Monitor profiling for soft proof: Mismatch beween visual match and calculated WP

[Hening Bettermann]

I have tried to calibrate my Eizo CG 243 for soft proofing and used the DTP94 and the Eizo software to measure the paper white of the Epson Premium Glossy. The software then sets the white point for the profile automatically. It resulted in 4849 K. However, when I compare the paper illuminated by a Solux lamp to the enlarged screen image of the ColorChecker White patch of Bruce Lindblooms synthetic ColorChecker, the screen image is grossly yellowish, and the visual match to the paper is much better with a 6500 K profile.

Whom should I trust?

Another problem concerns the choice between gray balance and contrast priorities. I tried both. The Contrast resulted in a contrast of 576:1, but in a ΔE of up to 3.4 in the blacks. Gray Scale resulted in a ΔE max of 0.5, for black 0-0-0, but in a contrast of only
236:1.

I remember from one of Andrew Rodneys articles, that a reflection print has a contrast of 350:1 at best. So unless the Epson Premium Glossy is a super paper, 576:1 is obviously too high.
Does this mean the software (or my measuring) are very unreliable in this respect?
How realistic then is the 236:1 figure?
And how reliable is either profile as far as contrast goes, white balance aside?

Thank you for your input - Hening.

[digitaldog]

Whom should I trust?

The setting that produces the best visual match to the print.

How realistic then is the 236:1 figure?

More realistic than 576! Again, whatever setting provides the best visual match is the one to go with.

[Hening Bettermann]

Thank you for your reply, Andrew.

White point:
> The setting that produces the best visual match to the print.

Well in this case it is not a print (yet), but a sheet of unprocessed paper. Would it hold for this as well? (I do not print myself, I'm trying to prepare my first test print for the print service.)

Contrast:
> Again, whatever setting provides the best visual match is the one to go with.

I can not choose the contrast rate directly, I can choose between Gray Balance or Contrast priority for the profile, then the software does the rest.
I can only control the luminance of the paper by adjusting the distance of the Solux lamp. But to which luminance?
To make up my own "standard", I went to my local library, held a gray card up in front of the exposed images there, and my spot meter showed EV 8. So I adjusted the illumination of my paper sheet to this value.
So I understand what I have to do is not *measure* the paper with this lighting, but find the monitor profile luminance (in cd) which gives the best visual match in perceived brightness to the paper. (The software chose 100 cd.)

So did I after trying 80, 100 and 120 cd; and I ended up with 5800K for the WP.

The validation for this profile states a contrast ratio of 371:1 (gray balance priority). If this does not happen to be a realistic value, there seems to be no way of adjusting contrast in a controlled way to mimic the contrast of paper.

Curious to see the first test print…

Thanks again for your help!  - Hening.

[digitaldog]

Well in this case it is not a print (yet), but a sheet of unprocessed paper. Would it hold for this as well? (I do not print myself, I'm trying to prepare my first test print for the print service.)

You really need one (or more) images, not just paper white. You need to have a soft proof on as well.

I can not choose the contrast rate directly, I can choose between Gray Balance or Contrast priority for the profile, then the software does the rest.

Not knowing the software, I’d think you would pick contrast priory and somewhere in the software, you can set a specific ratio. You’ll have to check the manual.

I can only control the luminance of the paper by adjusting the distance of the Solux lamp. But to which luminance?

You can move the light, sure. When it visually matches, you’re good.

The validation for this profile states a contrast ratio of 371:1 (gray balance priority).

That’s pretty high, certainly higher than a print, even on glossy paper.

[Hening Bettermann]

You are really fast!

> You really need one (or more) images, not just paper white. You need to have a soft proof on as well.

I'm not sure if you misunderstand me. My question meant: "Does an unprocessed sheet of paper qualify as paper white, or do I need to make an (overexposed ?) image of a white card, push it through my workflow and have it printed?" At this stage, I am inside the Color Navigator software, there is no soft proof to switch on. I am so far just trying to prepare a monitor profile for soft proofing.

> Not knowing the software, I’d think you would pick contrast priory and somewhere in the software, you can set a specific ratio. You’ll have to check the manual.

No there is no such choice. It looks like "contrast priority" means the software tries to maximize the contrast. The value of 576:1 might look like this.

> You can move the light, sure. When it visually matches, you’re good.

It is the other way around. I move the lamp until the gray card shows an EV of 8 (10+1/3 for the white paper). Then I try to pick the profile with the luminance that gives the best visual match between the paper and screen white. 100 cd looks quite good to me. (For the screen white, I returned to a blank document of TextEdit, which reads 255-255-255 in the DigitalColor meter, whereas the White patch of the synthetic ColorChecker says 245-243-239.)   

>> The validation for this profile states a contrast ratio of 371:1 (gray balance priority).
> That’s pretty high, certainly higher than a print, even on glossy paper.

