Confused about ColorChecker values

[Hening Bettermann]

1-
When I download Bruce Lindblooms synthetic ColorChecker, convert it to ProPhoto (abs col) and compare the readings of the DigitalColor meter on my Mac to the values given for ProPhoto on Bruce's site, the gray scale is all wrong. The middle gray for one reads 128-128-127 instead of 103-103-102. (BTW it reads almost the same in the original Lab file, 129-128-126.)

The CC looks visually good, as do my images viewed and edited on this screen.

This is on an Eizo CG243W calibrated with a DTP94 to WP 7000K, a bit high but that is the native WP. Calibrated to Lab, but the CC values don't change much if I choose a gamma 2.2 profile.

What is going on here?

2-
I have opted for an Lab workflow, amongst others because my 2 base applications, Raw Developer and PhotoLine, work much in Lab behind the scenes. So I have done my editing on a monitor calibrated to an Lab profile.
Now the print shop requires RGB files, so I will convert to ProPhoto. Question: Should I soft proof with the monitor in a gamma 2.2 profile?

Thank you for your input!

[Scott Martin]

When I download Bruce Lindblooms synthetic ColorChecker, convert it to ProPhoto (abs col) and compare the readings of the DigitalColor meter on my Mac to the values given for ProPhoto on Bruce's site, the gray scale is all wrong. The middle gray for one reads 128-128-127 instead of 103-103-102. (BTW it reads almost the same in the original Lab file, 129-128-126.)

The DigitalColor Meter is kinda worthless isn't it? I believe its showing you equivalent sRGB numbers *after* conversion. Like how it showed you RGB numbers for your Lab file? Right. It's not showing you the values within the file itself - it's showing you converted numeric values. The value of which is questionable. The fact that the numbers are the same for your colorchecker file regardless of color space actually speaks well of your display calibration. 7000K seems awfully high BTW - is your lighting that cool? Do your white onscreen visually match a  white piece of paper?

If you opened your files in Photoshop and set your eyedropper to "Actual Color" I think you'd see some more meaningful values, and those that you're seeking.

[Hening Bettermann]

Thank you Scott!

> If you opened your files in Photoshop and set your eyedropper to "Actual Color" I think you'd see some more meaningful values, and those that you're seeking.

Indeed! So I understand that the DigitalColor Meter shows values after conversion to screen profile. Good to know.

> 7000K seems awfully high BTW - is your lighting that cool? Do your white onscreen visually match a  white piece of paper?

Yes 7000K is a bit high. But it was the native WP of the monitor, and the only one I could make really neutral. I don't exactly recall how I figured that out, but I did before my first calibration of my then new monitor. Paper white, viewed under a Solux lamp, does indeed look a bit reddish in comparison. But the monitor calibrated to 5000K looks much worse. However, that is the printing standard in Europe, and the difference to that will be much larger anyway than the difference between 7000 and 6500.

Good light! - Hening.

[Ellis Vener]

Thank you Scott!
> 7000K seems awfully high BTW - is your lighting that cool? Do your white onscreen visually match a  white piece of paper?

Yes 7000K is a bit high. But it was the native WP of the monitor, and the only one I could make really neutral. I don't exactly recall how I figured that out, but I did before my first calibration of my then new monitor. Paper white, viewed under a Solux lamp, does indeed look a bit reddish in comparison. But the monitor calibrated to 5000K looks much worse. However, that is the printing standard in Europe, and the difference to that will be much larger anyway than the difference between 7000 and 6500.

Good light! - Hening.

What software are you using to calibrate and profile your display with? Eizo's Color Navigator or something else? I have an EIZO CG222W and have no problem calibrating and profiling it to 6500˚K,  80 fc & a gamma of 2.2 using Color Navigator. ¡The ∆E numbers and maximum spread are very tight  and well within the acceptable range.

[Scott Martin]

> 7000K seems awfully high BTW - is your lighting that cool? Do your white onscreen visually match a  white piece of paper?

Yes 7000K is a bit high. But it was the native WP of the monitor, and the only one I could make really neutral. I don't exactly recall how I figured that out, but I did before my first calibration of my then new monitor. Paper white, viewed under a Solux lamp, does indeed look a bit reddish in comparison. But the monitor calibrated to 5000K looks much worse. However, that is the printing standard in Europe, and the difference to that will be much larger anyway than the difference between 7000 and 6500.

7000K still isn't right. Appx 5500K-6100K (not 5000K) should give you a better paper white match depending on which Solux bulbs you have. Neutrality is relative to the color of white. If you're looking at a blue monitor and viewing prints under warm light you've got a problem. Your tones will be "neutral" when they match what your seeing on your prints, even if that seems warm in comparison to what you've been looking at onscreen. Long story short, don't be afraid of a warmer calibration. In the end, I'll think the benefit will be significant.

[Hening Bettermann]

I have now calibrated to 6.500 K (so as to comply with *some* standard...) and get equally neutral whites, according to measurements with the pipette and the DigitalColor meter. So I drop my believe that only 7000K will give me neutral whites, a believe based on measurements I can not recall. - The software I am using now is the Eizo Color Navigator that came with the screen.

I'm not sure how important it is (for me) to match paper white. As long as I don't send my images to the US for printing, they will look much warmer than the screen anyway. But will I (or the visitors of my planned exhibitions in local libraries) ever compare them side by side?

Thank you for your input! - Hening.

