Will Dell Calibration Software improve the calibration accuracy in later release?(U2713H related)

Yumichan, Thank you so much for your exceptionally comprehensive and helpful response.

I tried your method(force DCCS use RG_phosphor correction)

The result is greatly improved, compare to my calibrated Apple Macbook pro retina and Dell 2312HM.

The whitepoint and gamma still a bit off but I can live with that.(for now)

But still one problem hasn't really improved through this method.

The color still has very prominent green tint. It's Sooooooo GREEN.

I can get rid of the green in custom color preset. (which is not a very usable preset/gamut to me)

But not in sRGB and Adobe RGB preset.

I still hoping Dell can help us here.

I am aware I am not buying an Eizo or NEC but I have other cheaper Dell monitors like U2312HM and 2407WFP-HC which don't have  this prominent green tint problem(2312HM has blue tint and 2407wfp-hc has red and green tint but no where as prominent as U2713H)

Why this marketed Dell's high end attempt has?

Anyway, Thanks again Yumichan, you saved my day!

So the green tint is ONLY in OSD's sRGB and AdobeRGB presets, not in CAL1 or CAL2. Am I correct?

Which program do you use to soft-profile? Basiccolor won't work here either (same GB-LED backlight).

Try this:

ArgylCMS. It's a console(DOS-like) program for soft profile:

http://www.argyllcms.com/

An this is a graphical interface for it:

http://dispcalgui.hoech.net/

All are GNU. I asume that you have "Custom mode" soft profiled, with some value of Contrast and Brightness. Keeping these values, use Dell Display Manaer (or OSD) and switch to sRGB/AdobeRGB emulation modes.

1.-Open Dispcalgui, it will ask fro "bin" Argyll's folder.

2.-In "Tools"->"Import colorimetes corrections" (or somethng like tat depending on your language) import the RG_phosphor.edr. It will show as "Addition of Dell ultrasharp U2413 .." (GB-LED)

3.-Set these values:

*WP=6500,

*white &black level native (so you can keep brghtness from "custom mode" in Dell Dsiplay manager. CAL1 Y CAL2 have their own values but I'm not sure you can have diferent values for Custom and the emulation modes.

*gamma=2.2 absolute ("less agresive" correction from native gamma, "less banding" after, specially if you use a gamer nvidia)

*Black point correction, Auto=UNCHECKED, 0%, move the other value form 4 to 2% (less banding, less black correction too from native values)

*Calibration quality high

*profile type = "XYZ LUT + matrix", and in combo below check default preset for LUT (124 or 238 patches)

4.- click calibrate & profile. It will show a window to fix RGB gain but because yo cannot do anything here, just ckick next.

5.- Within 30minutes it will make a profile. Before installing you can see the gamut (not very coordinate-acurate this picture).

6.-Check banding with this URL (Ony IE ir Opera, not firefox, you want a browser that does not color manage):

www.lagom.nl/.../gradient.php

If it has a lot of coloured bands, do less black correction, or lower gamma jus a tiny amont close to preset's native (Tools->Un/calibrated screen report, I don't remember which)

You can check color acuracy from DispcalUI from "tools"->"verify profile", choose "verify_extended.ti1". The results will be in a HTML file (please enable Javascript).

If you made CAL1 or CAL2 profiles in ICC V2 version (sorry, I forgot that) you can also check color acuracy with DispcalUI and GB-LED EDR corrections. Just remeber to do not load any kind of ".cal1" files from your previous soft profiles in the upper combo box. It should show "Actual" or something like that.

With my U2413 it works, but Adobe WP more or less OK and sRGB was off only a little: 6600-6700. I can keep about

95% RGB values for Adobe an 93%-96% for sRGB. A little banding apears in that URL for grey gradient check. ATI GPU's LUT seems to work better than my nvidia.

As you say sRGB/AdobeRGB OSD presets inacuracies are Dell's fault (I acept that Standard or Custom comes uncalibrated).

