Eizo CS2740 Correction Profile

[Cameron Oatley]

Hello!

I recently got a new Eizo CS2740, and have the Calibrite DisplayPro HL colorimeter. However I’m uncertain which spectral correction I should be using and there doesn’t seem to be any in the database for this monitor.

Does anyone have a spectral correction appropriate for this monitor or guidance on which of the included ones I should be using?

Calibrite Display Pro HL on Amazon  
Disclosure: As an Amazon Associate I earn from qualifying purchases.

[Vincent]

It’s a widegamut W-LED PFS, AdobeRGB green primary flavor. Same as CS2731

Where can I find CS2731 EDR file?

Remember that ColorNavigator needs an EDR upgrade too. Otherwise it will use a generic matrix correction (non portable with i1d3) or GB-LED correction (not suitable for newer Eizos)
So you need to overwrite RG_phosphor edr in C:\ProgramData\EIZO\Sensor\i1DisplayProEDR (or equivalent forlder in /library/ app support for macOS). Then on Color Navigator before measurements choose “no compensation” (use EDR) instead of “color managed” (use generic matrix).

[Ben]

Thanks for asking.   Community needs to support developers.  Argyii is the heart of the Displaycal.   Not so funny how open source is made from someone and then gets a commercial job and income.   Paying open source helps.   Spectrometers are higher priced than colorimeter but make the correction.   People need to share the correction they make.   There is some at AVS forum.   I am dieing to try Colorspace Zro but my budget keeps going over from sickness.  They make there own correction with expensive equipement.   Free advice is going out of style except for AI.

I think the profile it needs is a “EIZO ColorEdge CS2731 (W-LED PFS) – Colorimetry Research CR-300 2nm”  .    A little clearer professional than the page at EIZO GmbH Display Technologies CS2731  .     I do not which is right.  “TECHNOLOGY “LCD PFS Phosphor IPS” vs w-led pfs from avsforum. The generic one from the meter software you can use a tool to get calibrations with has w-led pfs.  I would try that.   Maybe go to https://colorimetercorrections.displaycal.net/   and search EZIO .

Hope my meds are not talking.

[Ben]

Thanks Vincent.   I was researching while you was answering.

[Cameron Oatley]

I’m using displaycal (outputting through Resolve and a BlackMagic Decklink) rather than the eizo software as this way I can ensure everything works with my Resolve workflow and theoretically displaycal + 65x LUT should give more accurate results.
Can displayCal use these EDR files or do I need something different?

[Cameron Oatley]

Thanks, yeah I had seen stuff for the other Eizo monitors, but I cannot seem to find a clear answer as to whether those profiles can be used on the CS2740 or if it needs something different. I don’t want to be using the wrong profile and get erroneous results

[Vincent]

I’m using displaycal (outputting through Resolve and a BlackMagic Decklink) rather than the eizo software as this way I can ensure everything works with my Resolve workflow and theoretically displaycal + 65x LUT should give more accurate results.

You should use ColorNavigator to get desired whitepoint and brightness. Native gamut and whatever gamma you want (uually close to native -2.2- or close to target -2.4-)

Then create a LUT3D to transform from sourec profile (rec709 2.4 for example) to navte gamut (DisplayCAL profile). Better calibrate white in display than in LUT3D.

EDR replacement is needed to use CN7 for whitepoint.

Can displayCal use these EDR files or do I need something different?

EDR are binary files that store the same content as CCSS. EDR -> CCS is done by ArgyllCMS’ oeminst.exe.

Also as explained before, CS2740 is a WLED PFS with AdobeRGB green primary, plenty of CCSS samples.

• This reply was modified 1 month, 4 weeks ago by Vincent.
• This reply was modified 1 month, 4 weeks ago by Vincent.

[Cameron Oatley]

I’m getting whitepoint by adjusting the monitor itself, it has control to adjust RGB gain in 1/2000 increments (though it was extremely close to ideal out of the box, seems they do set those values at the factory). Brightness on the monitor was also set to 100cd/m2 and colorimeter was reading it as almost exactly that so I didn’t adjust further.

Using the “Spectral: LCD PFS Phosphor WLED IPS, 98% Adobe RGB/95% P3 (HP DreamColor Z24x G2)” correction in displaycal I’m currently getting these results (see attached), this is going through resolve, video level outputs, via a Decklink mini monitor 4k and using a display pro HL (I have a regular I1 Display Pro on the way as I’ve heard those are better at low levels)

Delta E gets down to about 0.3 which seems good, though run to run I’m sometimes seeing extremely high max delta E. I’m not 100% sure if this was just from not letting the monitor warm up enough or what but that seems odd. Sometimes it seems absolutely fine though?