[Scott Martin]

I have now calibrated to 6.500 K (so as to comply with *some* standard...) and get equally neutral whites, according to measurements with the pipette and the DigitalColor meter.

Great, now try calibrating to 6000K. :-]

I'm not sure how important it is (for me) to match paper white.

[from memory] "Try it, try it, and you will see. Try it, try it, and you may, I say!"

As long as I don't send my images to the US for printing, they will look much warmer than the screen anyway.

Sending them to the US won't make for cooler prints...

But will I (or the visitors of my planned exhibitions in local libraries) ever compare them side by side?

The idea is to make the process ideal for you, the printmaker, so that you can have reasonable expectations about how your images will look when printed and make the best image development adjustments possible prior to pritning. Your visitors may or may not compare them to your website but you want them to have a "Wow, these prints are incredible - that guy knows his stuff" moment. IMO, perfecting every aspect our printmaking environment helps set the stage for those moments. That said, I have clients that do amazing work in less than stellar environments so do what you like and let the prints speak for themselves.

[Hening Bettermann]

Great, now try calibrating to 6000K. :-]

[from memory] "Try it, try it, and you will see. Try it, try it, and you may, I say!"

I will, I will. Even though I don't quite understand the theory behind it. I do not make color adjustments by eyesight, but trust my profiles. (Exception: if the Automatic White Balance is obviously off, I adjust it by memory, if possible comparing to other images of the same shooting). I have almost no experience with softproofing so far. For the few images I tried, I got away adjusting the TRC only.

While we're talking screen to print:
My images are in Lab, and the screen is calibrated to L*. The print shop requires RGB. So I will convert to ProPhoto. Should I have a softproofing monitor profile RGB gamma 2.2?
Same for the Kelvin grades. The printing reference in Europe is 5000K. Would it make sense to have a monitor calibration of 5000K which I would use for softproofing only?

Good light - Hening.

[Hening Bettermann]

Now I have made a profile with WP 6000. I have also sneak read the thread opened by tony22: "So what IS the right white point for a monitor used for printing?"

So I also bought a pack of Epson Premium Glossy (the paper the print service will use). Of the 3 whitepoints: 6000-6500-7000, 6500 is the one that gives the best visual match to that paper, viewed under my Solux lamp.

So now there is only one question left: Since I will convert my images from Lab to ProPhoto for printing, should I make a profile with gamma 1.8 for softproofing? Or use 2.2, or the L* profile?

Good light - Hening.

[tony22]

Hening, you hijacked my thread! Just kidding. I'm curious as to what result you got on the Epson Premium when you had the monitor calibrated to 6000. What sort of tonal shift did you get on the print?

[Jon Maxim]

... Even though I don't quite understand the theory behind it. I do not make color adjustments by eyesight, but trust my profiles. (Exception: if the Automatic White Balance is obviously off, I adjust it by memory, if possible comparing to other images of the same shooting). I have almost no experience with softproofing so far. For the few images I tried, I got away adjusting the TRC only.

Hening, May I humbly suggest you read this exceptional article by Andrew Rodney in Lu-La:

http://www.luminous-landscape.com/tutorials/why_are_my_prints_too_dark.shtml

I (a long ago former professional film phototgapher but now decidedly amateur digital photographer) have read extensively about colour management, bought expensive hardware and software, but was never able to get what I considered a good match. That was until I read the article. You may find it will answer your concerns.

Jon

[tony22]

There's one thing about that article that I've always wanted to get more info about. Andrew writes:

"Note that altering a warm/cool bias using a CCT value is a viable solution, however, if a cast on the magenta/green axis is seen, this cannot be fixed with sole a Kelvin adjustment. Products that allow editing the white point using x and y axis of the white point on a CIE xy plot will allow colors on these axis to be adjusted."

I'd like to learn a bit more about what tools would allow for a correction of this sort.

[Hening Bettermann]

@ tony22 reply #9

Uff! There is obviously something I misunderstood. I do not have an inkjet, and not at all an Epson 9800 (which is the model the print shop will use). So I have not printed on this paper, just put it up next to the monitor, illuminated it with a Solux lamp and compared it visually to a blank page in TextEdit.

Good light! - Hening.

[Hening Bettermann]

@ jmaxim reply #10

Hi jon

thank you for the pointer. I am now studying it.

[digitaldog]

I'd like to learn a bit more about what tools would allow for a correction of this sort.

When altering WP with K values, the issue is you are using a 1 dimensional variable and you want to correct a 2 dimensional problem (color shift in CIE xy).

If the color cast happens to be warm/cool then adjusting the white point in Kelvin should work fine. But if your color cast is say a green/magenta for most of the black body curve, then you can't fix it with a simple Kelvin adjustment.

When people report their display is too green/magenta, the fix can be zooming in on the central white point of the CIE xy diagram in the SpectraView software. It will give you a better idea of how changing the x and y values will affect the white point. If the screen is too green, you‘ll want to move the white point down and to the right (vise versa for magenta). Make tiny adjustments (i.e. 0.002) and recalibrate. Understand that this only works if the calibration software allows you to enter specific CIE xy values.

[tony22]

Understand that this only works if the calibration software allows you to enter specific CIE xy values.

I'm not there yet (with a NEC). All I have are RGB Gain and Cut controls.

[Hening Bettermann]

So now I have studied Andrews LuLa article and the 2004 series of pdf articles on color management available on his web site. Together, they make indeed up what I needed: a hands-on tutorial on "From Master File To Print". Many thanks to Andrew for providing these articles!