I hope this will work for you to properly soft-profile these OSD presets.

yumichan, having found this thread, I have tried to follow your instructions for creating the missing XML file for my U2713. But I am not quite sure this has worked entirely as intended. Checking the logger.txt file again after calibrating with the XML file in place, shows two things: 1) the file is now much longer and now includes a detailed calibration report detailing RGB targest and Yxy values, which was not there previously, while the error message about not being able to parse the file is gone - so something definiely has changed. But 2) the beginning of the log file now reads:

[22.08.2013 21:42:35.927]  INFO  DELF047
[22.08.2013 21:42:35.927]  INFO  Generic Type 18
[22.08.2013 21:42:35.927]  INFO  DELA092
[22.08.2013 21:42:35.927]  INFO  Generic Type 18
[22.08.2013 21:42:35.942]  INFO  Display Name: DELL U2713H ---- Hardware ID: DELA092 --- DisplayIndex : 0
[22.08.2013 21:42:35.942]  INFO  Monitor Name: DELL U2713H
[22.08.2013 21:42:35.942]  INFO  Monitor Type: 8
[22.08.2013 21:42:35.942]  INFO  Serial Number: C6F0K32F0HDL
[22.08.2013 21:43:49.231]  INFO  Test Puck on Screen

Note how in the second part 'monitor type' reverts to 8 again. I have done calibrations specifying 18 for RG Phosphor and 8  for Wide Gamut CCFL in displayTechTypes.xml. Comparing them using ICCView shows that the resulting profiles are practically identical - so I am suspencting that the specified 18 setting may somehow be overridden in the second part, when the hardware seems to be read in. (I have edited to Name="DELL U2713H" Comment="U2713H", assuming that the omission of the first 'H' in your instruction was likely to be a typo - correct?) If you are still reading this and have any thoughts, I'd be grateful.

I would echo you last sentence about the sad state of affairs. It is ridiculous that users should have to go to this level of technical detail to get their monitor and software to work. It should obviously be possible to simply rely on DCCS to deliver the right calibration (reset LUTs, switching out of other presets if native is specified, correct calibration matrix, and all).

yumichan, many thanks indeed! Like darkinners earlier in this thread, I would really like to thank you for your exceptionally helpful and competent response.

 My Logger.txt shows the same "Monitor Type: 8" (hardcoded log it seems,

Yes, but the question is what the software then actually does. Does it use the 8 or the 18, given the contraditory logs? After having done everything else (see below), I have once again created profiles specifying first 8 and then 18 in the DisplayTechTypes.xml file; and as best I can tell (including comparing their Dispcal profile verififacation reports) they seem to be extremely close. That would seem to suggest that either the difference is negligible or that DCCS overrules the 18 and and both calibrations in fact use the same .edr (8). Have you ever compared yours?

I have used your instructions and verified the resulting profile with both i1Profiler (quite limited reporting functionality actually) and with DispcalGUI (much more informative). The result looks excellent, in my case including the grey range. Thanks again!

I was interested in your point about the BenQ, as I had been wondering why Dell/X-rite had not customised the i1profiler, which would have been the obvious thing to do for this monitor and target market, instead of coupling hardware calibation with something as simplistic (and flawed) as DCCS.

all "prosumer" widegamut are lost in "badlands" (asus does not have HW cal). Save your cents for ~1500euro NEC

Which is a shame: it’s not perfect, but given the price point, the U2713H continues to strike me as a great offering. It’s mostly a strange case of the software letting down the hardware, when the latter should have been the greater challenge. The prosumer market should be nothing to sneeze at considering the potential appeal to all those Lightroom and Photoshop using advanced amateur photographers out there. It seems odd to first bother to make a really quite competitive monitor for prosumers and professionals on a budget and then to forget about the need for equivalent software support.

Mine seems better: 0.35 average ΔC*00, and 0.88 combined Δa*00 and Δb*00 range. But it is worth noting that that is for a native calibration with native white point, which should show the monitor at its best (as long as the white point is tolerable), and of course we have different, if similar, monitors. I haven't tried calibrating for sRGB or AdobeRGB yet, as I am using native for image editing, while the facdtory sRGB emulation is good enough for non-colour critical work.

