Pages: 1 2 [3]   Go Down

Author Topic: Choices Between Printer Color Profiling Engines  (Read 12650 times)

Brad P

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 258
Re: Choices Between Printer Color Profiling Engines
« Reply #40 on: December 27, 2017, 11:30:17 pm »

. . . and saturation.
Logged

aaronchan

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 617
Re: Choices Between Printer Color Profiling Engines
« Reply #41 on: December 28, 2017, 12:30:17 am »

Thanks again Aaron.  I printed up the i1P samples and am going to let them dry overnight and shoot them tomorrow alongside the other samples, and will post the results.  For the moment though, Argyll remains the most impressive with the Z3200, followed by DropRGB.

Andrew, if you’re still around, I’m getting darker blue balls with i1P than I expected.  You might want to check that out in my pics tomorrow.  Not as dark/black as Copra, but much darker than those in your video, particularly in the darker ball.  I suspect it’s more likely a result of the inkset used by the Z, or possibly that we used i1P’s default settings, than an error in color management as I’ve been over the latter’s pipeline quite a few times.  But very interested in your take on that.  I’ll get a little better light on the final shot too because there is clear gradation in the i1P blue sample that’s not at all so clear in Copra’s and the darks in all of them simply need more light in the shots.

Hi Brad,

I have done a simliar test as you did before.
But mind was using different target patches and generate the icc in i1P
Like what Ethan said on the other post
One of my client use 1600 patches target to generate his icc profile which produce similiar blue as your, pretty dark in perceptual.
And I have made an icc profile with Ethan's method, 2371patches, the blue comes out a lot more "blue" compare to his.
So, I guess each profile generator has it's own algorithm which you need to find it's best target size to make an optimal icc profile.

aaron

Brad P

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 258
Re: Choices Between Printer Color Profiling Engines
« Reply #42 on: December 28, 2017, 09:26:10 pm »

So this brings up the more general issue of what colors should we be working with. I generally put these in several categories from the narrowest set, those that the printer can print, to increasingly larger gamuts.

These are stimulus colors. Colors are perceived based on context. That is what stimulus colors surround them and what adaptation has occurred.

1. Colors that are actually printable. By definition, these colors are within the printer's gamut.
2. Colors that can exist, theoretically at least, on a surface. These are the set of all colors with completely arbitrary spectral reflectances, and are bounded by the MacAdam limits. They are more limited than all possible visible colors because they are formed by subtraction of wavelengths. In a sense, these might be considered the limits of a "perfect" printer with an infinite set of inks.
3. Colors that exist and can be created by mixing all possible wavelengths of light. These require emissive sources. Think mixing light from a large number of lasers, each with a slightly different visible wavelength and adjustable magnitudes.
4. Finally, there are the imaginary colors. These are colors with xy coordinates that are outside the stimulus gamut. They don't exist because there are no combinations of wavelengths at any magnitude that produce them. These are the ones outside the classic CIEXY horseshoe gamut most often encountered in ProPhoto RGB where the Green and Blues are imaginary.

When evaluating a profile, my criteria varies. It's most critical for in printer gamut colors. There shall be no visible banding or other anomalies for these. Next, are Mac Adam colors that are outside the printer's gamut but physically realizable. These should be rendered smoothly and be pleasing though they will not be accurate. Then there are colors that are real but not possible on a surface. How they are rendered is not something I care about too much though my preference is that hue be maintained. These, can't ever be printed but can appear in photographs with things like red LED Christmas lights . I expect the luminance and saturation to drop but the hue to remain close.


Thanks for this explanation, Doug.  This, Andrew's chart, and actually seeing how the out of gamut colors are rendering in my test prints are challenging a perhaps tribal-like belief I've carried for a while that ProPhoto is the best image processing work space. 

I have been assuming in this post that if I regularly push ProPhoto RGB image data well out of printer and display gamut when processing, then try to pull it all back into print gamut, I'm likely to see image degradation in those out of print gamut colors.  And I've been assuming that the degradation can be seen particularly with problems printing out the charts in this post.  I've also been assuming that the smoothest OOG rendering profile (here to my eye Argyll perceptual) would do the best to minimize the negative affects of post processing manipulations.  If that's not correct, please correct me.

