Photoshop's Gamut Warning

[bjanes]

If you can't see it on your monitor, what can you see it on? The print. Print it out. Use the soft proof to help you pick a rendering intent and decide if you wish to edit (a copy) of the image further  based on the soft proof and original image in your working space (which I would submit is out of display gamut too).
[a href=\"index.php?act=findpost&pid=163382\"][{POST_SNAPBACK}][/a]

Yes, but does that not at least partially obviate the purpose of soft-proofing in the first place? And do you wish to make another print for each rendering intent and saturation edit whose effect you can't see on the monitor? Until our monitors match the gamut of the final output, there will be some problems with soft proofing.

If you find the overlay useful, great. Just keep in mind when it was invented and for what based on the current technology then consider what we have today. For most users, the feature could go away (and probably should but Adobe hates to ever remove anything despite nearly everyone agreeing its too complicated).
[a href=\"index.php?act=findpost&pid=163382\"][{POST_SNAPBACK}][/a]

The current PS overlay is not particularly useful, but one with a delta C such as I demonstrated would give some useful information, particularly if available real time and when used with rendering intents or saturation edits. Why soft proof it it still takes multiple print iterations? Better tools could cut down the number of test prints necessary to get the fullest gamut possible from the printer. Perhaps you can do it all with one test print, but others may not be so fortunate.

Bill

[digitaldog]

Yes, but does that not at least partially obviate the purpose of soft-proofing in the first place?
[a href=\"index.php?act=findpost&pid=163448\"][{POST_SNAPBACK}][/a]

Not unless you were under the impression that a soft proof perfectly matches the final which so far, no one should be saying.

[bjanes]

...and I'm POSITIVE he WAS using gradients...it's called "Bracket Proofing" where he uses a gradient on an adjustment layer where the adjustment goes from WAY too strong to  too weak. You choose the location in the gradient that gives you the optimal results and then fill the gradient with the level of tone to achieve that level of adjustment. He has a PDF available for free at Downloads look for the Bracket Proofing in the "Technique" section.
[a href=\"index.php?act=findpost&pid=163443\"][{POST_SNAPBACK}][/a]

Yes, but how is the bracket proof evaluated? With current soft proofing (where the result might not be visible on the screen) or with the useless gamut indicator?

Bill

[bjanes]

Not unless you were under the impression that a soft proof perfectly matches the final which so far, no one should be saying.
[a href=\"index.php?act=findpost&pid=163449\"][{POST_SNAPBACK}][/a]

From my original post, I was obviously aware that the soft proof is limited by the gamut of the screen, so I find your reply superfluous, except that it enables you to get in the last word as is your wont. If colors in the image are far out of gamut, I might want to make some adjustments before wasting time and paper.

Bill

[Schewe]

Yes, but how is the bracket proof evaluated? With current soft proofing (where the result might not be visible on the screen) or with the useless gamut indicator?
[a href=\"index.php?act=findpost&pid=163450\"][{POST_SNAPBACK}][/a]

Either soft proofing or print proof, or both...

Obviously, from the name "soft proof" nobody should be expecting a 100% accurate replication of the final print. DOH...and nobody I know expects that. What soft proofing DOES do is allow a gamut and dynamic range prediction of the print and, when used correctly, allows one to get a _LOT_ closer to what the final print will look like without ink hitting paper.

And, while it's not 100% accurate, I would say it's close to about 85-90% accurate prediction which is STILL more useful that looking at the image in its native color space on a display without soft proofing. Right? And soft proofing is more useful that merely looking at the gamut warning to see what's out of gamut.

So, while it ain't perfect, it's better than nothing and more efficient in time & costs to soft proof prior to printing than printing a ring-a-round and twiddling with images AFTER making a bunch of prints.

Do you actually soft proof? Do you print? Is it not useful to you? If not, I would suggest learning how to soft proof and determine for yourself its usefulness in your own work. Arguing about the limitations of display gamuts won't get a print done better/faster, will it?

[bjanes]

Either soft proofing or print proof, or both...

Obviously, from the name "soft proof" nobody should be expecting a 100% accurate replication of the final print. DOH...and nobody I know expects that. What soft proofing DOES do is allow a gamut and dynamic range prediction of the print and, when used correctly, allows one to get a _LOT_ closer to what the final print will look like without ink hitting paper.

And, while it's not 100% accurate, I would say it's close to about 85-90% accurate prediction which is STILL more useful that looking at the image in its native color space on a display without soft proofing. Right? And soft proofing is more useful that merely looking at the gamut warning to see what's out of gamut.