@ArneH. I have not tested DCCS 1.0.1.0 without XML patch. Good to know that is working properly. Next calibration at October/Nov I will try.

How good is your gray range? ("RGB+gray range" on upper combo box in Argyll HTML report) Current values of my HWcal are 1.4 for sRGB and 1.67 for AdobeRGB. They are OK, but I think that should be better.

I have tried to do what you instructed with making an xml file but I keep receiving the same message:

[05.09.2013 23:44:35.891] ERROR MatrixXMLParser.cpp:81 : Failed to parse xml file C:\ProgramData//X-Rite\Dell UltraSharp Color Calibration Solution\displayTechTypes.xml
[05.09.2013 23:44:35.910] INFO Display Name: H484N_140GW01 ---- Hardware ID: INL0002 --- DisplayIndex : 0
[05.09.2013 23:44:35.910] INFO Display Name: DELL U2413 ---- Hardware ID: DELF048 --- DisplayIndex : 1
[05.09.2013 23:44:35.910] INFO Monitor Name: H484N_140GW01
[05.09.2013 23:44:35.910] INFO Monitor Type: 9
[05.09.2013 23:44:35.910] INFO Serial Number: 0
[05.09.2013 23:44:35.910] INFO Monitor Name: DELL U2413
[05.09.2013 23:44:35.910] INFO Monitor Type: 8
[05.09.2013 23:44:35.910] INFO Serial Number: FJMKT2CNARYL

 I have pasted your code into notepad and saved it as displayTechTypes.xml at the end, changing Model to F048 and Key to DELF048.  I saved it to C:\ProgramData\X-Rite\Dell UltraSharp Color Calibration Solution\.

  Am I doing this right? 

Yumichan: I see that this thread is a year old. if you have a Dell U2413, you should be using the DellUltraSharpCalibrationSolutionSetup.exe created Jan 27 2014.  I want to calibrate my 2413 LUT using this, but I feel that it is pointless to do this on a Mac under Mavericks using Fusion 7 Windows 7 Virtual Machine. Windows can do 30 bit, but a Mac OS and graphics card cannot output 30 bit. If the Xrite and Dell software can do 30 bit the monitor won't show it, making it a useless exercise.

Xrite and DCCS do not work in 10bit workflow. Thay write 14bit internal LUT of Dell GB-LED monitors. Even measured in 8bit workflow LUT3D will be interpolated in 14bit precision, then dithered to 8bit native+2bit dithered AH-IPS panel.

Program (8 or 16bit image)

goes to

GPU output (8 or 10bit, with between 8 and 12bit LUT if GPU calibration is applied)

goes to

Monitor input (8 or 10bit depending model and selected input)

goes to

LUT3D (14bit, here is hardware calibration, but DCCS oversimplifies is measured so it could store more acurate corecctions by far, READ THIS, DELL?)

goes to

Dithering electronics (temproral dithering 14bit -> 8bit native + 2)

goes to

physical panel conections of LG AH-IPS 10bit (8+2) GBLED.

10 bit wirkflow is meant to other things (image representation and less troubles due rounding numers in color management under Photoshop), not to calibration.

So you MUST install a Win7/Win8 in your mac with dual boot, not pararlells or VMware, run DCCS, get CAL1 and CAL2 calibrated to your desired target, copy 2 ICM profiles (make sure they are version 2 if you want to fully validate them) to your OSX partition (or an USB drive).
Unless you do this, if your mac has an nvidia GPU or a integrated intel GPU you WON'T BE ABLE to properly calibrate ANY monitor with traditional calibration (not internal calibration), no matter what company you choose: Dell, Asus, HP, Viewsonic, NEC (in Europe)... widegamut monitors without internal calibration cannot get properly calibrated unless you have an AMD/ATI GPU or a nvidia Quadro. Banding will be horrible with other GPUs.