Thinking about your explanation and looking at this chart together with the Granger Rainbow/Bills Balls results, it appears that the largest problem areas in the charts I printed might be the imaginary colors in ProPhoto RGB.  It seems like the imaginary colors are the ones hardest to map to human vision and, for the reasons you state, excluding those in a working space might lead to better results.  Why would we want to include those imaginary colors in a processing work space? 
« Last Edit: December 28, 2017, 09:59:39 pm by Brad Paulson »
Logged

Doug Gray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2197
Re: Choices Between Printer Color Profiling Engines
« Reply #43 on: December 28, 2017, 11:14:14 pm »


Thanks for this explanation, Doug.  This, Andrew's chart, and actually seeing how the out of gamut colors are rendering in my test prints are challenging a perhaps tribal-like belief I've carried for a while that ProPhoto is the best image processing work space. 

I have been assuming in this post that if I regularly push ProPhoto RGB image data well out of printer and display gamut when processing, then try to pull it all back into print gamut, I'm likely to see image degradation in those out of print gamut colors.  And I've been assuming that the degradation can be seen particularly with problems printing out the charts in this post.  I've also been assuming that the smoothest OOG rendering profile (here to my eye Argyll perceptual) would do the best to minimize the negative affects of post processing manipulations.  If that's not correct, please correct me.

Thinking about your explanation and looking at this chart together with the Granger Rainbow/Bills Balls results, it appears that the largest problem areas in the charts I printed might be the imaginary colors in ProPhoto RGB.  It seems like the imaginary colors are the ones hardest to map to human vision and, for the reasons you state, excluding those in a working space might lead to better results.  Why would we want to include those imaginary colors in a processing work space?

The basic problem is that RGB colorspaces are additive while printer spaces are subtractive. As a practical matter if you want to have a gamut that can print all possible printable colors ProPhoto RGB is the only game in town. One just needs to be a bit careful with it.

I do think there is value in having profiles that produce reasonable looking results, even with extreme OOG colors. Fortunately, you don't need to make prints to do this. Soft proofing will show almost all the effects of extreme color mapping. It's quite accurate only failing when the printed color itself is outside the monitor's gamut.

Also, even sRGB has a lot of colors that can't be printed just like it clips a lot of colors that can be printed. There are about the same number of both. Adobe RGB has far more colors that can't be printed but still has many that can be printed but are outside the Adobe RGB gamut.

See this topic for details and examples:
http://forum.luminous-landscape.com/index.php?topic=107015.0
« Last Edit: December 28, 2017, 11:17:23 pm by Doug Gray »
Logged

Brad P

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 258
Re: Choices Between Printer Color Profiling Engines
« Reply #44 on: December 28, 2017, 11:35:46 pm »

Thanks, but let me state my question a little differently.  Seeing in Lab the imaginary blue color's luminosity, it was easier to understand why the blue balls turned out so dark and why the different color engines resolve them differently.  Looking at ProPhoto against CIE XY, I can see the imaginary colors in ProPhoto are the blues, and to a lesser extent the greens and even lesser still yellows.  I'm extrapolating (but didn't look in Lab) that one would have the same difficulty bringing those into any visible light colorspace. 

I get the fact that additive and subtractive light color spaces are different and why, but intuitively (and maybe my intuition is quite wrong) in a visible additive and subtractive workspace, it seems as you wrote an easier job to modify saturation, luminance then hue to bring the color closer to "fitting in" a printed gamut.  With the blue colors at least in Lab, I just don't get how one would do that other than playing wth luminosity, and seems like it would be pretty destructive to what the color really is imagined to be anyway.

So actually for the last hour I've been looking for the color space that most closely approximates human vision, which I'm surprised to see turning up in google searches still is 1931's CIE XY, is that right???  I was thinking of toying around with whatever color space I find and granger rainbows that I can make in PS to see how it prints with Argyll using that colorspace as the reference colorspace in the line item command.  Does that make sense?   
Logged

Doug Gray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2197
Re: Choices Between Printer Color Profiling Engines
« Reply #45 on: December 29, 2017, 12:25:40 am »

Thanks, but let me state my question a little differently.  Seeing in Lab the imaginary blue color's luminosity, it was easier to understand why the blue balls turned out so dark and why the different color engines resolve them differently.  Looking at ProPhoto against CIE XY, I can see the imaginary colors in ProPhoto are the blues, and to a lesser extent the greens and even lesser still yellows.  I'm extrapolating (but didn't look in Lab) that one would have the same difficulty bringing those into any visible light colorspace. 

