Luminous Landscape Forum

Raw & Post Processing, Printing => Colour Management => Topic started by: The View on January 05, 2018, 12:28:51 am

Title: High Delta E after monitor calibration
Post by: The View on January 05, 2018, 12:28:51 am
I just re-calibrated my NEC PA 271W, 5 years old, 15 000 hours running (but with screen to turn off after 5 minutes of non-usage). Calibrated with i1 Display Pro and Spectraview II 1.1 36 with a Mac running 10.12.6 Sierra.

Calibrated towards 140.0/cd/m2  a Gamma of 2.20 and a D65 white point. contrast ratio monitor default, color gamut native (full)

I haven't done it in a long time and now the delta e of all values over a luminance value of 20 is in the red with an average delta e of 11.74

Does this mean my monitor is burnt out or has the calibration not worked?

Delta E is the relation of color to gray scale value - what does 7.7 signify in a comparable fact?
Title: Re: High Delta E after monitor calibration
Post by: Doug Gray on January 05, 2018, 12:49:05 am
I just re-calibrated my NEC PA 271W, 5 years old, 15 000 hours running (but with screen to turn off after 5 minutes of non-usage). Calibrated with i1 Display Pro and Spectraview II 1.1 36 with a Mac running 10.12.6 Sierra.

Calibrated towards 140.0/cd/m2  a Gamma of 2.20 and a D65 white point. contrast ratio monitor default, color gamut native (full)

I haven't done it in a long time and now the delta e of all values over a luminance value of 20 is in the red with an average delta e of 11.74

Does this mean my monitor is burnt out or has the calibration not worked?

Delta E is the relation of color to gray scale value - what does 7.7 signify in a comparable fact?

Those are very large errors. You might try lower luminance settings. As they age monitors tend to drop in max luminance but they still can be profiled with fairly low dE errors. Unless there is some compatibility issue or the monitor is failing.
DeltaE is the geometric distance between two points in L*a*b* space. Math wise it is sqrt((L1-L2)^2+(a1-a2)^2+(b1-b2)^2). One can visualize it as the distance between 2 points inside a 3D box.
Title: Re: High Delta E after monitor calibration
Post by: The View on January 05, 2018, 01:18:24 am
Those are very large errors. You might try lower luminance settings. As they age monitors tend to drop in max luminance but they still can be profiled with fairly low dE errors. Unless there is some compatibility issue or the monitor is failing.
DeltaE is the geometric distance between two points in L*a*b* space. Math wise it is sqrt((L1-L2)^2+(a1-a2)^2+(b1-b2)^2). One can visualize it as the distance between 2 points inside a 3D box.

But what does it mean in colors. That colors appear too dark or is it the contrast of the colors? I know when a tire has only 15 psi in it I better not drive.

So, is my monitor low on air and will I be soon riding on the rim?

I'm shooting professionally and do post production as well.

I'll reduce the luminance a bit and see what I'll get. A lower luminance is OK at night, but during the day it can make editing tough.
Title: Re: High Delta E after monitor calibration
Post by: The View on January 05, 2018, 01:26:13 am
I just did another calibration and it totally failed (even though spectraview OK'd it). I can see it the whites are too blue.

In three instances I had noticed that after a mouseclick the display got brighter...

I reduced the intensity to 130 cd/m2 and got the Delta e down to 6.71, average 3.7

I'll see how daytime editing is with this or if I can turn down the intensity even more.

It just did that thing again where the display decreased intensity in a step just like it does when profiling it... 
Title: Re: High Delta E after monitor calibration
Post by: Doug Gray on January 05, 2018, 01:46:43 am
But what does it mean in colors. That colors appear too dark or is it the contrast of the colors? I know when a tire has only 15 psi in it I better not drive.

DeltaE is used as an indication of how different two "colors" are but say nothing about the colors involved. They can be neutral grays, or variably saturated colors of any hue.
 
https://en.wikipedia.org/wiki/Color_difference
https://en.wikipedia.org/wiki/Lab_color_space
Title: Re: High Delta E after monitor calibration
Post by: Mark D Segal on January 05, 2018, 08:49:14 am
...............

Delta E is the relation of color to gray scale value - what does 7.7 signify in a comparable fact?

