- This topic has 20 replies, 2 voices, and was last updated 7 years, 1 month ago by
.
-
Posts
-
Hello Florian, I have a problem creating a matrix profile with DisplayCal.
I used a colorimeter i1DisplayPro and one i1Pro2 spectrophotometer on a LG OLED 55B6.
Unfortunately, however, despite I have repeated several times measurements, the results are always improper.First I would like to ask if I understand correctly the image that I enclose.
The table represents the measures spectro (reference) and the measure of the colorimeter with the applied correction? so with the relative difference in deltaE? or it is a comparison of the measures spectro colorimeter but without the correction?I specify that I have made several attempts to measure both in contact mode and in distance mode (13 cm for i1DisplayPro and 18 cm for the i1Pro2).
As a pattern generator I use davinci resolve but I also tried plugging the TV directly to the PC directly with the same results.
The spectro is new never used so I would say that it is accurate.
I do not understand what I’m neglecting or where I’m wrong.
What you do you think?Thanks
Calibrite Display Pro HL on Amazon
Disclosure: As an Amazon Associate I earn from qualifying purchases.Hi,
Unfortunately, however, despite I have repeated several times measurements, the results are always improper
What leads you to this conclusion?
The table represents the measures spectro (reference) and the measure of the colorimeter with the applied correction?
Correct.
Hi,
Unfortunately, however, despite I have repeated several times measurements, the results are always improper
What leads you to this conclusion?
The table represents the measures spectro (reference) and the measure of the colorimeter with the applied correction?
Correct.
The tollerance is not 0.001 for x and y, while 1,5% for Y?
The tollerance is not 0.001 for x and y, while 1,5% for Y?
I’m not sure what you’re asking. The main point of a colorimeter correction is to make two different instruments read a display the same (within repeatability of the instruments involved and limitations imposed by a 3×3 matrix), and this is what this matrix achieves very well (delta E of 0.05 for white is imperceptibly small, and max delta E of about 1 for fully saturated red is also just at the perceptibility threshold). The difference on white for Y is 0.02%. Due to a matrix being strictly additive, it is not possible to have a matrix fix white and saturated colors to the same extent though if the underlying display is not strictly additive as well (TVs are usually poorer at additivity due to internal processing). Usually this is not a problem because it is far more important that the white point readings (and thus neutrals) are corrected, as the human visual system perceives colors in relation to the white, and the human eye is quite good at making out even small differences in neutrals. The more chromatic (saturated) a color gets (moves away from neutral), the less the eye will be able to make out color differences.
Florian, thank you for your explanation.
The fact is that I’ve read elsewhere that differences between the reference spectro and the colorimeter (corrected) should be thinner.
For example, if I insert the results obtained in this sitehttp://displaycalibrationtools.com/display-profiling-calibration-tools/meter-profile-comparison/
which should validate the quality of the DisplayCal matrix does not pass the test.
Furthermore with the same workflow and the same measuring instruments, on another display that is a oled rgb Odyssey, I got from DisplayCal much better results as you can see in the attached image.For all these reasons, I had concluded that there is a problem on this specific product lg oled b6.But you’re telling me that the results are good the same and that there is no problem?
on another display that is a oled rgb Odyssey, I got from DisplayCal much better results
Yes, more typical for matrices produced for a display with good additivity is a deltaE 2K of <= 0.05 for all four measured colors.
For all these reasons, I had concluded that there is a problem on this specific product lg oled b6.
The “problem” you’re seeing highlights a limitation of this particular TV in its current state (less good additivity).
But you’re telling me that the results are good the same and that there is no problem?
What I’m telling you is that there is no problem with the matrix itself per se.
Thanks Florian, one last question please:
If the problem is not in the matrix that calculates DisplayCal why the matrix that calculates Lightspace not detect the limit of this tv (less good additivity)?
In Lightspace comparing the measurements of the reference and the colorimeter (correct) there are no alarming differences in x and y, but only in Y on red and green.Finally I would like to ask you if you can suggest me some tricks or methods to try to improve the result of the matrix.
There is a difference, for example, if instead of LCD (generic) Adaptive imposed LCD (generic) HiRes?Thanks
If the problem is not in the matrix that calculates DisplayCal why the matrix that calculates Lightspace not detect the limit of this tv (less good additivity)?
It will show up in the matrix, but developers are of course free to choose which compromise to take if display additivity is not perfect. Argyll does favor white as I mentioned and for the reasons I outlined.
In Lightspace comparing the measurements of the reference and the colorimeter (correct) there are no alarming differences in x and y, but only in Y on red and green.
This depends on what you consider “alarming”. DeltaE should be the deciding factor, not simple percent differences in either x, y, or Y.
There is a difference, for example, if instead of LCD (generic) Adaptive imposed LCD (generic) HiRes?
HiRes or not may make a difference with the i1 Pro, depending on how peaky the measured spectra are, adaptive probably not as it mainly helps with dark colors, and these are not measured as part of the four color matrix method (adaptive is the default though, so if you’re going HiRes, use adaptive HiRes).
Hello Florian, trying to compare my DisplayCal data with those of others, the Ted user tells me that even given the consumer level of this OLED, it is not normal to get my results.
http://www.avsforum.com/forum/139-display-calibration/2644073-official-b-c-e-g6p-oled-calibration-thread-38.html#post51666737
His suggestion is to ask how DisplayCal calculates errors because there is a considerable difference in the values of translation with other programs.
Thanks for your patience, I would like to understand where is the problem.I replied in the thread.
I would like to understand where is the problem.
I don’t think there is one, but please attach both the reference and colorimeter measurements file you created the correction from.
I do not know if the attached files are what you asked me.
That’s the one half of the files I need, the others are those from your reference spectro (i1 Pro 2 according to your screenshot above).
