- This topic has 23 replies, 3 voices, and was last updated 3 years, 4 months ago by
.
-
Posts
-
Hi,
I use the BenQ SW271 monitor. First of all, using internal firmware slots, I calibrated the monitor with Pallete Master Element to native color space 120nit, Gamma 2.2, blackpoint – Absolut Zero.
My final target is creating LUT REC.709 for Davinci Resolve. With this step, everything is clear. But first I need to check the correctness of the native calibration of the monitor at all.I am trying to verify the calibration by DisplayCAL, but I face the following difficulties:
1) which profile to choose for verification? Obviously, I can take the maximum possible Adobe RGB color space for this. Or should I choose something else when verifying native color space?
2) When I went to the “Verification” tab, I found that the check button is not active. I couldn’t activate it no matter what settings I choose.Before I go into calibration and creating a LUT for Davinci Resolve, I want to make sure that the monitor is properly calibrated. Help me deal with this problem. Thank you!
- the one made by PME. Check white “whiteness” dE aims for a* errors, assumed CCT for b* deviation from your target. Check also RGB+gray balance, range value <1 against true neutral grey (if profile reference has no neutral grey because a table 3TRC curves profile, choose “use gray calibration…” checkbox on top of resulting HTML report.
- I think that happens when 1st run and no DisplayCAL profiles are installed, hence I do not remember well. Try to install PME’s profile through DisplayCAL (file menu) or choose “current” settings preset on to combo box. That should enable it.
2) When I went to the “Verification” tab, I found that the check button is not active. I couldn’t activate it no matter what settings I choose.
Did you install Argyll CMS ? Just go to Tools – Port and install from there
- I think that happens when 1st run and no DisplayCAL profiles are installed, hence I do not remember well. Try to install PME’s profile through DisplayCAL (file menu) or choose “current” settings preset on to combo box. That should enable it.
Thank you for your answer, Vincent!
I found no other way to activate the calibration check button than to make a new quick calibration. We can ignore it. Only then did the button become active.
I enclosed the measurement screenshots. I also had questions, despite the fact that I read a lot of information in this forum, but understood little:
1) what type for my monitor should I choose? RG Phoshor or LED RGB? (I chose the first one)
2) whether I specified the measurement parameters correctly if the PME had the following: “native color space 120nit, Gamma 2.2, blackpoint – Absolut Zero”?Thank you!
2) When I went to the “Verification” tab, I found that the check button is not active. I couldn’t activate it no matter what settings I choose.
Did you install Argyll CMS ? Just go to Tools – Port and install from there
Sure. CMS was offered for download immediately after installing DisplayCAL.
- the one made by PME….
But with these check settings the number of contrast sharply increased…
Correction is unknown, check community corrections plot them, choose closest.
Simulation is to check how it behaves in color managed apps. Grey range is pretty bad as expected from a Benq, low cost low quality monitors. Panels are nor quite neutral in grey ramp, thus a lot of measurements are neede to calibrate grey. PME does not do that. DisplayCAL can do it in medium/slow calibration speed, but it will be GPU calibration. Uless you have a GPU with high bitdepth LUTs and dithering (mostly AMDs) banding is expected (although there are some registry tricks in windows for nvidia+dithering, google it).
Bad grey range means that some greys are pink , others are green. Altough each one may be close to its reference, seen all these greys in a phoro may reveal grey tints which are ugly.
Low cost poor quality monitors like Benq SW are expected to behave this bad. Return for refund and get an Eizo CS2731 (QHD) or CS2740 (UHD) or try to calibrate in GPU on top of your HW calibration just to fix grey.Also color uniformity of Benq SW are between the worst quality widegamut displays out there (SW271, Photography Life review):
So check it and plan to return for refund if it is bad and/or you do not want to try GPU calibration to fix grey.If you disable simulation profile just validates if profile (PME) and display match. Your last screenshoot. Range remains as bad as expected. More than 1.5 is easily spoted on a non color managed grey gradient like lagom lcd test. Use MS Paint for testing.
MadVR/Resolve LUT3D may be able to solve taht kind of grey issues as long as XYZLUT profile used to make them was ver very detailed (so it can capture each of thses deviations from grey to colored grey).
If you want to keed monitor and you have an AMD GPU card, IMHO it’s easier to make a GPU calibration (calibration speed medium or slow, and matrix profile for general use on desktop). Then make detailed XYZLUT profile for Resolve/madVR but keep that GPU calibration (calibration tab = all to native). It will as to keep current calibration curves (the ones loaded into GPU because of DisplayCAL calibration). Then when making LUT3D do not apply VCGT.
Desktop gets grey calibrated on all apps this way. This ont the only way, just a hint for a veratile solution. You can choose others.So check it and plan to return for refund if it is bad and/or you do not want to try GPU calibration to fix grey.
Unfortunately, it is too late to return the monitor, I will work with what I have.
If I have understood you correctly, in my situation I need to get the most correct gray color by creating an ICC DisplayCAL profile that will work over calibration with LUT PME?I have Mac Pro 2013, dual AMD FirePro D700 6 GB, so there is no registry manipulation here.