So, while it ain't perfect, it's better than nothing and more efficient in time & costs to soft proof prior to printing than printing a ring-a-round and twiddling with images AFTER making a bunch of prints.

Do you actually soft proof? Do you print? Is it not useful to you? If not, I would suggest learning how to soft proof and determine for yourself its usefulness in your own work. Arguing about the limitations of display gamuts won't get a print done better/faster, will it?
[a href=\"index.php?act=findpost&pid=163455\"][{POST_SNAPBACK}][/a]

Of course, I print and I also soft proof. Why do you always assume that anyone other than your and your gang are ignorant? If you had taken the trouble to view my post, you would have seen a soft proofing screen. As it was, you made unjustified implications.

BTW, I purchased and have viewed carefully the From Camera to Print video by you and Michael, so I can't be completely ignorant, or else you are a poor teacher . IMHO, gamut mapping software is a valuable addition to soft proofing, at least if you know how to interpret the output. Do you? I am not arguing about the limitations of display gamuts, but am simply pointing out a different way to view by what extent a color may be out of gamut, something that the OP or one of the follow on posters inquired about.

Bill

[Schewe]

Of course, I print and I also soft proof.
[a href=\"index.php?act=findpost&pid=163461\"][{POST_SNAPBACK}][/a]

From your posts and the fact you were bemoaning the inability of displays to approach the gamut of prints, I couldn't tell. Sounded like you were arguing the technicalities without the benefit of experience...

[bjanes]

Post withdrawn by poster.

[bjanes]

From your posts and the fact you were bemoaning the inability of displays to approach the gamut of prints, I couldn't tell. Sounded like you were arguing the technicalities without the benefit of experience...
[a href=\"index.php?act=findpost&pid=163481\"][{POST_SNAPBACK}][/a]

Really? That is your uncharitable and biased take. Typical bull in a china closet, lacking in any subtlety or nuance. Sometimes, it is better to actually read and try to understand what others are saying, rather than jumping to unjustified conclusions from your narrow point of view.

[David Sutton]

Really? That is your uncharitable and biased take. Typical bull in a china closet, lacking in any subtlety or nuance. Sometimes, it is better to actually read and try to understand what others are saying, rather than jumping to unjustified conclusions from your narrow point of view.
[a href=\"index.php?act=findpost&pid=163491\"][{POST_SNAPBACK}][/a]

Oh, for heavens sake. Go get some sleep. David

[shewhorn]

He has a PDF available for free at Downloads look for the Bracket Proofing in the "Technique" section.
[a href=\"index.php?act=findpost&pid=163443\"][{POST_SNAPBACK}][/a]

       Well I guess that settles that then.     If only I could upgrade MY OWN memory.

Cheers, Joe

[bjanes]

I wanted to create a very basic test file. I created a 2400x3000 pixel file in the sRGB color space at 8 bits (I actually started out with ProPhoto but after seeing the results I was getting I dumbed it down to sRGB). I then created 14 divisions, each with a gradient in it (the divisions were running vertically on the page). I created two sets of gradients for each color those being X to black (where X is the color) and X to white. The values I used are as follows (numbers are in the format of Red, Green, Blue)

255,0,0
0,255,0
0,0,255
255,0,255
255,255,0
0,255,255

So with two sets of that (one going to black, the other going to white) I had 12 gradient strips. I also created a white to black gradient and finally one strip that was grey (128, 128, 128).

[{POST_SNAPBACK}][/a]

Joe,

I made my own testimage and got somewhat different results with the gamut warning as shown below for perceptual rendering with the Canon Matte Profile which I downloaded from the Canon USA site. The map is not useful because it does not indicate if the color error is large or small:



The standard way to express a color error is with ΔE (which measures both luminance and chrominance in the 3 dimensional relatively perceptually uniform L*a*b space. Some workers prefer ΔC which ignores luminance and takes only color into account). There are several varieties of these measurements, but I am using *CMC. One loads the image in to Gamutvision and a soft proof is given on the upper right. ΔE or ΔC magnitudes are shown by a pseudo color display with the key on the right. Note that the ΔE on the top has a gamut problem because of with low luminance in the shadows. This does not show on the ΔC plot or on the Photoshop Gamut display, but it should show up with softproofing with black point compensation turned on.





Correlation between the Gamutvison result and PS is not that good.

The whole test is not that useful because the sRGB color space contains colors that can't be printed and there are many colors that the printer can handle, but are out of the sRGB gamut.