No. Delta E is an error term that measures the difference between a target reference value for a colour patch (gray or any other colour) and the value the device achieved trying to replicate that target value. Normally it is measured using the LAB scale as Doug indicated above. I agree with Doug that the errors you report are much too large. If 7.7 is the average error the profiling evaluation is reporting, it means that on average for all the patches measured there are large disconnects between the target reference values and the corresponding colours the display is projecting. They should be less than 1.0 on average. It is impossible to know the cause of these large errors based on the information provided. It could be a hardware problem with the display, or with the measuring instrument, or with a setting in the display or the colour management software causing this. If you can try the exercise using a different colorimeter on the same display or a different display with the same colorimeter, that would help diagnose whether there is a problem with either the display or the colorimeter. Also check all your calibration settings for the display and the measurement conditions in the software -  I don't know Spectraview so I can't advise on the details, as I use BasicColor for my display profiling. But Spectraview should also give you fine results if all is working properly.
Title: Re: High Delta E after monitor calibration
Post by: John Hollenberg on January 05, 2018, 09:58:14 am
Sounds like your monitor is failing.  I have 2,000 hours on my NEC PA272W and average delta E of 0.48 and maximum 0.98.  There isn't much to go wrong with Spectraview software.
Title: Re: High Delta E after monitor calibration
Post by: digitaldog on January 05, 2018, 10:29:52 am
Delta-E and color accuracy
In this 7 minute video I'll cover: What is Delta-E and how we use it to evaluate color differences. Color Accuracy: what it really means, how we measure it using ColorThink Pro and BableColor CT&A. This is an edited subset of a video covering RGB working spaces from raw data (sRGB urban legend Part 1).

Low Rez: https://www.youtube.com/watch?v=Jy0BD5aRV9s&feature=youtu.be (https://www.youtube.com/watch?v=Jy0BD5aRV9s&feature=youtu.be)
High Rez: http://digitaldog.net/files/Delta-E%20and%20Color%20Accuracy%20Video.mp4 (http://digitaldog.net/files/Delta-E%20and%20Color%20Accuracy%20Video.mp4)
Title: Re: High Delta E after monitor calibration
Post by: Frodo on January 05, 2018, 12:52:54 pm
I had a broadly similar issue on an AOC screen using the Spyder 4 and datacolor profili g software. The screen gained a magenta tone despite multip!e profiling. I thought my Spyder was dying (I bought an X-rite display pro), but uninstalling and reinstalling the Datacolor software solved the problem.
Title: Re: High Delta E after monitor calibration
Post by: Alan Goldhammer on January 05, 2018, 01:33:10 pm
I had a P222 NEC monitor in use for over 8 years and it was for both photo and general use.  I never saw dE values that were as high as the OP observed.  I always calibrated to 120.  If you are seeing visible changes it could mean the monitor is beginning to fail but also might mean that your Display Pro is going bad as well.  If you know of someone in your area who has SpectraView compatible puck, you could at least run a calibration using that to rule out the latter point.
Title: Re: High Delta E after monitor calibration
Post by: lhodaniel on January 05, 2018, 03:24:34 pm
I would first completely uninstall (I use CleanMyMac or CleanApp for this) then reinstall Spectraview and see if that fixes it. To check the colorimter, try an manual calibration using the OSD and i1Profiler. If the DisplayPro is wonky it should show there.

My 13300 hour P241 showed signs of the backlight yellowing, so I just ran the Self Color Correction routine in the Advanced Settings OSD (category B) and that seemed to help. But, I never had bad delta E's from SVII anyway. It was just having to make a fairly large correction for white balance at the beginning of the calibration.

Lloyd
Title: Re: High Delta E after monitor calibration
Post by: The View on January 05, 2018, 06:16:27 pm
What worries me is that up and down of brightness.

I had been writing for several hours and not experienced it.

Then I come to this website, and within ten seconds I have had three up-down movements. Not random. Always the same two levels.

It could be that the display is adapting for the huge white areas of the post windows, because after it did this a few times brightness remained stable.

A bit like what the display does when you power up the computer.
Title: Re: High Delta E after monitor calibration
Post by: The View on January 05, 2018, 06:19:03 pm
I would first completely uninstall (I use CleanMyMac or CleanApp for this) then reinstall Spectraview and see if that fixes it. To check the colorimter, try an manual calibration using the OSD and i1Profiler. If the DisplayPro is wonky it should show there.

My 13300 hour P241 showed signs of the backlight yellowing, so I just ran the Self Color Correction routine in the Advanced Settings OSD (category B) and that seemed to help. But, I never had bad delta E's from SVII anyway. It was just having to make a fairly large correction for white balance at the beginning of the calibration.

Lloyd

That self Color Correction - is it part of Spectraview II?

I don't think you need a cleaning app on a mac - you just pull the app into the trash.
Title: Re: High Delta E after monitor calibration
Post by: The View on January 05, 2018, 06:20:29 pm
Sounds like your monitor is failing.  I have 2,000 hours on my NEC PA272W and average delta E of 0.48 and maximum 0.98.  There isn't much to go wrong with Spectraview software.

Yes, I have that impression.