I get the fact that additive and subtractive light color spaces are different and why, but intuitively (and maybe my intuition is quite wrong) in a visible additive and subtractive workspace, it seems as you wrote an easier job to modify saturation, luminance then hue to bring the color closer to "fitting in" a printed gamut.  With the blue colors at least in Lab, I just don't get how one would do that other than playing wth luminosity, and seems like it would be pretty destructive to what the color really is imagined to be anyway.

So actually for the last hour I've been looking for the color space that most closely approximates human vision, which I'm surprised to see turning up in google searches still is 1931's CIE XY, is that right???  I was thinking of toying around with whatever color space I find and granger rainbows that I can make in PS to see how it prints with Argyll using that colorspace as the reference colorspace in the line item command.  Does that make sense?

1931 CIEXY isn't the best and it is, to some degree a simplification of human vision but it is the basis for all the colorspaces in current use. Adobe RGB, ProPhoto RGB, CIELAB are all based on 1931 CIEXYZ. and there are precise conversions between each. The use of negative numbers allows any RGB colorspace to cover the entire CIEXY and there have been efforts to extend sRGB to do so. The Academy of Motion Picture Arts and Sciences has done a great deal of work extending both gamut and dynamic range with RGB spaces and has published a lot of open source material to facilitate people working in that biz.

Another thing to consider is that printer profiles don't know about RGB colorspaces. They work in a truncated version of CIELAB called ICCLAB where the a* and b* are limited to +-128. Lots of colors in ProPhoto RGB are clipped when converted to ICCLAB. Still, there are many colors in ICCLAB that also are imaginary. There are also colors that exist but exceed ICCLAB. It can be messy.

But at least all printers (that I have heard about) have gamuts that are inside ICCLAB.

I generally work in ProPhoto RGB unless I know the image gamut fits in a smaller space. But I soft proof and find that's the best way to work and avoid the more peculiar things that can occur with large spaces. But I still soft proof even with sRGB. Lots of colors visible on an sRGB capable monitor that won't print without some shifting.
Logged

Brad P

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 258
Re: Choices Between Printer Color Profiling Engines
« Reply #46 on: December 29, 2017, 03:18:30 am »

Well I’ve spent the night googling and reading to find an RGB color working space best approximating human vision (stimulus colors, no imaginary colors) and the best I could find appears to be CIE RGB.  That ICC profile is in my ColorSync directory so that’s readily accessible.  If there’s a significantly better one I’d appreciate any steers.

Interesting point about the printer workspace, ICCLAB. I imagine that the paper ICC profile would reflect the clipping that ICCLAB would do. If so, that part of the puzzle might be solved in software before the image data is sent to the printer.

I created two granger rainbows in PS both in ProPhoto and ColorSync and, just because it’s getting late, soft proofed them using the Argyll-generated, ProPhoto linked ICC profile.  I was somewhat stunned to see the differences, particularly in hue. The luminance banding did appear to be moderately reduced in the CIE RGB file.  Tomorrow I’ll do a more proper experiment and create an Argyll paper profile linked to CIE RGB and print out the two sets of rainbows for comparison, each printed using the appropriate printer profiles. 

I remain a bit concerned that CIE RGB may not be the right colorspace to be using to approximate stimulus colors. Mapping it out in ColorSync against ProPhoto RGB, although it looks like it may be clipping the correct dark blue colors in ProPhoto RGB (and the greens/yellows), the gamut volume rendered in 3D is roughly only about 2/3rds that of ProPhoto RGB, and from what I’ve been reading tonight and seeing in 2D renderings, I would have expected something more like 90+% overlap.  Hmm.
« Last Edit: December 29, 2017, 03:22:19 am by Brad Paulson »
Logged

Doug Gray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2197
Re: Choices Between Printer Color Profiling Engines
« Reply #47 on: December 29, 2017, 11:09:04 am »

Well I’ve spent the night googling and reading to find an RGB color working space best approximating human vision (stimulus colors, no imaginary colors) and the best I could find appears to be CIE RGB.  That ICC profile is in my ColorSync directory so that’s readily accessible.  If there’s a significantly better one I’d appreciate any steers.

