2021-08-28 at 0:38 #31401
Hi. I’ve been trying on an off to get my TV/PC-monitor calibrated for a while now.
I’ve focused on 3D LUTs for madVR and therefore haven’t imported any profiles in windows yet.
The tv is a Samsung UE82NU8005 (simply put an NU8000 set.)
Today I finally managed to get results I’m happy with, and this has been my process:
Display & instrument chose madvr calibration use white level drift compensation spectral correction for a similar but smaller TV (49NU8000) Calibration Interactive display adjustment: no Tone curve: As measured Profiling Profile quality: High Testchart: Very large testchart for LUT profiles 3D LUT Create 3D LUT after profiling: yes Source colorspace: DCI-P3/SMPTE-431-2 D65 Tone curve: Rec. 1886 Rendering intent: Absolute colorimetric with white point scaling 3D LUT file format: madVR (.3dlut) Input encoding: TV RGB 16-235 *When Profile only is pressed: Use linear calibration instead: checked Embed calibration curves in profile: unchecked MadTPG: Image area: 10% Background: 25% APL linear light stay on top/use fullscreen/disable osd Verification Testchart or reference: Extended verification testchart Simulation profile: Rec709 ITU-R BT.709 Use simulation profile as display profile: check Enable 3D LUT: check Tone curve: Rec. 1886
I’ve done the same to make a different profile for using madVR HDR -> SDR tonemapping, but with these changes:
3D LUT Source colorspace: DCI-P3/SMPTE-431-2 D65 Profiling Simulation profile: unchecked BT.2020 activated in madVR to activate wide gamut on the tv TV brightness set to max
*These settings shouldn’t have any implicatin for my use right?
But what about when I make a calibration for my desktop?
On my TV it’s impossible to completely deactivate the local dimming, but I’ve set it to low.
What finally seemed to help me was to set the background in madTPG to APL 25% linear light.
If I understand correctly this should make the backlight stay at the same level during the whole calibration.
My issues comes when I try to verify these results.
The first verification of my bt.709 profile was very good and I had a tick marking recommended tolerances passed 😀
The next verification was less good. everything is the same as the first time, but only nominal tolerances passed this time.
The .bt2020 verification had Maximum ΔE*00 error at 5.09! And this is just verifying my profile, not against the bt.2020 gamut.
Have I done something wrong, or is my set just this unstable?
Is there something I can do to improve this? Would it help to use more than 4954 test patches?
I’ve added the Measurement report in case they’re interesting.
Attachments:You must be logged in to view attached files.2021-10-10 at 16:19 #31950
Does anyone have any thoughs on my workflow and results?