Ad
Ad
Ad
Pages: [1]   Bottom of Page
Print
Author Topic: X-Rite Color Checker Passport  (Read 3581 times)
JessicaLuchesi
Full Member
***
Offline Offline

Posts: 128


WWW
« on: January 15, 2010, 02:45:44 PM »
ReplyReply

I have a crazy question, which may be completely absurd, but this is what I was thinking:

If you calibrate your screen, take a properly exposed photo of the color checker, use the plugin to generate a DNG profile, use the profile, white balance the image, and place the actual physical target besides the screen, we should see pretty much in the screen the same target we have in our hands, right?

If that is correct, couldn't the passport be used to check the actual calibration of the monitor?
Logged
ErikKaffehr
Sr. Member
****
Offline Offline

Posts: 7630


WWW
« Reply #1 on: January 15, 2010, 03:02:46 PM »
ReplyReply

Hi,

There are some issues, too many variables! I would have another suggestion:

Download one of those synthetical Color Checker targets, preferably one in LAB or at least with an embedded color profile.

I would recommend this one: http://www.brucelindbloom.com/downloads/Co...culator.tif.zip
Now open it in photoshop and try:

View->Proof->Setup->Custom

Choose any good paper profile or "Fogra Coated" or something useful. Check the same boxes as I have below:

[attachment=19477:Screenshot.jpg]

This should look as a good approximation of your CC-card. If it looks OK, fine, if it does not you need to do something!


Best regards
Erik

Quote from: JessicaLuchesi
I have a crazy question, which may be completely absurd, but this is what I was thinking:

If you calibrate your screen, take a properly exposed photo of the color checker, use the plugin to generate a DNG profile, use the profile, white balance the image, and place the actual physical target besides the screen, we should see pretty much in the screen the same target we have in our hands, right?

If that is correct, couldn't the passport be used to check the actual calibration of the monitor?
« Last Edit: January 15, 2010, 03:08:43 PM by ErikKaffehr » Logged

JessicaLuchesi
Full Member
***
Offline Offline

Posts: 128


WWW
« Reply #2 on: January 15, 2010, 05:20:45 PM »
ReplyReply

Quote from: ErikKaffehr
Hi,

There are some issues, too many variables! I would have another suggestion:

Download one of those synthetical Color Checker targets, preferably one in LAB or at least with an embedded color profile.

I would recommend this one: http://www.brucelindbloom.com/downloads/Co...culator.tif.zip
Now open it in photoshop and try:

View->Proof->Setup->Custom

Choose any good paper profile or "Fogra Coated" or something useful. Check the same boxes as I have below:

[attachment=19477:Screenshot.jpg]

This should look as a good approximation of your CC-card. If it looks OK, fine, if it does not you need to do something!


Best regards
Erik

Hey Erik,

Thank you so much for your help, your tips did help a lot.

I know it sounds crazy, but base idea is, if the CC Passport (and it's plugin) works, then it has to be a basis to generating a image comparable to the target. Also, comparable to the several 24-patch images available.

Here's the thing, it worked, weird enough. The CC Passport was properly exposed with a simple softbox, no reflexes or shades on it, done with flash, and with a flashmeter, not simply camera metering. Once the screen was calibrated to 110cd/m2, and black point to 0.35cd/m2 ( the lowest I could go ), I ran the plugin, created the profile for the DNG image, White Balanced in Lightroom, the image of the target did indeed come very very close to the actual target, being held side by side to the screen.

Also, the actual target came very very close to the image posted by Tim on the other thread. The downloaded 24-patch LAB image still has a "washed out" feeling to it tho. But the actual colors are spot on.

But, the actual calibration was done using ColorEyes Display Pro, not the actual Spyder software.
Logged
digitaldog
Sr. Member
****
Offline Offline

Posts: 9185



WWW
« Reply #3 on: January 15, 2010, 05:24:43 PM »
ReplyReply

Quote from: JessicaLuchesi
If that is correct, couldn't the passport be used to check the actual calibration of the monitor?


Maybe, not necessarily etc. Should get you closer to your goal but humans need to view these big plies of numbers on the different devices (and real life) and render the image to result in what they fell is a match. And even with the best color management in the world, you can nail some colors while others are off (due to something known as metameric failure). There are big differences in what are known as a reference media. Just think about the huge contrast ratio/dymamic range of that passport in daylight compared to that of your display (and its backlight compared to actual sunlight filtered however). Think about the surrounding colors in a scene versus those when you view a display or print. Plus the science that all this stuff is based upon is kind of old and was never designed way back when, for what we are doing today with computer color management. Until we have color appearance models being used, take many of the issue above into account, there are all kinds of areas that can bite us in the butt and require human intervention (image editing).
Logged

Andrew Rodney
Author “Color Management for Photographers”
http://digitaldog.net/
JessicaLuchesi
Full Member
***
Offline Offline

Posts: 128


WWW
« Reply #4 on: January 15, 2010, 05:47:08 PM »
ReplyReply

Quote from: digitaldog
Maybe, not necessarily etc. Should get you closer to your goal but humans need to view these big plies of numbers on the different devices (and real life) and render the image to result in what they fell is a match. And even with the best color management in the world, you can nail some colors while others are off (due to something known as metameric failure). There are big differences in what are known as a reference media. Just think about the huge contrast ratio/dymamic range of that passport in daylight compared to that of your display (and its backlight compared to actual sunlight filtered however). Think about the surrounding colors in a scene versus those when you view a display or print. Plus the science that all this stuff is based upon is kind of old and was never designed way back when, for what we are doing today with computer color management. Until we have color appearance models being used, take many of the issue above into account, there are all kinds of areas that can bite us in the butt and require human intervention (image editing).

