- This topic has 17 replies, 3 voices, and was last updated 10 months, 4 weeks ago by
.
-
Posts
-
So I’m trying to calibrate an LG OLED A1 by creating a 3D LUT (gamma 2.4, 709, 100 nits etc.) with i1 Display Pro and can’t seem to get it done. Since A1 does not have an internal 3D LUT capabilities, I’m trying to create the LUT with Resolve’s own pattern generator, connected via Decklink Mini 4K. The computer is PC, the A1’s black levels were set to limited. For colorimeter correction I used a generic one for LG OLED. White drift compensation was on.
However, after initial successful (I think so at least) calibration couple of months ago with DisplayCAL, I wasn’t able to replicate those results even when using exactly the same workflow. Couple of yellow hues are way off base according to the verification, and the results just got worse and worse after every profiling. The highest deltaE was around 5. The amount of patches was small since I was only trying to test things out and get the devices warm. Something troubling I noticed was the fact that the measured luminance jumped around quite a lot between profiling attempts and was even affected by the size of the patch window. Small one = high white luminance and reverse.
Right now I’m not sure how to proceed. I tried to find every available power setting that could have an impact on the brightness, but they were very limited, and don’t have any acces to the internal hidden menus.
Calibrite Display Pro HL on Amazon
Disclosure: As an Amazon Associate I earn from qualifying purchases.Something troubling I noticed was the fact that the measured luminance jumped around quite a lot between profiling attempts and was even affected by the size of the patch window. Small one = high white luminance and reverse.
Looks like ABL kicking in. It it worked in the past you may be aiming to bighter white than before, or fimrw change or some ABL option in OSD.
Are there any steps I could take to combat ABL, outside of getting a service remote to turn some of those features off.
two things i find essential for oled are full field pattern insertion and white drift compensation, i dont know if resolve tpg supports that, if not try madtpg which allows full field pattern insertion. also for patterns try maximize luma difference. there is a long lg c8 lut thread here check that for many good tips.
Okey so the full field pattern insertion (default values) and maximazing luma difference didn’t seem change the outcome. Pure yellow R255 G255 B0 seems to be problem, the measured one seems to be a bit dimmer as you can see in the report.
I’m only using the small testchart for LUT profiles and verifying using the 55 patch set so who knows how many problem colors there actually are but I’ll cross that brigde if I’ll ever come to it Right now I’d like to know if there’s any way re-profile the problem yellow without having go through the full patch set? Or are there any other steps I could take?
The latest report attached.
On Resolve’s profile folder there should be a ti3 file, try to locate measured yellows used for making a profile.
im also experiencing the yellow issue on the C1, but not as severe
here is the ti3 measurements for 255,255,0 yellow:
SAMPLE_ID RGB_R RGB_G RGB_B XYZ_X XYZ_Y XYZ_Z
400 100.000000 100.000000 0.00000000 70.6951725 87.8172861 5.98039355how can i interpret this?
On Resolve’s profile folder there should be a ti3 file, try to locate measured yellows used for making a profile.
Okey, what do I do next?
On Resolve’s profile folder there should be a ti3 file, try to locate measured yellows used for making a profile.
Okey, what do I do next?
Like NoVoiceMail, see if predicted value by profile matches ti3, if not something went wrong. Post info in argyllcms maillist
im also experiencing the yellow issue on the C1, but not as severe
here is the ti3 measurements for 255,255,0 yellow:
SAMPLE_ID RGB_R RGB_G RGB_B XYZ_X XYZ_Y XYZ_Z
400 100.000000 100.000000 0.00000000 70.6951725 87.8172861 5.98039355how can i interpret this?
HTML report, abs val, XYZ. Keep in midn that they are normalized to Y=100
Predicted by profile
76.9978
92.7814
13.8557Measured in report
72.8387
88.4292
13.6597Measured in ti3 for profile (IDBK if they are stored in abs or transformed to PCS)
70.6951725
87.8172861
5.98039355
with ti3 used fro making the profile and repost seek if something went wrong in “colprof” or “collink”in Argyllcms mailist
@Thomasin
you should use bscpylgtv to put the tv into bypass mode when calibrating/profiling (or native gamut with equivalent bypass settings 2.2 gamma 192,192,192, if A1 doesnt support bypass)
for the verification error, I did some tests with different types of verification on the same calibration. the calibration was done in 0-255 patches full range on the TV at 100 contrast/50 brightness, fast calibration speed and 175 patches. the yellow errors are always present when verifying using 16-255 patches and TV in limited range 85 contrast/50 brightness and also present when verifying in 0-255 full range patches at 100 contrast/50 brightness.
If you verified using 16-235 patches with TV in limited 85 contrast/50 brightness the report looks much better, why? its a bit of a mystery. what is the correct contrast setting at limited and full range settings in SDR is also a bit of a mystery. the convention is 100 contrast clips wtw when in SDR limited but is correct when in SDR full range.