In order (if not correct – correct me):
1. I do primary calibration in PME “native color space 120nit, Gamma 2.2, blackpoint – Absolut Zero”.
2. Then, using the built-in LUT PME, I calibrate to DisplayCAL and create an ICC which I use to replace the ICC PME.
3. I calibrate with the following settings (check my screenshots with the settings)
4. Then, having substituted ICC PME for ICC DisplayCAL, I proceed to calibration of Davinci Resolve in REC.709. I leave the same settings in the first tab of the monitor and colorimeter. In the calibration tab I leave all the settings “as measured” (you said “all native”, I understood that); the type of profiling is on XYZ LUT, the rest is by default. The 3D LUT tab is the default on Rec709/Rec. 1886 with 2.4 gamma but without vcgt.Correct me if something is wrong. Thank you in advance!
Step 2 only if grey color issues bothers you on desktop, Photoshop, Indesign or any other software.
DisplayCAL LUT3D creator as long as target profile is very detailed (fine grain , a lot of patches) and covers all these issues, should be abe to correct it.Let me explain. PME HW calibration show non corrected grey issues because of Benq poor software and native behavior of poor QC’d panel used by Benq. This is what you show us in #26830.
There are 2 ways to correct it… if it bothers you. (IMHO 1.5 range should be very visible but YMMV):
option a) correct in GPU with a common DisplayCAL GPU calibration on top of PME. You want to keep this GPU calibration when you make Resolve XYZLUT ICC that will be used for LUT3D generation. Since grey GPU calibration is applied at desktop level, you should not include VCGT data into LUT3D
option b) live with current (bad) PME results in desktop, photoshop, Indesign or any other software. Grey issues will be present, grey coloration and all those issues that your report shows.
Then make a very detailed profile, with lot of patches so it can capture all issues on screen after PME. Then make a LUT3D with that profile. You can include VCGT data here because it is not applied at whole desktop. Resolve will be corrected by LUT3D, Photoshop/GIMP won’t.Choose a or b depending on your goals/easy of use.
b) is easier, just follow DisplayCAL Resolve wiki step by step. Resolve output should be corrected as long as ICC profile used to make LUT3D was so detailed that those issues were measured.
a) allows you to fix grey issues for Photoshop or other appications, but you should be more careful to keep that GPU calibration on Resolve profile and do not apply it twice (one on GPU LUT, another one in software LUT3D for Resolve).
YourAMD GPUs in Imacpro should dither by default (that’s the cool thing about MAD cards) so no banding, 10bit end to end, etc.Then make a very detailed profile, with lot of patches so it can capture all issues on screen after PME. Then make a LUT3D with that profile. You can include VCGT data here because it is not applied at whole desktop. Resolve will be corrected by LUT3D, Photoshop/GIMP won’t.
This method seems to me logical and more intuitive. If I understand the sequence of actions correctly, as I wrote above, I need to calibrate the monitor using PME in the native color space and then recalibrate with DisplayCAL for using the new ICC over the LUT PME installed in the monitor. Should I use the generated ICC profile to generate 3DLUT for Davinci Resolve and I don’t need to make additional measurements via the Davinci Resolve GUI viewer? Just generate it from the correct new desktop-ICC? So in 3DLUT I narrow it down to REC709 and my measurements should be enough? I have a weak idea of how to make it in DisplayCAL.
One more question. When recalibrating a desktop with what settings can I calibrate as much color space as possible to make the most of PME’s Calibration in Native?
Then make a very detailed profile, with lot of patches so it can capture all issues on screen after PME. Then make a LUT3D with that profile. You can include VCGT data here because it is not applied at whole desktop. Resolve will be corrected by LUT3D, Photoshop/GIMP won’t.
This method seems to me logical and more intuitive.
That is B)
If I understand the sequence of actions correctly, as I wrote above, I need to calibrate the monitor using PME in the native color space and then recalibrate with DisplayCAL for using the new ICC over the LUT PME installed in the monitor.
That is A)
Choose one.
One more question. When recalibrating a desktop with what settings can I calibrate as much color space as possible to make the most of PME’s Calibration in Native?
GPU recalibration with DisplayCAL on top of PME just modifies grey & white. Same for i1Profiler or other GPU calibration apps.
GPU calibration = correct grey ramp = VCGT LUTs in GPU. That’s all. After that a ICC profile is made with monitor behavior after calibration.b) is easier, just follow DisplayCAL Resolve wiki step by step. Resolve output should be corrected as long as ICC profile used to make LUT3D was so detailed that those issues were measured.
I decided to follow method “b” using the wiki instruction. After calibrating the PME, I proceed to carefully calibrate the desktop with DisplayCAL, creating ICC and 3DLUT for Davinci Resolve. Hopefully the number of patches will be enough for that). Please check the screenshots, any errors?
-steps in wiki = OK. IDNK which is spectral power distribution uses SW271, check colorimeter database.
-last screenshot => MacOS color management engine is extremely poor, do not use that ICC profile as “display profile” in OS settings. MacOS can only work properly with simplified profiles = matrix, 1 TRC, ideal contrast. PME profile should fall in this category.
Just make sure that PME profile remains as default in OS settings.
DisplayCAL XYZLUT is just for generating an accurate LUT3D from PME HW calibration.CCSS BenQ SW271 (ColorMunki) 3.3nm, 380-730nm (2019-07-26 15:34:59)
Looks you are right, it’s a GB-LED: the one you choose, U2413 correction bundled with DisplayCAL or that one from community database.
PME does not support it, it uses RGBLED for all SW series (wrong), same for WLED PFS or QLED, hence WP & primaries from PME always will have an error, but actual error depends on YOUR colorimeter sensivities stored in firmware. The further colorimeter drifts from std observer where actual SPD and RGBLED SPD drift from each other the bigger the error will be. If they are close on red wavelengths error will be low.
Cannot be predicted without dumping colorimeter firmware data.
