- This topic has 89 replies, 5 voices, and was last updated 3 years, 1 month ago by
.
-
Posts
-
Thank you Vincent that is great info! I do not want to match apple retina, I want to match true colours and white balance, then I assume I should trust the measurement (and just see apple’s image as one of the many looks my clients may experience). Does the meaurement report (attached in my previous post) seem good to you?
It is not a report, it’s just profile summary, and TRC still looks as L* which is weird although Resolve LUT3D should undo it to your desired gamma (2.4? 2.2?).
You need to check if display matches that profile. Explained before.
I guess, based on your previous comment, that a correctly calibrated D65 Adobe RGB monitor should (have a slightly warmer white and) be a bit more more vivid/saturated than a out out the box Apple retina display?
No. Plugged into a macOS computer and with a profile that describes its actual behavior it should look the same (if we exclude the cooler white & greyscale of tupical macbooks).
Plugged into a Windows machine any widegamot calibrated to or using a wider colrospace than sRGB will look oversaturated ON DESKTOP. In color managed apps it should look the same (excluding whitepoint)
Thank you Vincent for your patience.
I did not know that colours are oversatureted on windows after calibration.
In photoshop the colours look good after calibration (native gamut hardware calibration 2.2 gamma and then Displaycal 2.2 calibration)
In Resolve the colours look oversaturated.
After creating Resolve 3d lut with displaycal, the colours still do not look correct.
Is the reason that I use clean out instead of having a Blackmagic card for outputting to the monitor?
Or is there any other settings I should try (I do not understand your tech words, please make it simple to understand) :). The aim is to follow your guideline to get 2.4 gamma in Resolve.
Thank you Vincent for your patience.
I did not know that colours are oversatureted on windows after calibration.
In photoshop the colours look good after calibration (native gamut hardware calibration 2.2 gamma and then Displaycal 2.2 calibration)
Windows desktop is not color managed. Raw RGB numbers on each icon or wallpaper or pixel in game are set as is. Photoshop is color managed.
In Resolve the colours look oversaturated.
After creating Resolve 3d lut with displaycal, the colours still do not look correct.
Read the FAQ you need to link LUT3D to each Resolve configuration.
Or is there any other settings I should try (I do not understand your tech words, please make it simple to understand) :). The aim is to follow your guideline to get 2.4 gamma in Resolve.
If this seems too complicated, calibrate in TCP to Rec709/sRGB, D65, gamma 2.4. Maybe LG TCO “rec709” preset is not what you want, there should be a custom mode, put CIE xY coordinates of sRGB/Rec709 primaries, set D65 and g2.4
This should render non color managed SDR video more or less accurate.Thank you. Where is the FAQ that describes the linking you describe (I find faqs but not describing this)? I am following all the guidelines how to create and apply 3dlut in resolve, but I can not apply a lut in resolve to the monitor since I use “clean out” to get full screen. I can however (while grading, then of course turn off) apply a lut to the timeline, but the lut created by Displaycal does not make sense in colours. Please give a how to description or a link to where this is described.
Since the colormanaged Photoshop got good colours with the already done calibrations I just want to get Resolve to work also 🙂
Since LUT3D is created skip to linking (“Using the 3D LUT in Resolve”)
Thank you Vincent, great link. I now calibrated with those settings and get almost 100% sRGB and Adobe gamma. I apply the lut, but when I go into Resolve and add the lut (temporarily) in the grading timeline (as mentioned before I do not know how I could apply the lut on the monitor in resolve since I use clean out to get full screen and the monitor lut does not support the clean out (I do not have a blackmagic hardware monitor card)) then it seems that the resolve lut is applied on top of the calibration of the screen, so it goes to far and gets very low saturation. Any suggestion on what I should do instead?
In Resolve forum: “In DisplayCAL Profile loader (in the taskbar) you must define an exception for Resolve with Reset video card gamma table option so that calibration gets reset when Resolve is running since calibration is already written into the LUT.”
My question:: I can not understand where exactly I will find the profile loader, can you direct me exactly where I should find this in DisplayCal?
In Resolve forum: “In DisplayCAL Profile loader (in the taskbar) you must define an exception for Resolve with Reset video card gamma table option so that calibration gets reset when Resolve is running since calibration is already written into the LUT.”
It is not correct. You MAY have set LUT3D with VCGT data or you MAY have choosen to keep gobal 1D grey calibration for all apps (hence LUT3D for resolve has no VCGT data because it relies on system wide 1D LUT in GPU)
My question:: I can not understand where exactly I will find the profile loader, can you direct me exactly where I should find this in DisplayCal?
On tray, right taskbar on Windows. As said above if it is not needed, it depends on what option did you choose.
For a multipurpose computer+widegamut for photo & video & web better to keep 1D calibration globally, hence the opposite that they recomend you (but Resolve LUT3D MUST not have VCGT data)Niklas, I think you can apply the LUT to the program monitor in Resolve as you can see here: youtu.be/ilGVwSvs674 (at 4:23). I haven’t tried it yet, let me know if it works.
Hi, I’m struggling to properly calibrate the LG 31MU97 all by myself… Even though I spent the last two days doing calibrations, tests and reading up on the forum and the whole internet about almost every advanced setting in displaycal. It’s like whenever I think I understood how to proceed, it’s not so, it’s exhausting.
Hope to find help here once and for all.
