Unsolved
This post is more than 5 years old
47 Posts
0
173778
October 27th, 2014 16:00
U2413, Real gamut coverage
Hello,
Although I am still fighting to obtain a correct U2413, I am calibrating it to check the coverage. I find that calibrating in native mode with DCCS results in a quite smaller space than AdobeRGB (in greens) as seen visually. If I calibrate -then in GPU- (dispcalGUI) in standard mode the best I get is 93-94% (50% brightness) of adobeRGB (value given by dispcalGUI) instead of the 99% of adobeRGB as it should, according to DELL specs. I follow directions here and here (specially the comments too). Could it be another fault of these monitors? Has someone found this too? Another thing, I am becoming crazy trying to find it: HOW can I know/calculate the gamut coverage of a profile vs other (for example related to AdobeRGB)? The only way I know is calibrating (by GPU, so not internal LUT) with dispcalGUI, (it shows a value of coverage when it finishes), but if I have another profile (i.e. obtained in i1profile or DCCS), then how does people calculate the coverage? Which software gives that value?
Thanks!
0 events found
Darkbluesky
47 Posts
0
November 15th, 2014 06:00
I don't have Dell Display Manager installed. I uninstalled it some weeks ago....
yumichan
4 Apprentice
•
739 Posts
0
November 15th, 2014 07:00
Weird behaviour...
Plug it to another computer and if you can with another OS (live DVD or something similar), select manually sRGB/AdobeRGB/Standard OSD mode and with a non color managed browser check "contrast ramp" in previous link.
green & cyan should look under saturated in sRGB
red & magenta shoudl look under saturated in Adobergb and sRGB
Darkbluesky
47 Posts
0
November 15th, 2014 13:00
I have tested in the same computer booting from cd with Ubuntu 14.01, I tested the lagom.nl tests with firefox (as it is installed in Ubuntu, I don't think it has the color correction plugin), but I downloaded also Opera, which AFAIK, is not color managed. In both browsers the lagom.nl tests appear exactly the same between aRGB and sRGB.
Tomorrow morning I'll do the same test in another computer, just in case.
yumichan
4 Apprentice
•
739 Posts
0
November 16th, 2014 04:00
If it doesn't work... return it and try to get a new one from Dell online shop.
Darkbluesky
47 Posts
0
November 16th, 2014 05:00
I have done today the same test with ubuntu, but now in the other pc (8600gts), as well, with firefox and opera. They both, the aRGB and sRGB are exactly the same...
On that Pc I calibrated Custom @6500 2.2 120. and getting 97.6% aRGB coverage for 6500K, and 98.2% aRGB coverage if WP results in 6430K
Now I am back to my PC, and trying to do a calibration with DCCS, just to compare with DCGUI.
This friday DELL sent me a replacement monitor that should arrive in monday, and I have to return it (or other one) tuesday. Will see to do tests monday afternoon.
Darkbluesky
47 Posts
0
November 16th, 2014 08:00
BTW, I have just calibrated in my PC, with DCCS using xy coordinates for primaries and WP. The coordinates I took the ones from the previous DCGUI calibration in Custom, that gave me WP of ~6400 and 98% coverture aRGB. Now after calibration with DCCS (automatic) and DCGUI profile only, I got 95.4% of coverage.
Why is it like that? Is that normal? It seems too much difference for me.
yumichan
4 Apprentice
•
739 Posts
0
November 16th, 2014 09:00
It isn't such difference
In first case you are computing intersection in cubic units of a bigger volume against AdobeRGB. There is a 2% (green-cian) of AdobeRGB that falls outside intersection.
The second case is based on a computed coordinates you set as "valid" after LUT3D calibration, maybe you are limiting native gamut which results in a smaller intersection. That's the main cause these % numbers do not have the utility that lots of users give to them.
If you truly want to see how it performs when displaying AdobeRGB images in a color managed enviroment, just dispcalGUI test a simulation profile and check for dE2000 outlayers (non "dark" outlayers)
Anyway, unless working on eciRGB2 is not very wise to hardware calibrate to native gamut, such a waste of reds on a 8bit workflow for sRGB/AdobeRGB images... If this unit performs well as it seems (if no taking account sRGB boken preset), just set in DCCS "AdobeRGB" preset and let DCCS do the process.
yumichan
4 Apprentice
•
739 Posts
0
November 16th, 2014 11:00
As said before, validate profile against AdobeRGB to see where and how severe are the issues you see.. in dE2000, not %, (and not very dark ones due to device measurement limits)
Darkbluesky
47 Posts
0
November 16th, 2014 11:00
I think I understand more or less what you say.
In order to avoid the 'limitation' I may introduced by putting the xy coordinates I did two calibrations:
Just to test, I have done a DCCS calibration in AdobeRGB and 120 cd/m2, then I did a profile only in DCGUI, the coverage is 95.6%.