Interesting point about the printer workspace, ICCLAB. I imagine that the paper ICC profile would reflect the clipping that ICCLAB would do. If so, that part of the puzzle might be solved in software before the image data is sent to the printer.

I created two granger rainbows in PS both in ProPhoto and ColorSync and, just because it’s getting late, soft proofed them using the Argyll-generated, ProPhoto linked ICC profile.  I was somewhat stunned to see the differences, particularly in hue. The luminance banding did appear to be moderately reduced in the CIE RGB file.  Tomorrow I’ll do a more proper experiment and create an Argyll paper profile linked to CIE RGB and print out the two sets of rainbows for comparison, each printed using the appropriate printer profiles. 

I remain a bit concerned that CIE RGB may not be the right colorspace to be using to approximate stimulus colors. Mapping it out in ColorSync against ProPhoto RGB, although it looks like it may be clipping the correct dark blue colors in ProPhoto RGB (and the greens/yellows), the gamut volume rendered in 3D is roughly only about 2/3rds that of ProPhoto RGB, and from what I’ve been reading tonight and seeing in 2D renderings, I would have expected something more like 90+% overlap.  Hmm.

Refer to the CIEXY chromaticity chart show here
https://en.wikipedia.org/wiki/Gamut

There is no RGB colorspace that approximates the CIEXY gamut. It's always a compromise. The RGB colors can all be real colors but they will always leave out some colors outside the triangle they make. RGB colorspaces made from real colors are largest when composed of 3 points on the CIEXY horseshoe. These three points each have only a single wavelength of light. Since all others have zero magnitude any possible print with colors at the corners will be black. This is because only that single wavelength is reflected. Since an illuminant is a continuum, the reflected light will be zero. Colors at these points are only possible with emissive sources such as lasers.

So colorspaces like ProPhoto RGB expands the gamut by introducing imaginary colors. While these don't exist, the larger space allows us to create real, printable colors that otherwise would not be representable in a RGB  colorspace defined only by real colors.

It might be useful to examine the discussion of the CIE 1931 colorspace and how XYZ values are derived from spectral data.

https://en.wikipedia.org/wiki/CIE_1931_color_space
Logged

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20630
  • Andrew Rodney
    • http://www.digitaldog.net/
Re: Choices Between Printer Color Profiling Engines
« Reply #48 on: December 29, 2017, 11:51:41 am »


Color, is a perceptual property. So if you can't see it it's not a color. Color is not a particular wavelength of light. It is a cognitive perception, the excitation of photoreceptors followed by retinal processing and ending in the our visual cortex, within our brains. As such, colors are defined based on perceptual experiments.
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20630
  • Andrew Rodney
    • http://www.digitaldog.net/
Re: Choices Between Printer Color Profiling Engines
« Reply #49 on: December 29, 2017, 11:53:46 am »

As a practical matter if you want to have a gamut that can print all possible printable colors ProPhoto RGB is the only game in town.
On the other hand, no printer can produce all of sRGB.
It's a lot about fitting square holes in round pegs.
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".

Brad P

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 258
Re: Choices Between Printer Color Profiling Engines
« Reply #50 on: December 29, 2017, 12:46:20 pm »

On the other hand, no printer can produce all of sRGB.
It's a lot about fitting square holes in round pegs.

Exactly.  And square holes are probably what I'm knowingly attempting.  Anyway Don Quixote marches on.   

Here's a snapshot of CIE RGB from one of the wikipedia posts Doug sent showing CIE RGB within CIEXY.  What I imagine is there's a bigger RGB workspace that someone smart has drawn within the CIEXY chromaticity diagram that has been thoughtfully optimized to contain the broadest array of colors while still containing no imaginary colors.  I saw earlier suggestions of Don RGB and Beta RGB.  Anyway, I'm going to try one of them of CIE RGB later today.

