few seconds to "crunch the numbers", then you will see a cloud of colored points in your Grapher, corresponding to the colors in the file as interpreted via the working space profile embedded in it. If you dial back the opacity of the monitor display's 3D graph, you will see the color dots that live *inside* the monitor display's gamut, whereas the ones that exceed its gamut will be visible *outside* of it.

As I said, this second way (using ColorThink) is very likely to be far more accurate, as well as infinitely more detailed, than the way that uses Photoshop's somewhat blunt Gamut Warning tool.

Best of luck, and regards.

--------------
Marco Ugolini
Mill Valley, CA
___________________________________________________________________________

Re: Wide working space and monitor gamut...
    Posted by: "Chris Murphy”
    Date: Sun Jul 23, 2006 10:42 am (PDT)

The amount of space most images take up in a wide-gamut RGB space is comparatively small. So this isn't a factor most of the time. But when there are colors in an editing space that don't exist on the display, they are "clipped" rather than compressed. So it is possible to see artificial posterization, or lack of detail, where there really is detail and no posterization.

There are a number of solutions to the problem. The first is Desaturate Colors By option in Color Settings which provides for a crude level of compression from edit space to display space. Another is to get a wider gamut display. But the most practical and useful solution is true perceptual rendering to the display, which is something the v4 ICC specification allows for, in particular with a mostly defined perceptual reference medium. So now we just need to see vendors taking advantage of this; and also we need to get v4 wide-gamut RGB profiles (which will improve their rendering to print significantly as well).


Chris Murphy
Color Remedies (TM)
www.colorremedies.com/realworldcolor
___________________________________________________________________________

Re: Wide working space and monitor gamut...
    Posted by: Dan Margulis
    Date: Mon Jul 24, 2006 2:04 am (PDT)

The theoretical gamut of the monitor can be deceptive because it is extremely easy to create settings where the monitor won't show detail in saturated colors, even if the working space is sRGB where the monitor ought to have no problem.

Fortunately, there's an easy solution. Whenever you see a blob in a highly saturated color, look at the individual channels. If they have detail, then the file has detail regardless of what the screen shows. Whether the detail will hold when it hits the output device is another story.

Dan Margulis
___________________________________________________________________________

Re: Wide working space and monitor gamut...
    Posted by: "Raphael Bustin"
    Date: Mon Jul 24, 2006 5:51 am (PDT)

I'm just as concerned about posterization and other "bad profile" effects that you might see (for example) in deep blue skies.

But mainly the context of my question had to do with the alleged benefits of very large working spaces like PhotoPro RGB.  My point is -- if the working space is much larger than the monitor's space, aren't you working "blind"?

So far, I haven't worked up the courage to try anything larger than AdobeRGB as a working space.

The good news is that soft proofing can actually reveal some of these problems (as far as output profiles are concerned.)  But I'd never really understood what happens under the hood at the interface between working space and monitor.  The texts I've seen don't say much about this.

Rafe Bustin
www.terrapinphoto.com
___________________________________________________________________________

Re: Wide working space and monitor gamut...
    Posted by: Stephen Marsh
    Date: Mon Jul 24, 2006 7:04 am (PDT)

Rafe Bustin wrote:

I'm just as concerned about posterization and other
"bad profile" effects that you might see (for example)
in deep blue skies.

Rafe, as per the reply from Chris, are your skies (or other areas of interest) out of your editing space gamut and or monitor gamut? Common images do not always contain wider gamut colours.

It may also be wise to test print when you see such artifacts (or moire) to help verify if it is monitor related.

But mainly the context of my question had to do with
the alleged benefits of very large working spaces like
PhotoPro RGB.  My point is -- if the working space is
much larger than the monitor's space, aren't you working
"blind"?

Yes, but we live in a world where users can plug their camera directly into the printer - granted these users are not the same ones who edit their photos or who would use wide gamut RGB.

It is a two part issue, in my mind. There is the theory, and practical application.

i) Capture/Raw Conversion/Archival
ii) Editing/Conversion to other spaces
 
If a user does not with to clip colour values, a wider gamut may be required on capture/processing. There does not appear to be much technical argument here. There is debate on whether one really needs to capture such colours and that will depend on the workflow and user, as with high bits and other hot topics.

Problems arise on the second point. Many get so caught up on the first point, that they toss it all out when performing the second. Perhaps this is because all one hears about is the theory, rather than putting it to good use.

As you note, if editing in the wide gamut space, it can be blind or take experience learning the output process and the previewing tricks etc. If the output device can make meaningful use of the gamut then you  may only really see the final result in hard copy form and then decide if things are worth it or not.

High bits are generally advised in such a wide gamut space.

Converting from the wide gamut RGB space to a smaller matrix based RGB editing space will clip the out of gamut colours as there is no perceptual transform.

So far, I haven't worked up the courage to try anything
larger than AdobeRGB as a working space.

Would your output be capable of showing any difference? If it could, would the images have this gamut, and if they did, would it be of any real value?

and monitor.  The texts I've seen don't say much about this.

I am not the one to answer this, but I would presume RelCol intent. I am not sure if the monitor profile was table based and not matrix would make any difference in this case.

Regards,

Stephen Marsh.
___________________________________________________________________________

 Re: Wide working space and monitor gamut...
    Posted by: Andrew Rodney thedigitaldog
    Date: Mon Jul 24, 2006 8:15 am (PDT)

On 7/24/06 5:36 AM, "Raphael Bustin"  wrote:

But mainly the context of my question had to do with
the alleged benefits of very large working spaces like
PhotoPro RGB.  My point is -- if the working space is
much larger than the monitor's space, aren't you working
"blind"?

Yes but that1s true for smaller color spaces too. The gamut of SWOP CMYK has colors that are NOT visible on an sRGB display.

So far, I haven't worked up the courage to try anything
larger than AdobeRGB as a working space.

Unless you have one of the few and very expensive Adobe RGB gamut displays, you1re still flying 3blind2 with that working space too. Of course, we have to figure out if the gamut of the scene falls within the color space we want to use for encoding first (something you do in a RAW converter).

99% of displays fall within an sRGB gamut. LOTS of images, scenes and color spaces have colors that fall outside that gamut. So we1ve all been flying blind for years assuming the image contains colors that fall outside sRGB.

Andrew Rodney
___________________________________________________________________________

Wide working space and monitor gamut...
    Posted by: "colorman042000"
    Date: Wed Jul 26, 2006 4:29 pm (PDT)

Chris Murphy wrote:

Chris, if my monitor is displaying an image with a *lot* of out-of-gamut colors will it display more posterization and lack of details than if the image had had those out-of-gamut colors compressed by conversion ( w/perceptual rendering) into a smaller space like sRGB? Will the perceptual method display more or less or same posterization and loss of details than other form of compressions.

Andre Dumas