Then I repeated in DCCS with Native (although I understand what you say about reds) and 120 cd/m2. Coverage in DCGUI, 95.8%.
Likely I am missing something here (?), but I don't see what can be...
yumichan
4 Apprentice
•
739 Posts
0
November 16th, 2014 13:00
The only issue is not daylight white (3dE towards cyan). There were similar issues with DCCS raising brightness after writing LUT3D
Gamut coverage and AdobeRGB simulation are very good.
Darkbluesky
47 Posts
0
November 16th, 2014 13:00
Well, not sure about it, but in general it does not seem too good to me, but I don't know.
[View:/cfs-file/__key/communityserver-discussions-components-files/3529/Measurement-Report-2_5F00_5_5F00_0_5F00_0-_1420_-DELL-U2413-DCCS-AdobeRGB-vs-AdobeRGB-_2800_sim-DCGUI_2900_.zip:550:0]
These are the settings to do the simulation:
Darkbluesky
47 Posts
0
November 16th, 2014 14:00
Ah, so the criteria for coverage and aRGB simulation would be the DE*00 which has to be as low as possible (<3?) for primaries (in my case max 1.33)?
The gamma and gray balance graphics are ok too? I thought they where somewhat not too ok (?)...
On the other hand I am a bit confused as I expected to see in the simulation report, for example, for green primary (RGB 0 255 0), the xy green primary coordinates of AdobeRGB in the first set of columns at the left (and the ones measured, in the last set of columns at the right), but instead I see 0.2302 and 0.7016. (I think that adobeRGB xy coordinates for green primary are 0.21 and 0.71) (?)
If it is ok, what a pity the presets problems...
Tomorrow I should receive the DELL one. I'll see and report.
yumichan
4 Apprentice
•
739 Posts
0
November 16th, 2014 15:00
Coords are in PCS coords, CMS colorspace centered in D50.
L*=100, a*=0,b*=0 is D50. To see measured values in xyY which is not "percetually uniform" (that means distances in xy has no real "visual" equivalent , dinstances must be taklen in L*a*b* dE2000) use "absolute values". D65 in this report is not L100a0b0, but a=-1, b=-20 which is refered to D50 centered coordinate system.
Its important to notice the difference between "PCS world" and "measured world", between xyY distance, which not matters and dE2000 wgic is visual distance.
Validation is done uin PCS coords because PCS is what GIMP or photoshop use to color manage, this is where color inaccuracies will arise. Then color managed coords are transformed back to monitor with CHromatic Adaptation.
Darkbluesky
47 Posts
0
November 17th, 2014 14:00
Thank you, I understand it now.
I calculated the intersection area of these both 'triangles' (AdobeRGB and the profile), and I found 96.2% of coverage (it was 95.x when profiling only in DCGUI) for the DCCS profile, and 98.2% for the DCGUI one. So I guess that the coverage is around that (~96) although it is not perceptible (DE<3)...
On the other hand I have received the DELL replacement. This monitor is also manufactured in Sept14, but this time it HAS a difference between presets aRGB and sRGB, sRGB is clearly less saturated (specially in reds, than the aRGB preset). This monitor seems of the same 'family' than the first I had form Amazon: they both do NOT have a green dominant.
I have made some calibrations with it (in the other PC, which is inline with mine, but gives a little more coverage -just like my recent fresh Win7 install in other partition in my PC- I'll have to do a clean install of my PC when I finish all that tests):
1. DCGUI calibration of Custom preset D6500 120 2.2:
When calibrating (D65) the monitors in DCGUI (custom preset), except in this one and the first I got from Amazon (manuf. date: May 2014), the green was dominant, and I ended with something like R100G92B98. But the first one and this one are quite neutral, and when calibrating this one, I ended with R99G97B100 (brightness 24, contrast 50), instead of the other typical 'set of parameters'.
[View:/cfs-file/__key/communityserver-discussions-components-files/3529/5367.Measurement-Report-2_5F00_5_5F00_0_5F00_0-_1420_-DELL-U2413-Sept14_2800_2_2900_-DCGUI-Custom-Cal_2600_Prof.zip:550:0]
(Continues...)
Darkbluesky
47 Posts
0
November 17th, 2014 14:00
I forgot that:
I did also an AdobeRGB simulation of the profile obtained, after Cal&Prof in DCGUI in Custom preset (6500, 120, 2.2). Here seems that the primaries are closer than in the DCCS AdobeRGB calibration (but maybe this can be improved using imposed xy coordinates?).
I mean, it seems that the monitor has the potential, and is more related to DCCS limitations (how I wish I could do DCCS calibration manually, and not in automatic, or that DCGUI or other soft could write the hardware LUT).
[View:/cfs-file/__key/communityserver-discussions-components-files/3529/Measurement-Report-2_5F00_5_5F00_0_5F00_0-SIMULATION-_1420_-DELL-U2413-Sept14_2800_2_2900_.zip:550:0]