Originally Posted by mywaycn
my question is what setting i might wrong to cause 3D LUT so bad but RGB balance in post cal check is looks good in SDR cal.
if you have any suggestion, i am all ears. Thank you so much.
TV: LG OLED 65 B8P
SW: Calman 22.214.171.124 Home Enthusiast
Meter: i1Display pro retail 2011 model. (i bought it aroud 2012, so it should be. 0.1 - 1000 cd/m2)
TPG: Rpi PGenerator with 2.0.16 DeviceControl on Rpi 3B+
TPG: 1080p24, RGB Full
TV: Active grayscale points = LG 26 pts 16-235
PG control interface options: Windows size 10%, Pattern APL 25, Pattern size 8
Workflow baisc option: Colorspace target= D65, HD, BT.709; Gamma fomula=Power 2.2; Luminance level=Video (16-235)
Application measurement options: Full Fieid Pattern insertion= Freq 30 (1D LUT), 15 (3D LUT); Duration 5; level 15% ; AutoCal targets DeltaE formula=de2000_jNDab, Cube deltaE formula=de2000
3D LUT: iRP point based LUT = 5000 pts
Post Calibration View
is confuse me that is RGB balance looks flat so i assume which is good. Howerer, de2000 color delta error is huge.
Hi, when you want to generate APL with custom size patterns, you need to select user defined and set APL and size.
When you select 10% Windows, it will generate 10% window and ignore the APL & size setting.
PGenerator always need to have RGB-Full (as device settings from inside its interface, as this is the only correct setting for patch generation.)
Then to the software you will use, for example to CalMAN you are using, if you want 16-235, you set from inside CalMAN to generate 16-235, if you want 0-255 then you set 0-255 from CalMAN. For LG use always video levels.
I have examined your LG 3D LUT file generated from CalMAN.
Since CalMAN don't have any tool to examine any 3D LUT data file, the only way of verification is measurements of grayscale/saturation/colorchecker etc.. measurement or checking visually some color reproduction patterns or real content.
For that reason I used LightSpace where it has 3 tools to examine 3D LUT files. (for not entering to more details about these tools, you can see about 3D LUT Viewer
, LUT Preview
and 1D LUT Viewer details to older posts)
The data are not looking good, the cube shape has many problems to various areas, its completely useless LUT generation, while the same time the 1D LUT viewer looking good (except the usual unnecessary luminance drop (top right), explained again there
), I mean there no strange issues (see 1D LUT picture there
, because has large size for posting it here)
Its very visible also the Yellow over-saturation bug, which exist to any LG 3D LUT generation from CalMAN for at least 1 year, see more details about the problem there: https://www.avsforum.com/forum/139-d...l#post56558088
or there: https://www.avsforum.com/forum/139-d...l#post56555486
Generally CalMAN can produce better results when you are using less points, as user feedback has shown over the years if you are watching this thread... It's better to use up to 3500 or less (not sure if it will measure 3500 different, but if you enable full logging and open the data files and check you will see if its measured 2-3 times the same patch).
Its not normal with less points to produce better results but its the up to capabilities of the software.
The problems you experience are easily noticed when I used LUT Preview, it will be the same issues as you see to your LG (the following files are animated APNG files, they will swap your LUT ON/OFF per 1 sec....if you don't see animation try other browser which support APNG files):
BTW the issues you have are similar with the other user with projector also there: https://www.avsforum.com/forum/139-d...l#post57787698