I can lessen the errors by reducing brightness.
Title: Re: High Delta E after monitor calibration
Post by: The View on January 05, 2018, 06:30:06 pm
Delta-E and color accuracy
In this 7 minute video I'll cover: What is Delta-E and how we use it to evaluate color differences. Color Accuracy: what it really means, how we measure it using ColorThink Pro and BableColor CT&A. This is an edited subset of a video covering RGB working spaces from raw data (sRGB urban legend Part 1).

Low Rez: https://www.youtube.com/watch?v=Jy0BD5aRV9s&feature=youtu.be (https://www.youtube.com/watch?v=Jy0BD5aRV9s&feature=youtu.be)
High Rez: http://digitaldog.net/files/Delta-E%20and%20Color%20Accuracy%20Video.mp4 (http://digitaldog.net/files/Delta-E%20and%20Color%20Accuracy%20Video.mp4)

Good video, thank you.

I noticed that with bigger patches a 5.66 delta E is clearly noticeable, with smaller patches it is subtle.

So I'll likely get my brightness down by another 5 cd/m2 and I'll be OK for now.
Title: Re: High Delta E after monitor calibration
Post by: digitaldog on January 05, 2018, 06:53:38 pm
I can lessen the errors by reducing brightness.
Raise it, no harm done but you'll need to adjust the print viewing conditions next to the display for a new match.
I run my PA272W at 150cd/m2 because I can easily adjust my print viewing station and very low levels are harder to hit, certainly as displays age.
Title: Re: High Delta E after monitor calibration
Post by: lhodaniel on January 05, 2018, 06:56:30 pm
That self Color Correction - is it part of Spectraview II?

I like the apps because they clean out all the prefs etc that might be causing the problem. And, I was Windows for 23 years so it might be inertia.  :D

No, it's in the OSD, RGB section "Advanced Settings, then scroll to tab "B". I read somewhere long ago that this was to recal the monitor from age effects. I have upgraded to the latest firmware of the monitor, BTW.

Lloyd
Title: Re: High Delta E after monitor calibration
Post by: xpatUSA on January 05, 2018, 07:23:22 pm
I haven't done it in a long time and now the delta e of all values over a luminance value of 20 is in the red with an average delta e of 11.74

Does this mean my monitor is burnt out or has the calibration not worked?

No idea but 11.74 is unacceptable for serious work.

Google for "just noticeable difference" + CIE. The value varies but 2.3 seems popular in the literature.

[/quote]Delta E is the relation of color to gray scale value[/quote]

No, it is not, as others have already stated.

The only thing I am bringing extra to the table is the "Just Noticeable Difference", well worth learning about . .

Ted
Title: Re: High Delta E after monitor calibration
Post by: Mark D Segal on January 05, 2018, 07:53:03 pm
Ya, but the "JND" is supposed to be a dE of 1.0 in LAB scale.
Title: Re: High Delta E after monitor calibration
Post by: Tim Lookingbill on January 05, 2018, 10:27:51 pm
What worries me is that up and down of brightness.

I had been writing for several hours and not experienced it.

Then I come to this website, and within ten seconds I have had three up-down movements. Not random. Always the same two levels.

It could be that the display is adapting for the huge white areas of the post windows, because after it did this a few times brightness remained stable.

A bit like what the display does when you power up the computer.

This is when you start typing or just hitting the keyboard, right?

You might just have a USB connection issue if at any point you are utilizing that type of connection during and after profiling. I'm not familiar with how your display setup is but just as it has been suggested, test by hooking up another monitor (cheapo or old used one) and don't even calibrate it or hook up any device through the USB or any other connection and see if you get this increase/decrease in brightness associated with keyboard activity.

If it doesn't do the brightness fluctuation, then it may be the internal electronics of the NEC display at issue or near or at failure. If it does with the test display, then it's most likely with your internal electronics of your device connection pathway be it USB, Firewire or whatever is used during NEC connects and profiling routines.
Title: Re: High Delta E after monitor calibration
Post by: Jack Hogan on January 06, 2018, 12:23:37 pm
Ya, but the "JND" is supposed to be a dE of 1.0 in LAB scale.

It seems that more recently folks have been using 1 dE2000 as a JND.  dE2000 is in general less sensitive than plain old dE(76) - that is it takes more dE's to make 1 dE2000.

Jack
Title: Re: High Delta E after monitor calibration
Post by: digitaldog on January 06, 2018, 12:49:21 pm
The formula is important! I prefer dE2000 because in most cases, I'm interested in small(er) differences (distances) between two sets of colors. And it's important to consider where the colors are in overall color space (neutrals vs. saturated colors) and if the number provided is one dE calculation or an average of more than one. The report is kind of important and few reports are better than what we can get in ColorThink Pro and BabelColor (the later is great visually):



Title: Re: High Delta E after monitor calibration
Post by: Doug Gray on January 06, 2018, 03:59:44 pm
It seems that more recently folks have been using 1 dE2000 as a JND.  dE2000 is in general less sensitive than plain old dE(76) - that is it takes more dE's to make 1 dE2000.