Taking the ‘good’ result below (Delta E 0.5 and max 2.7) does that imply this spectral correction is OK? Or should I be doing something different?

Attachments:

You must be logged in to view attached files.

[Vincent]

I’m getting whitepoint by adjusting the monitor itself, it has control to adjust RGB gain in 1/2000 increments (though it was extremely close to ideal out of the box, seems they do set those values at the factory). Brightness on the monitor was also set to 100cd/m2 and colorimeter was reading it as almost exactly that so I didn’t adjust further.

Why don’t use ColorNavigator, targeting native gamut

Using the “Spectral: LCD PFS Phosphor WLED IPS, 98% Adobe RGB/95% P3 (HP DreamColor Z24x G2)” correction in displaycal I’m currently getting these results (see attached), this is going through resolve, video level outputs, via a Decklink mini monitor 4k and using a display pro HL (I have a regular I1 Display Pro on the way as I’ve heard those are better at low levels)

Why don’t use the ones made for those Eizos… althogh that one from default bundle should work too.

Delta E gets down to about 0.3 which seems good, though run to run I’m sometimes seeing extremely high max delta E.

where

I’m not 100% sure if this was just from not letting the monitor warm up enough or what but that seems odd. Sometimes it seems absolutely fine though?

Taking the ‘good’ result below (Delta E 0.5 and max 2.7)

verifying what? dispay vs native gamut profile or LUT3D + monitor vs Rec709? if the latter which gamma & black ouput settings

does that imply this spectral correction is OK? Or should I be doing something different?

No. You have no way to verify if spectral correction is ok unless
-spectrophotometer
-visual assesment of whitepoint

But CS2740 is a WLED PFS with native AdobeRGB green.

[Cameron Oatley]

Why don’t use ColorNavigator, targeting native gamut

Because I’m outputting via a BMD decklink using DaVinci resolve. So I need to use DisplayCal with Resolve output.

Why don’t use the ones made for those Eizos… althogh that one from default bundle should work too.

I can’t find any for this specific monitor in the DisplayCal/Community database

where

Seems this was just insufficient warmup, with an hour or so of warmup the issue went away

But CS2740 is a WLED PFS with native AdobeRGB green.

Does this mean this correction profile in DisplayCal is suitable or do I need to be using something else?

Spectral: LCD PFS Phosphor WLED IPS, 98% Adobe RGB/96% P3 (HP DreamColor Z24x G2)

The main issue I’m now getting is that despite using the DisplayCAL Resolve preset, which uses Tone Curve “As Measured”, after calibration it seems to lift blacks noticeably. Giving higher contrast or darker scenes a notably different look. I’m not sure what setting I’m doing wrong to cause that?

Attachments:

You must be logged in to view attached files.

[Ben]

Black Point Correction in Display Cal calibration if you have enabled a tone curve.    Or else the native hardware blacklevel is high.   What is the gamma and contrast ratio ?   Run verification and check simulation profile, check tone curve you use and black out offset as calibrated.    Black output offset is like blacklevel but effects shadows.   Posting a verification report will let us see the problem.

Black lifting considerable could be from black point compensation in display cal profileing.

Auto is the best way to me to do black point correction in the Display Cal calibration, but I do not do Display Cal calibration except for giggles and hopes.  I use HCFR to check the hardware OSD controls for calibration but for a TV set and not professional monitor.

[Cameron Oatley]

I’ve attached a verification report (though note I just ran the ColorNavigator calibration as per Vincent’s suggestion as I was preparing to do another calibration overnight so that may have slightly changed some things, I didn’t adjust monitor brightness or gamma though)

I didn’t have black point compensation enabled. Should this be enabled? It’s not enabled by default in the Resolve preset of DisplayCAL. I’ve attached screenshots of my settings just in case. But I’m also wondering if maybe the DisplayPro HL is just causing this issue?

I assumed since I’m not using an OLED and the monitor isn’t the highest contrast thing in the world it wouldn’t matter. But I don’t know if this is just me, the colorimeter, the spectral calibration, or some config issue

NOTE: These verification reports are with no correction applied, since as far as I’m aware you can’t stop resolve from disabling the monitoring LUT when running the calman calibration mode, so this shows the monitor ‘as is’, NOT with the issue I’m seeing post-calibration

• This reply was modified 1 month, 1 week ago by Cameron Oatley.

Attachments:

You must be logged in to view attached files.