This all may be a fools errand.  But in the prints here I can visibly see problems likely at least exacerbated by imaginary colors in ProPhoto RGB.  It may well be that those colors rarely or even never will be touched in a normal image processing workflow, but I can imagine in very dark blacks there's some blue.  And I know how I can destroy data in files with pretty powerful tools we have. It very well may be that loosing a larger gamut working space to gain purity in these areas is a trade not worth making.  Certainly that seems to be the opinion of people smarter than me about these things.  But at least it helps me understand a little more about the terms of the trade.
« Last Edit: December 29, 2017, 01:49:19 pm by Brad Paulson »
Logged

Doug Gray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2197
Re: Choices Between Printer Color Profiling Engines
« Reply #51 on: December 29, 2017, 02:50:08 pm »

Exactly.  And square holes are probably what I'm knowingly attempting.  Anyway Don Quixote marches on.   

Here's a snapshot of CIE RGB from one of the wikipedia posts Doug sent showing CIE RGB within CIEXY.  What I imagine is there's a bigger RGB workspace that someone smart has drawn within the CIEXY chromaticity diagram that has been thoughtfully optimized to contain the broadest array of colors while still containing no imaginary colors.  I saw earlier suggestions of Don RGB and Beta RGB.  Anyway, I'm going to try one of them of CIE RGB later today.

This all may be a fools errand.  But in the prints here I can visibly see problems likely at least exacerbated by imaginary colors in ProPhoto RGB.  It may well be that those colors rarely or even never will be touched in a normal image processing workflow, but I can imagine in very dark blacks there's some blue.  And I know how I can destroy data in files with pretty powerful tools we have. It very well may be that loosing a larger gamut working space to gain purity in these areas is a trade not worth making.  Certainly that seems to be the opinion of people smarter than me about these things.  But at least it helps me understand a little more about the terms of the trade.

Brad,

Download the zip file attached to this post:
http://forum.luminous-landscape.com/index.php?topic=114653.msg998229#msg998229

It contains LAB circles every L* 5 from 5 to 95. It's in LAB colorspace so when you view it in proof mode with a printer profile you can see what LAB color gets printed for pretty much any LAB color. It's also useful to see what colors are printable by using the view proof mask which masks out OOG colors. Good way to compare printer profiles.

You can also use it to see where the various RGB color spaces clip by selecting one in soft proof mode and using the gamut mask. It will give you a good idea where printers diverge from RGB colorspaces.
Logged

Brad P

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 258
Re: Choices Between Printer Color Profiling Engines
« Reply #52 on: December 31, 2017, 06:03:45 pm »

Thanks for that file Doug.  I spent an hour monkeying around with it and all types of profiles, old and new, on my computer.

I got to the point of printing Granger Rainbows in CIE RGB, ProPhotoRGB and a few others except for pressing the print button, feeling economic.  Instead that started me off reading color science materials and assorted web pages about Argyll.  I've decided to skip that step of very good, but still generic print spaces, and pursue as Graeme suggested directly producing TIFF specific profiles using tiffgamut, collink, and cctiff (the latter to embed the profile in the files, then later print in noncolormanaged work flow).  That looks like by all accounts what will get me at least very close to an optimal printing workflow. 

I'm happy with the generic Argyll icc printer profile generated earlier and that should work well for noncritical images.  I still have yet to produce my first print that way, but am working on a set of commands now in a mac textedit file (thanks Alan) today and hope to make my first in a few hours. 

After that, I'll probably use the targen program assuming I get the print workflow worked out and regenerate a smoothed out test patch kit.

One interesting twist on all this is I shoot with Hasselblad and do my raw development in Phocus.  They have their own proprietary workspace they call Hasselblad RGB and Hasselblad L* which, according to their marketing literature approximate Adobe RGB and ProPhoto RGB.  They only approximate those spaces in size, looking at them with ColorThink, and are actually very different.  The Hasselblad L* is missing a lot of suspect dark blue colors (I'm suspecting imaginary colors) in ProPhoto RGB and differs a little in greens and some yellows . . .  I've decided in the interim to bring that colorspace into Photoshop as my working space and see how that works out.  It's probably more geared toward my raw processing too.

Assuming I actually get a print after all this, I guess I'll have to make the donation to Argyll!
« Last Edit: December 31, 2017, 06:14:22 pm by Brad Paulson »
Logged
Pages: 1 2 [3]   Go Up