Jack

Hi Jack,

Yes. Generally dE00 is smaller than dE76 except for the critical neutrals where dE00 can exceed dE76 by as much as 50% but only in the hue shift critical yellows and violets.

OTOH, outside of neutrals and very near neutrals, dE00 is almost allways a fraction of dE76. Take the two greens L*a*b*: (50,-65,0) and (50,-70,0). By inspection these have a dE76 of 5.0 but the dE00, which is perceptually better (and computationally messy), is the barely noticeable 1.25.

I explored a bit about how dE00 responds to saturation and hue here:
http://forum.luminous-landscape.com/index.php?topic=111560.0
Title: Re: High Delta E after monitor calibration
Post by: The View on January 06, 2018, 06:36:39 pm
This is when you start typing or just hitting the keyboard, right?

You might just have a USB connection issue if at any point you are utilizing that type of connection during and after profiling. I'm not familiar with how your display setup is but just as it has been suggested, test by hooking up another monitor (cheapo or old used one) and don't even calibrate it or hook up any device through the USB or any other connection and see if you get this increase/decrease in brightness associated with keyboard activity.

If it doesn't do the brightness fluctuation, then it may be the internal electronics of the NEC display at issue or near or at failure. If it does with the test display, then it's most likely with your internal electronics of your device connection pathway be it USB, Firewire or whatever is used during NEC connects and profiling routines.

No, I discovered that it's not connected to typing. That was a chance correlation.
Title: Re: High Delta E after monitor calibration
Post by: Tim Lookingbill on January 06, 2018, 10:58:17 pm
No, I discovered that it's not connected to typing. That was a chance correlation.

This is what I was referring to as well which hasn't been addressed here by other contributors...

Quote
I just did another calibration and it totally failed (even though spectraview OK'd it). I can see it the whites are too blue.

In three instances I had noticed that after a mouseclick the display got brighter...

I reduced the intensity to 130 cd/m2 and got the Delta e down to 6.71, average 3.7

I'll see how daytime editing is with this or if I can turn down the intensity even more.

It just did that thing again where the display decreased intensity in a step just like it does when profiling it...

This is highly unusual behavior between display and computer circuitry. My suggestion since it doesn't appear you're going to troubleshoot this with the spare display, is to tell this to an NEC support technician to see if they've encountered similar behavior with other NEC owners.

Since you mentioned you are making money as a photographer, time is money and the trouble you describe in this thread is so rare that troubleshooting may send you down a long rabbit hole that might wind up being total waste of time with no solution.

Or just buy you new NEC replacement. They're only around $1200.
Title: Re: High Delta E after monitor calibration
Post by: Garnick on January 10, 2018, 08:12:46 am
That self Color Correction - is it part of Spectraview II?

I don't think you need a cleaning app on a mac - you just pull the app into the trash.

"I don't think you need a cleaning app on a mac - you just pull the app into the trash".  I realize this is not in direct reference to this thread, but a rather disturbing notion nevertheless I believe.  How often have you trashed an app by simply "Trashing" it?  When you have some spare time, check your Libraries, both Macintosh HD(apps etc) and User.  Go to preferences in each case.  I think you will find an abundance of prefs(.plist etc.) from the apps you have supposedly trashed, still residing there.  If that is the case you should definitely Trash those prefs, since they can wreak havoc when you are reinstalling an app, for whatever reason.  Simply trashing an app generally does exactly that, but it does not trash the supporting prefs etc.  There are a number of app "cleaners"(trashers) available, some of which have been mentioned previously.  For many years I've been using "App Zapper", and have never found any prefs left behind that should not be there.  Just a thought.

In reference to the main topic, I've been using PA series NEC displays for many years, along with the Spectraview calibrating & profiling software.  I originally used the Spyder hardware, but have recently opted for the i1 Display Pro.  My main working display(PA27W...) is now 5+ years old and is still clicking along as expected, with many hours of use.  I normally calibrate to a luminance of 100 cd/m2 and produce a dE of .30.  I also have the screen saver kick in after 10 minutes, and if I'm leaving the computer for a short time I initiate the screen saver manually.  I generally produce two calibrations for different situations, one at 6500K and my main working cal at 5500K, and all at Native Contrast Ratio.  On my most recent cal I also did a 5500K with a contrast ratio of 350:1.  That cal produced a dE of .14.  The NEC seems to be cruising along as expected and I'm sure I'll be able to get a few more years out of it.   

Gary