Home › Forums › Help and Support › Direct-Load 3D LUTs for Eizo CG277
- This topic has 6 replies, 2 voices, and was last updated 6 years, 1 month ago by Vincent.
-
AuthorPosts
-
2018-03-14 at 23:09 #11066
Hi Just a quick quest as I am struggling to find the answer.
I am wondering if I can Direct-Load 3D LUT to an Eizo CG277?
I am running 2017 iMac 27″ and have the monitor connected via Black Magic UltraStudio Mini Monitor via HDMI as a reference, I am running Davinci Resolve.
I will be using an i1Pro 2 spectrophotometer and am aiming for the following target:
REC709 max luminance 80 and gamma to 2.2.
Cheers
Peter
2018-03-15 at 14:00 #11069AFAIK the only way to interact with these internal lut features is to use Eizo propietary libaries and AFAIK there is no “GNU-like” application that uses them.
But I think that Eizo’s ColorNavigation (CN) allows yo to load 3DLUT in “.3dl” format.
So calibrate your CG277 with CN to native gamut, D65 and 2.2 and store it in the CALx slot used with HDMI. This “should” modify CG277’s internal 1DLUT but keep LUT3D with output values with the same value as input (native gamut).
Then use your i1Pro2* and DisplayCAL to make a profile of the display without calibrating, just profile it.
Once you have such profile, use DisplayCAL 3dlut maker, use Rec709 2.2 as input profile and your new nativegamut profile as output. Select 3dl format and compute that LUT3D. Load it to your monitor with ColorNavigation.For a more accurate result, try to measure while profiling with DisplayCAL through Resolve output. It should be in DisplayCAL documentation.
*=too slow, consider i1DisplayPro colorimeter + CCSS or a faster device if you want a XYZLUT profile with a lot of patches.
Calibrite Display Pro HL on Amazon
Disclosure: As an Amazon Associate I earn from qualifying purchases.2018-03-15 at 22:58 #11085Thanks Vincent,
I’ll give it a go, sounds fairly complicated but your instructions seem very clear, I’m not the most tech savvy person out there! 🙂
Cheers
2018-03-17 at 7:48 #11101Hi Vincent,
I gave it a go and all went well and the reference monitor looks fairly neutral but is there now a way to test the monitors accuracy?
Cheers
Peter
2018-03-17 at 16:31 #11103Use DisplayCAL verification against simulation profile using that simulated profile as target: Rec.709 ITU-R BT.709, Gamma 2.2
Your *.3dl/*.cube LUT3D accuracy is dependent on the acuracy of the profile you used as input. An “idealized” CG277 calibrated D65 2.2 native gamut does not require a “complex” (big XYZLUT) profile of course, but ussually you want a very accurate profile for making a LUT3D because we have non ideal displays, nor idel devices for measurement.
IDNK what settings you used for that profile and LUT3D but i1Pro2 poor dark readings and slow measurement speed doesn’t seem a good starting point for this task. That’s why I said best to use a i1DisplayPro: keep your i1Pro2 for paper/textile readings and for generating CCSS (ArgyllCMS spectral corrections), for display measurements these Xrite graphics art spectrophotomerters are not up to the task.
So when you validate you may find some misbehaviours near black a*b* deviations, or your validation may end OK but visually you could see that a*b* deviations in lower end of grayscale. In that particular situation (only if that situation happens) you may want to validate CN calibration with DisplayCAL and/or inspect the profile you used for making LUT3D (TRC curves, with DisplayCAL profile info), looking for that kind of issues in both of them. Some may be caused by CN logic, some by profile type and complexity (size) you chose, or LUT3D creation parameters (for Rec1886 vs 2.4 gamma & black output offset for example), and some may be caused by your measurement device. That fine tune task is up to you…- This reply was modified 6 years, 1 month ago by Vincent.
2018-03-18 at 7:57 #11107Thanks for your reply once again, really appreciate it.
I actually have access to i1DisplayPro, I will go through the process again…
Just confirming a few things though….
Do I just repeat the whole process you mentioned even though a 3DLUT is already installed in the monitor?
Could you look at the steps below and let me know if all is well please?
- Calibrate CG277 with ColorNavigator NX to native gamut, D65 and 2.2 and store it in the CALx slot used with HDMI. (I do this without the BMD UltraStudio Mini Monitor, just a HDMI cable with a USB-C to HDMI converter hooked up to the iMac.
- Then use the i1Pro2 (will be using the i1DisplayPro) and DisplayCAL to make a profile of the display without calibrating, just profile it. (Connected via HDMI with BMD UltraStudio Mini Monitor and via Resolve)
- Use DisplayCAL 3dlut maker, use Rec709 2.2 as input profile and your new native gamut profile as output. Select 3dl format and compute that LUT3D. Load it to the monitor with ColorNavigation NX.
Thank you for your time and generous advice, I really appreciate it.
Cheers
Peter
2018-03-18 at 11:33 #11109First of all verify with DisplayCAL and your i1Pro2 your results. Also try some video sample with grayscale, to visually inspect if there is something weird.
If they are OK, you are done.If there is something that looks “wrong”, then try to locate the source:
-CN’s “visual” or measured a*b* deviations in dark greys after calibration, validating profile with DisplayCAL (you have several priority choices in CG’s HW calibration: contrast, greyscale…etc)
-Profile misbehaviour (TRC), the one you created with DisplayCAL to feed LUT3D creation
-low light i1pro2 lack of accuracy in these 2 previous checks
Once you spotted the possible source of your problems, redo your 3 steps (your last message) taking account what failed.But if you verify your results right now and they are OK, you are done. On previous message I just pointed out possible issues and how to proceed.
Edit: It is also possible that CN’s hardware calibration to Rec709 gamut and 2.2 gamma is good enough for your needs, and just with CN’s calibration to that target you wil be done.
But you asked for custom LUT3D loading… hence this thread.- This reply was modified 6 years, 1 month ago by Vincent. Reason: Alternative answer
-
AuthorPosts