Yes that is exactly my question. (How) is it possible to build a profile that emulates the true contrast ratio of (the particular) paper? It sounds like I would need new software? And maybe even a new monitor ??, since the combo would need to allow the contrast adjustment the way you describe it. I thought I had made a lifetime investment with that Eizo...

Uff!

[digitaldog]

My question meant: "Does an unprocessed sheet of paper qualify as paper white, or do I need to make an (overexposed ?) image of a white card, push it through my workflow and have it printed?" At this stage, I am inside the Color Navigator software, there is no soft proof to switch on. I am so far just trying to prepare a monitor profile for soft proofing.

You need one or more images printed and being viewed while you soft proof those images using the paper profile inside of Photoshop.

It is the other way around. I move the lamp until the gray card shows an EV of 8 (10+1/3 for the white paper). Then I try to pick the profile with the luminance that gives the best visual match between the paper and screen white. 100 cd looks quite good to me. (For the screen white, I returned to a blank document of TextEdit, which reads 255-255-255 in the DigitalColor meter, whereas the White patch of the synthetic ColorChecker says 245-243-239.)   

The numbers are meaningless. Only a visual match is meaningful. If 100 looks good, 100 is correct.

(How) is it possible to build a profile that emulates the true contrast ratio of (the particular) paper?

Its not the profile, its the calibration of the display. I don’t have an Eizo. But on the NEC, I can set the display contrast ratio to any value I wish. The calibration of the black and white luminance are then adjusted in an attempt to produce that.

[Hening Bettermann]

> Its not the profile, its the calibration of the display. I don’t have an Eizo. But on the NEC, I can set the display contrast ratio to any value I wish. The calibration of the black and white luminance are then adjusted in an attempt to produce that.

Right before I ordered a NEC to be able to adjust the contrast manually, I wondered if there was a workaround. And I learned that simulating paper white and black ink in the proof are supposed to do just that: to display the image in the contrast range of the paper. But (how) can the software do this regardless of the point of departure, i.e. the contrast range of the monitor BEFORE the proofing is enabled? since this range does not seem to be linked to the paper contrast in any way.

My luminance levels under test have the following contrast ranges:
80 cd - 410:1; 100 cd - 374:1; 120 cd - 384:1.
(gray scale priority, which seems to be the same as what other applications call "calibrated to deepest NEUTRAL black".)

 Do I really need a new monitor?

[Hening Bettermann]

> Do I really need a new monitor?

Well so far I have bought a new puck, Spyder3, to cope with the wide gamut monitor, and ColorEyes software. 

I'm getting ever more confused on this.

My plan was/is to use Andrews test image, compare the native image to the image with softproof enabled, then adjust the white point in Photoshop (rather than in the monitor profile), then later to transfer the adjustments as a baseline to my images.
Proof setup: RelCol, +BPC, +simulate black ink, +simulate paper color
Softproof disabled to begin with.

1-When I open Andrews test image by double-click, it opens in Preview. The White field in the step wedge reads 255-255-255 in the DigitalColorMeter.

2-When I open the image in Photoshop, the White reads 187-187-187 in the DCM, and the visual impression is accordingly. PS Info shows 255-255-255.

??

3-When I leave the image open, but click in the Finder window so PS is not active,  the DCM, too, shows 255 for the White field. --?

Now with soft proof enabled, the white field reads 174-176-182; with the image still open in PS, but PS not active, the values are 237-240-249.--??

Before I had discovered the difference between these latter 2 readings, I had used the latter triplet as input in an adjustment layer curves, setting output to 255, and got a real nice match between the proofed image and the native image opened in PS.

Still, I feel I need to understand better what is going on before I proceed. In particular, the difference between Preview and Photoshop puzzles me.

I understand that softproofing the white of glossy/OBA papers like "my" Epson Premium Glossy is questionable, and am prepared to have to modify these adjustments later.

Hoping for your help.

[digitaldog]

It opens in Preview because that’s the application you have associated in the Finder to open such files. Either fix that (get info) or drag and drop over Photoshop.

2-When I open the image in Photoshop, the White reads 187-187-187 in the DCM, and the visual impression is accordingly. PS Info shows 255-255-255.

DigitalColorMeter is useless for whatever you’re trying to do here.

[Hening Bettermann]

Thank you for your fast reply.

> It opens in Preview because that’s the application you have associated in the Finder to open such files. Either fix that (get info) or drag and drop over Photoshop.

I see that I have to re-phrase that question: "Why does Preview show colors different from Photoshop? I thought either was color-managed?"

[digitaldog]

If tagged, it should show the same previews.

[Hening Bettermann]

The difference I saw and see is subtle, and I can now see that it must have something to do with my sight. Whatever image I place to the right seems slightly darker  to me, also if I change the location on the screen. Strange.

[Hening Bettermann]

> Now with soft proof enabled, the white field reads 174-176-182; with the image still open in PS, but PS not active, the values are 237-240-249.--??

Mystery solved. That was with the Marquee tool in PS selected, so the black crosshair interfered with the reading. With the Move tool selected, readings are the same under all conditions.