I know Andrew. It's just that I read in several places that Spyder devices could not be trusted, and then, well, fear factor. The last thing I needed was reason to second guess how accurate my calibration was when I am starting out as photographer for a new magazine.

I mean, all I wanted to know was if I would be seeing something quite different than the actual product, or worse, evaluating exposures in an incorrect way and either sending over or under exposed images to clients as a result. In a short, I wanted to be comfortable that I know what I have a really good parameter of what I'm sending to the client, even if I know I cannot fully predict how an image will come out of press, or even how they're seeing it in their computers.

I know I went into a lot of drama, but in the end, I think I can trust my calibration device. And found a software ( the ColorEyes ) that gave me a result I think is more spot on.

Besides, like a friend reminded me, it's a notebook screen, not an Eizo monitor.
Logged
ErikKaffehr
Sr. Member
****
Offline Offline

Posts: 7630


WWW
« Reply #5 on: January 15, 2010, 11:45:56 PM »
ReplyReply

:-) !



Quote from: JessicaLuchesi
Hey Erik,

Thank you so much for your help, your tips did help a lot.

I know it sounds crazy, but base idea is, if the CC Passport (and it's plugin) works, then it has to be a basis to generating a image comparable to the target. Also, comparable to the several 24-patch images available.

Here's the thing, it worked, weird enough. The CC Passport was properly exposed with a simple softbox, no reflexes or shades on it, done with flash, and with a flashmeter, not simply camera metering. Once the screen was calibrated to 110cd/m2, and black point to 0.35cd/m2 ( the lowest I could go ), I ran the plugin, created the profile for the DNG image, White Balanced in Lightroom, the image of the target did indeed come very very close to the actual target, being held side by side to the screen.

Also, the actual target came very very close to the image posted by Tim on the other thread. The downloaded 24-patch LAB image still has a "washed out" feeling to it tho. But the actual colors are spot on.

But, the actual calibration was done using ColorEyes Display Pro, not the actual Spyder software.
Logged

probep
Full Member
***
Offline Offline

Posts: 149


« Reply #6 on: January 16, 2010, 12:33:08 AM »
ReplyReply

Quote from: digitaldog
Maybe, not necessarily etc. Should get you closer to your goal but humans need to view these big plies of numbers on the different devices (and real life) and render the image to result in what they fell is a match. And even with the best color management in the world, you can nail some colors while others are off (due to something known as metameric failure). There are big differences in what are known as a reference media. Just think about the huge contrast ratio/dymamic range of that passport in daylight compared to that of your display (and its backlight compared to actual sunlight filtered however). Think about the surrounding colors in a scene versus those when you view a display or print. Plus the science that all this stuff is based upon is kind of old and was never designed way back when, for what we are doing today with computer color management. Until we have color appearance models being used, take many of the issue above into account, there are all kinds of areas that can bite us in the butt and require human intervention (image editing).
I have recalled the old report: "Our investigations have shown that visual matches between a hard-copy and its reproduction
on a CRT monitor can show considerable colorimetric differences." The color difference reached dE=14.
I think it is true not only for CRT monitors.
Logged
digitaldog
Sr. Member
****
Offline Offline

Posts: 9185



WWW
« Reply #7 on: January 16, 2010, 10:57:29 AM »
ReplyReply

Quote from: probep
I think it is true not only for CRT monitors.

Why would that be the case?
Logged

Andrew Rodney
Author “Color Management for Photographers”
http://digitaldog.net/
probep
Full Member
***
Offline Offline

Posts: 149


« Reply #8 on: January 16, 2010, 12:08:13 PM »
ReplyReply

Quote from: digitaldog
Why would that be the case?
Maybe I am wrong. But LCD as well as CRT monitors have high peaks in the spectral power distribution curves.
Logged
Tim Lookingbill
Sr. Member
****
Offline Offline

Posts: 1193



WWW
« Reply #9 on: January 16, 2010, 01:09:09 PM »
ReplyReply

Quote from: probep
Maybe I am wrong. But LCD as well as CRT monitors have high peaks in the spectral power distribution curves.

I would agree with you on this even though your white paper on the subject was too complex for me to understand.

I can tell you strictly from observation and implementation of color management that I don't think any complex mathematical colorimetric calculation can correct for the display's "filter effect" and its influence on perception of a print lit by a full spectrum light source compared to the spikey LCD screen to allow a perfect match.

This photo.net thread:

http://photo.net/digital-darkroom-forum/00UbBI  

I started a while back illustrates what I call this "filter effect" using the Solux lamp to illuminate surfaces that exhibit different spectral responses under a controlled color temp appearance in this case the Solux's supposedly D50. Color wasn't an issue because adaptation under these lights corrected for it, but it wasn't happening on the display because the color tables needed to be corrected for and weren't.

The profile that would have to be created for forcing the display to correct for this filter effect would require major color table manipulation because expecting adaptation to the display's warm cast clearly isn't working to make the hue and saturation of the colors calibrated to 6500K look the same under 5000K calibration.

Or maybe the EyeOne Match software is just expecting to only calibrate to a 6500K standard and any deviation from it will introduce errors in the final transforms that are built into the 5000K profile. I'm not sure, but no one has yet explained why my prints match on my 6500K display viewed under a warmish Solux's D50.

Does the math in these profiles anticipate the effects of color constancy is what I'ld like to know.

This is what color looks like under D50 Solux of standard color targets. No filter effect.

[attachment=19496:OleNoMoi...kinSolux.jpg]
« Last Edit: January 16, 2010, 01:38:52 PM by tlooknbill » Logged
Pages: [1]   Top of Page
Print
Jump to:  

Ad
Ad
Ad