Another option is to get a PC from a friend and calibrate your Dell's, then copy 2 ICMs to your MAC.

I see my thread got bump up.

I have a little update as well.

First, I was really fed up and ended up bought a Eizo CG277 and the U2713H became a monitor for entertainment(xbox,ps3 and stuffs like that)

Second, I recently decide to calibrate the U2713H, plug it in computer again, bam!

USB hub not working. 

Called Dell support, today they delivered a new/refurbished U2713H to me(full box)

plug it in, fire it up. do the calibration.

WOAH! It's much more acceptable now, ghosting/overshoot still here

But the green tint improved A LOT! still noticeable green tint but miles better than my old one.

And I can get very close to D65 with DCCS now, my old one always sitting somewhere 66-68k after calibration with DCCS.

The replacement U2713H is A07, my old one is A02. 

I don't know what Dell did in these revisions but they did something right, at least in the right direction.

Well, that CG277 is A FLAGSHIP MODEL, best 27" GBLED you can buy (2200 euro). Not sure if it is even comparable to U2713H.

For entertaiment purposes in your U2713H, best CAL1 or CAL2 in sRGB or sRGB/Rec709 primaries (custom) with 2.2/2.3 gamma. Unless you are playing "toon" games like Super Mario Bros, not wise to use widegamut modes.

First, you're saying that the hardware cal will not work through VMWare on a Mac but offer no reason why.

Second, you're saying that the DCCS (Dell Color Calibration Solution) sends an image to the computer which sends it to the monitor to be observed by the colorimeter. Because this starts in 8 bit and  involves several transformations and approximations, what good is it? Why send a 16 bit image, even a Windows machine can only output 10 bit?

The first reason was exposed in other theads regarding the same subject, it's your fault have not read it since it's ALL THE SAME PROBLEM reported again and again because PEOPLE DO NOT USE SEARCH in this forum to find an answer to a common topic. Since I do not work for Dell, I do this kind of support on my owm beacuse I want to, and aI not going to copy and paste all explanatiosn again and again.
BTW, the reason is that because whatever reason Xrite and Dell decided, DCCS checks serial number and models twice. First by DCC/CI on display output (that why it needs access to physical GPU not a vistualized) and on USB port used for comunication (LUT uploading and such things I suppose)

The second argument of yours is false.

Calibration is done by measurement of native display response on several points (colors) of its gamut, then interpolated, then corrected in a LUT (GPU for normal monitors or internal LUT3D these dells).
Becasuse of THAT, IT IS NOT NEEDED TO WORK IN 10BIT TO MEASURE NATIVE RESPONSE, since the points measures are only a FEW IN ANY MONITOR (NEC & Eizo Included), then interpolated. (128,128,128) native response in bative uncalibrated 8bit workflow will be tha same than (512,512,512) beacuse it is the same number with two 0 padding. SAME APPLIES to Basiccolor Display, for example (NEC Spectraview).
After calibration is performed, measure a few points, interpolate to get full response, then corrected & uploades) we need to profile the current state of the monitor. It is done in the same way: take a few measuremets, then interpolate and compute a profile that describes monitor's current state so color managed apps can deform images RGB values in order taht monitor response matches what theses numerbrs represent.
As in calibration, much of ot is interpolated and it is no so noticeable diference between 8 and 10bit which are padded values from 8bit.

TRUE 10bit advantages IS IN COLOR MAMAGEMENT. When Photoshop reads a 16bit sRGB photo and you have a NEC spectrview or a U2413 which are calibrated to a "AdobeRGB"-like gamut and gamma.

You do not loose a lot of green ramp (from black) in this sRGB image because color managment has 2 extra bits to compute the deformation of "RGB values of a SRGB image" to "RGB values of an AdobeRGB gamut".

THERE is when 10bit is needed, to avoid such "banding" introduced when you apply color managemet between mismatch profiles.

In this example even 10bit are hepful when displayed an AdobeRGB tagged image in this AdobeRGB-like gamuts because monitor profile takes acount of deviations from actual monitor response from its desired one.