- So, only a few hours ago I realized that displaycal doesn’t change the measured gamut at all. With that said, I see the reason why I have to set the widest gamut possible for my monitor BEFORE running displaycal (I tried to start with an LG hardware calibration by setting sRGB/Rec.709 chromaticity coordinates of primaries, instead of leaving the native gamut, but then in displaycal I get 96% of sRGB gamut coverage… With the native gamut I had 99.7%), but I just don’t want such a saturated windows desktop, my eyes would get used to the wrong gamut compared to when I work inside color managed softwares… So what I need would be doing all the displaycal stuff with a native gamut HW cal as a basis, but then restrict my final gamut to rec.709, how do I do this? I was thinking to do a final HW calibration (after displaycal) with the rec.709 gamut, leaving it stored in the monitor’s HW, and then apply the previous displaycal ICC profile to windows system. Make sense? Or would I still get a lower gamut coverage percentage? All this rigmarole just to have the highest percentage possible at the end of the process…
- In all the tests I’ve done, if I open the profile information I have this type of plotted gamut right at its R G B corners (see attachments), are they normal? Shouldn’t they be more “linear”?
Thank you!
1-Use 2 HW cals, your LG model may have 2 slots CAL1/CAL2.
For non color criticl work, use CAL1 => sRGB/rec709 but 2.2 gamma
For apps with color management 1) switch to CAL2 (HW calibrated to native), 2) set its profile as default display profile, 3) open color managed app.
Alternatively if you have an AMD card, these cards have an sRGB emulation from EDID data, so starting with CAL2 setting (native) described and configured as above, you can switch on/off SRGB emulation on demand (off with color managed app, on with everything else. Don’t use GPU ‘s sRGB emulation on CAL2 with all profile stuff configued with color managed apps, all wil look undersaturated!)2-Maybe you are seeing tables in the opposite direction, maybe it’s so irregular, maybe those vbunds are small and do not matter (calibrating & validating same OSD preset settings with a matrx single curve profile without BPC will give you a hint about the culprit & severity).
Vincent, I often get similar gamut corners when I use XYZLUT+mtx profile type and 600-800 patches for 3D LUT profile part. But if I set 175 patches, smoothing is more strong, so the corners are more even. I also think that Profile Info 2D plot is either done using inverse tables, or with some error, its 3D plot looks more smooth. Any ideas?
To Googlois:
Beware, LG has recently published Tru Color Pro upgrade (dated 26-02-2021).
Don’t rush for perfection. You don’t realy need strict sRGB for surfing and office work, so for general work make hardware calibration for CAL1 with sRGB gamut, gamma 2.2, some good-looking white point near 6500K (set by xy, I use 0.3123 0.3280, but this depends on your colorimeter and its correction variation) and table-based profile type with maximum patch count (usually 200-400 in proprietary software), use v.2 ICC profile (some applications still don’t support v.4, and it’s not actually better). Yes, you may do DisplayCAL calibration upon harware one if it is difficult to find good xy for white point or DisplayCAL shows bad test results. Read about DisplayCAL settings bellow.
For color-critical tasks and wide gamut make hardware calibration for CAL2 with native gamut, gamma 2.2, 6500K WP or WP by xy that you’ve already found good (but I use 5000K for printing), table-based profile with maximum patch count. If you need high contrast (>700 is good for photo, >1200 is OK for consumer video), don’t use uniformity correction. Test it in DisplayCAL. You should get good WP, low error at greayscale, probably smoothly rising up to dE~1.3 in black, but dE~1.5 is OK for saturated dense colors. If test is failed, use DisplayCAL with the next settings: proper colorimeter correction (if you don’t have spectrophotometer), white/black point drift = on if external light may change (totally avoid switching bulbs, don’t wear saturated shirt), signal range 0-255, white point – better to use native (but if you haven’t got good one after HW calibration, set it by xy near 6500K, that is 0.3128 0.3292), gamma 2.2, black correction=auto, level 3.0 (if test shows dC>0.4 in darks), calibration=medium, profile type=XYZLUT+mtx (single curve is only usable to get good greyscale and to fasten output process, it makes weak correction for odd displays but is enough for well-tuned ones), black point compenation = on, press gear icon and set relative colorimetric intent as default, patches count = 175 (use 833 if test fails), testchart = auto-opimized. Check your monitor mode before start.
Another approach is not to use HW calibration, but DisplayCAL only upon sRGB, AdobeRGB or Rec.2020 (the last one means native gamut, but you don’t have Rec.2020) modes. It is more simple and may produce more smooth gradients, but gamut RGB corners may appear quite away of ICC standards.
Note on display brightness setting in DisplayCAL. I usually build “quick” profile with native brightness setting and test it to get brightness fall when I don’t set WP with RGB bars in display menu (notebooks and AIOs don’t have RGB control). I see that DisplayCAL tends to get wrong WP if calibration curves bend up. Some brightness reserve is needed here. By example, test shows 6630K WP and 160 cdm. So I tune up display brightness to 166 cdm at set target brightness to 160cdm in DisplayCAL.
Vincent, I often get similar gamut corners when I use XYZLUT+mtx profile type and 600-800 patches for 3D LUT profile part. But if I set 175 patches, smoothing is more strong, so the corners are more even. I also think that Profile Info 2D plot is either done using inverse tables, or with some error, its 3D plot looks more smooth. Any ideas?
Less patches, less detail. If you wish to find if this is an artifact from profilling or actual non ideal behavior jsut calculate dE with more idealized colospace (like 175 lut or matrix). If after checking you think that it is a profilling artifact send to argyll mailist, Graeme is who knows what is going on.
Regarding <- or -> direction, check in advanced settings (gear icon) it you set low detail in one of them.