[Vincent]

Why don’t use ColorNavigator, targeting native gamut

Because I’m outputting via a BMD decklink using DaVinci resolve. So I need to use DisplayCal with Resolve output.

Plug it to your computer (USB_C/DP/HDMI), calibrate WP & brightness & native gamut using CN7, then plug to decklink and create profile & LUT3D.

If white point does not match (using the same EDR/CCSS) then you have a problem in your deckling or Resolve configuration.

Why don’t use the ones made for those Eizos… althogh that one from default bundle should work too.

I can’t find any for this specific monitor in the DisplayCal/Community database

EDR is a binary for CCSS. If you have EDR, you have the CCSS: Argyllcms “oeminst” app (console)

where

Seems this was just insufficient warmup, with an hour or so of warmup the issue went away

But CS2740 is a WLED PFS with native AdobeRGB green.

Does this mean this correction profile in DisplayCal is suitable or do I need to be using something else?

Spectral: LCD PFS Phosphor WLED IPS, 98% Adobe RGB/96% P3 (HP DreamColor Z24x G2)

In practical terms is the same but blue spike may be a bit moved to shorten wavelengths. Since you need the CS2731/CS2740 EDR for ColorNavigator (my 1st quote) then is very easy to use the same file translated to CCSS (oeminst)

https://www.argyllcms.com/doc/oeminst.html

The main issue I’m now getting is that despite using the DisplayCAL Resolve preset, which uses Tone Curve “As Measured”, after calibration it seems to lift blacks noticeably. Giving higher contrast or darker scenes a notably different look. I’m not sure what setting I’m doing wrong to cause that?

Using a LUT3D for resolve has 2 steps:

1- Get an accurate description of display, capturing all its innacuracies and non perfect additivities (XYZLUT profile instead of matrix profile) with optional grey calibration.

2- Compute a LUT3D that reencodes some colorspace (rec709) intp monitor colorspace (the newly created XYZLUT profile).

Your problem seems likely related to 2nd.
A 1000:1 display tracking Rec1886 lifts blacks because rec1886 is NOT 2.4 in a “low contrast display” (anything less than 2000 or 3000:1 is low contrast). This is Rec1886 BY DEFINITION. So on a low contrast display you simulate (LUT3D) or calibrate (ColorNavigator) to 2.4 gamma (not Rec1886) with black output 100% and “assume” that in an ideal colorspace with “infinite contrast” like Rec709 gamma 2.4 there will be “out of gamut colors” near black and you will try to “map” them to the low contrast colorspace in monitor by mapping Rec709 RGB 000 (0nit) directly to Display RGB 000 (0.1 nit @100nit white), so the few RGB 1, 2,3 will be compresed a little.

Of course you may want gamam 2.2 for other purposes. no problem. 1st step stays the same, you change the target in 2nd.

Of couse an HTML report shows all these issues to it’s easy to find the culprit and the solution. Use them.

PS: Same if you used Calman or HCFR for verification. Rec1886 BY DEFINITION depends on contrast (black level) and it’s very very very unlikely that you want Rec1886 on a display that ones calibrated will be between 800:1 and 1000:1.

PS: ColorEdges CS should have near perfect “color tint uniformity” (delta C) with a mild brightness drop on corners for a 100% white field. Do not enable uniformity compensation, assume that drop in brightnss on corners, you do not have as much contrast  as the CG-X models to spend on uniformity, and without uniformity compensation CS models have no color tint.

[Vincent]

I’ve attached a verification report (though note I just ran the ColorNavigator calibration as per Vincent’s suggestion as I was preparing to do another calibration overnight so that may have slightly changed some things, I didn’t adjust monitor brightness or gamma though)

I didn’t have black point compensation enabled. Should this be enabled? It’s not enabled by default in the Resolve preset of DisplayCAL. I’ve attached screenshots of my settings just in case. But I’m also wondering if maybe the DisplayPro HL is just causing this issue?

I assumed since I’m not using an OLED and the monitor isn’t the highest contrast thing in the world it wouldn’t matter. But I don’t know if this is just me, the colorimeter, the spectral calibration, or some config issue

I was answering you as you were writing.

Short answer: DO NOT USE REC1886. It s not what you think it is. Put directly gamma 2.4 black output 100% without any mention to Rec1886.

[Vincent]

Report shows lifted blacks because…. it is tracking Rec1886 on a 700:1 display. Google Rec1886 spec. Good for OELD and VA TVs, out of consideration on “low contrast” IPS.

[Author]

Posts