The 3D plot below demonstrates the sRGB (the wireframe) contains high luminance greens and a few yellows that can't be printed, and that some yellows and oranges at high luminance are in gamut for sRGB and out of the printer gamut. By contrast, at lower luminance there are greens, blues, and cyans that are out of sRGB but well within the printer gamut. There is no rendering in these views, which only show the two gamuts. The wire frame is sRGB and the solid color represents the printer space.



Here is what happens with perceptual rendering. The wire frame is the ouput (printer) space and the vectors show how the rendering affects low luminance values (the matte paper does not have deep blacks) and high luminance saturated primaries. There are a great many out of gamut colors, which are clipped by the relative colorimetric rendering. Perceptual rendering can not handle such large out of gamut conditions, and the results are similar.



Research at Kodak and elsewhere has shown a gamut of colors which represent real work surface colors which actually appear in photographs. Excluded are highly saturated self-luminous or florescent objects, neon lights, and computer generated display. Here is a saturation plot demonstrating the real world surface colors along with the gamuts of ProPhotoRGB, aRGB, and sRGB drawn up by Prof. Dr Gernot Hoffman in Germany. The dotted line shows the gamut of the surface colors, and the gamuts of three common spaces are also shown. ProPhotoRGB covers the whole surface gamut, but the smaller spaces do not.



These are 2 dimensional plots showing mid luminances that are most important. At high and low luminances, even sRGB has some colors which can not be printed, but these are not real world colors and are not important. This is shown in this Kodak paper on Color.org. Download and view at your convenience.

[a href=\"http://www.colour.org/tc8-05/MetricsUpdateNov01.pdf]http://www.colour.org/tc8-05/MetricsUpdateNov01.pdf[/url]

Note how the Canon printer colors include important greens at mid luminance that are well outside the sRGB gamut shown by most monitors and not fully in gamut for aRGB. This eplains why ProPhotoRGB is the preferred working color space for digigal photography with modern inkjet printers. Glossy papers have an even better gamut. Softproofing might be problematic with these colors. Herr Schewe can indicate how he handles them.

[standard_observer]

[a href=\"index.php?act=findpost&pid=164118\"][{POST_SNAPBACK}][/a]

bjanes,

Your contributions are very much appreciated.

DPL

--

[bjanes]

bjanes,

Your contributions are very much appreciated.

DPL

--
[a href=\"index.php?act=findpost&pid=164175\"][{POST_SNAPBACK}][/a]

Thanks for the kind words, DPL. I think that my post did clear up some of the OPs concerns, but neither he nor anyone else has responded. Disappointing.  

Bill

[Chris_T]

There is a read out mode option in the Info Palette for Proof Color. The first read out can be selected for Actual Color, and the second read out for Proof Color (rgb in italic). At any given sample point, the two read out values are different, regardless of whether I'm in Soft Proof or not.

Questions:

- What is the definition of Proof Color?

- How can we make use of this information?

[Schewe]

- What is the definition of Proof Color?

- How can we make use of this information?
[a href=\"index.php?act=findpost&pid=164549\"][{POST_SNAPBACK}][/a]

The info palette can read out what the actual color WILL be based on the proof set up. However, it's not particularly useful when printing RGB to drivers that want RGB unless certain RGB readouts may mean something to you. It's primarily designed to be used with RGB images and CMYK output (or CMYK files being cross-rendered in a different CMYK space) where certain CMYK numbers may be needed and useful.

[bjanes]

There is a read out mode option in the Info Palette for Proof Color. The first read out can be selected for Actual Color, and the second read out for Proof Color (rgb in italic). At any given sample point, the two read out values are different, regardless of whether I'm in Soft Proof or not.

Questions:

- What is the definition of Proof Color?

- How can we make use of this information?
[a href=\"index.php?act=findpost&pid=164549\"][{POST_SNAPBACK}][/a]

The first readout is that of the color in the working color space. The second readout is that of the data that will be sent to the printer in terms of the printer color space. This can be verified by converting the image to the printer color space, giving the actual colors that will be sent to the printer. In this case, the two readouts are exactly the same.

Since printers are highly non-linear, the color that will be produced in the print (output color) is not easily predicted from the input RGB values. I don't really know of what use such data might be. What would be of interest would be the output in terms of a device independent space, such as L*a*b. You could compare with the original data, also in L*a*b, and determine the Delta E.

[shewhorn]

bjanes...

Thanks for putting the time into your excellent response!

Cheers, Joe