Rec. 1886 calibration questions

Home Forums General Discussion Rec. 1886 calibration questions

This topic contains 16 replies, has 2 voices, and was last updated by  Alex Sikorsky 6 months ago.

Viewing 15 posts - 1 through 15 (of 16 total)
  • Author
    Posts
  • #15696

    Alex Sikorsky (@alex-5)
    Participant
    • Offline

    I monitored the development of the program and from time to time calibrated it to check the state of affairs, but always rolled back to another software for calibration and there were enough reasons. After reading the next update, I downloaded the latest version 3.7.1.4 and tried to calibrate the monitor on EOFT Rec.1886. In this post, similar calibration is also considered when OEFT Rec.709. Using the * .ccmx file built back in the last year with an iPro2 spectrophotometer for i1Display + Nec2690WUXi. (1) When setting up, the system constantly suggested reducing the backlight point, which contradicts the logic, since Initially, the black and white points were set in the program and defined in the calibration tab. (2) Constructed a profile in 3 and iteration with compensation from possible temperature drift – the operation went on for a very long time, almost 3 hours and 20 minutes. (3) The dark profile turned out to be littered with EOFT Rec.1886, and there is something to check the result with than software. (4) I tried to correct the contents of the video adapter LUT, according to the target function Rec.1886, but saw that the profile auto-loading program does not turn off and almost immediately after a reset (cyclically) does restore the video adapter’s LUT, after which it had to be disabled and the profile removed from the system and roll back. And now the questions: (1) How to clarify only the gamma function by building a matrix profile? (2) and then make only the characterization of the monitor on the existing configurations with the construction of the final tabular profile? (3) Is it possible for this to create your own target reference.txt files created in the same i-Profiler?

    #15698

    Florian Höch (@fhoech)
    Administrator
    • Offline

    Hi,

    (1) When setting up, the system constantly suggested reducing the backlight point, which contradicts the logic, since Initially, the black and white points were set in the program and defined in the calibration tab.

    If you set a white level target, then this becomes the target during interactive adjustment. If you don’t set a white level target, then the target during interactive adjustment is just the initial luminance. It is up to you whether you want to use it or not.

    (3) The dark profile turned out to be littered with EOFT Rec.1886, and there is something to check the result with than software.

    Sorry, I don’t understand what you mean here. Is it a question?

    (4) I tried to correct the contents of the video adapter LUT, according to the target function Rec.1886, but saw that the profile auto-loading program does not turn off and almost immediately after a reset (cyclically) does restore the video adapter’s LUT, after which it had to be disabled and the profile removed from the system and roll back.

    Altering the videoLUT after the display has been calibrated and profiled invalidates the profile. Therefore, altering the videoLUT contents manually is diametrically opposed to a color managed workflow that makes use of the videoLUT for calibration.

    (1) How to clarify only the gamma function by building a matrix profile?

    Sorry, I’m not sure I understand the question. So you want to create a matrix profile? In that case, move the patch amount slider on the “profiling” tab all the way to the left (34 patches). If you want more control, you can enable advanced options in the options menu.

    (2) and then make only the characterization of the monitor on the existing configurations with the construction of the final tabular profile?

    Not sure what you are asking here either, please clarify. You want to create a matrix profile first, and then a table-based one? Why not only the table-based one?

    (3) Is it possible for this to create your own target reference.txt files created in the same i-Profiler?

    Using the testchart editor, you can create your own testcharts (small button next to dropdown on the profiling tab, same on verification tab).

    #15699

    Alex Sikorsky (@alex-5)
    Participant
    • Offline

    * For calibration, I used the * .ccmx file created last year, with an iPro2 spectrophotometer (EFFI-2000) for i1DisplayPro + Nec2690WUXi. Then I repeated everything with another device – a Spyder4 colorimeter and the result was similar.

    BEGIN_ARGYLL_DISPCAL_ARGS
    -v2 -Q2012_10 -qm -t6500 -b100.71 -G2.4 -f0 -a4.9 -k0 -A0.25 -B0.276200 -d1 -c1 -yr -P0.549077490775,0.507086614173,1.61891117479 -X “NEC LCD2690WUXi (i1 Pro 2).ccss” -Ibw
    END_ARGYLL_DISPCAL_ARGS

    EOFT Rec.1886 can be calculated in the spreadsheet editor, and then get a simple gamma function for each point of the curve to control what is happening. Regardless of what the monitor itself is calibrated for, it can be calibrated to any EOFT for display. What did I do wrong when calibrating?

    Attachments:
    You must be logged in to view attached files.

    i1Display Pro on Amazon   i1Basic Pro 2 on Amazon  
    Disclosure: As an Amazon Associate I earn from qualifying purchases.

    #15701

    Alex Sikorsky (@alex-5)
    Participant
    • Offline

    Yes, if you edit LUT after calibration, this is against logic, on the one hand, but it can show how accurate the system is calibrated by the system, taking into account the target gamma function. Then you can recalibrate the system, making only a characterization taking into account the new LUT changes. From this, the question arose to first construct the matrix profile but taking into account the selected function EOFT, check the quality of the gamma function itself, and then characterize and obtain the final profile. Sorry for my bad english.

    #15703

    Alex Sikorsky (@alex-5)
    Participant
    • Offline

    Sorry, I don’t understand what you mean here. Is it a question?

    After building the profile, I got a completely different gamma function, which in fact turned out to be darker than rec.1886 should have been.

    #15707

    Florian Höch (@fhoech)
    Administrator
    • Offline

    After building the profile, I got a completely different gamma function, which in fact turned out to be darker than rec.1886 should have been.

    -v2 -Q2012_10 -qm -t6500 -b100.71 -G2.4 -f0 -a4.9

    You are using ambient light level adjustment (-a4.9). this is intended to influence the calibration, so if your goal is to have plain Rec. 1886 EOTF, disable that ambient light level adjustment.

    You can also probably cut down calibration time quite a bit by disabling white level drift and black calibration drift compensation (the latter is only useful for spectrometers, the former only needed if the display light output is not stable over time).

    #15708

    Alex Sikorsky (@alex-5)
    Participant
    • Offline

    After building the profile, I got a completely different gamma function, which in fact turned out to be darker than rec.1886 should have been.

    -v2 -Q2012_10 -qm -t6500 -b100.71 -G2.4 -f0 -a4.9

    You are using ambient light level adjustment (-a4.9). this is intended to influence the calibration, so if your goal is to have plain Rec. 1886 EOTF, disable that ambient light level adjustment.

    You can also probably cut down calibration time quite a bit by disabling white level drift and black calibration drift compensation (the latter is only useful for spectrometers, the former only needed if the display light output is not stable over time).

    Thanks for the recommendations. How to remove from the registry profile loading in LUT? (Which branch to clean?)

    #15709

    Florian Höch (@fhoech)
    Administrator
    • Offline

    Thanks for the recommendations. How to remove from the registry profile loading in LUT? (Which branch to clean?)

    You can disable the profile loader via its tray icon right-click menu (remove check from “load & preserve calibration”).

    #15712

    Alex Sikorsky (@alex-5)
    Participant
    • Offline

    Already figured out – your program is loaded not from the registry, but from the task manager Windows.

    Calibrated according to your recommendation (disabled the drift correction for current), reduced the number of scales to 34. Medium calibration mode, characterization – High. Enabled antialiasing addition. The result is better, but in the shadows the curves spread out and at the same time did not correspond to the declared function EOFT.1886. To test, I have curves built in a spreadsheet editor, like the program for checking correctly calibrated gamma functions. while I study the program more deeply, it’s too early to draw conclusions, but it calibrates faster with a colorimeter than a spectrophotometer – it does this with big errors. A large number of patches does not improve the result.

    Attachments:
    You must be logged in to view attached files.
    #15714

    Alex Sikorsky (@alex-5)
    Participant
    • Offline

    I will still try to understand the pros and cons of the program. I suppose that a large number of interractions and operating time contribute to the overheating of the devices, which makes the shadows bluish at spectrophotometers and the accuracy of curve construction is disturbed even at colorimeters in the shadows …

    Attachments:
    You must be logged in to view attached files.
    #15716

    Alex Sikorsky (@alex-5)
    Participant
    • Offline

    //In ( OEFT rec.709) := := if L<0.018 then V=4.5*L else V=1.099*L^(0.45)-0.099 , (Gamma 0,45=1/2.22222)

    //Out (EOFT rec 1886): = (a * (MAX ((N_Count / 255) + b; 0)) ^ (2.4)) * 255, where a = (L_wp ^ (1 / 2.4) -L_bp ^ (1 / 2.4) ^ 2.4, b = L_bp ^ (1 / 2.4) / (L_wp ^ (1 / 2.4) -L_b ^ (1 / 2.4)), and L_wp and L_bp are the brightness points of white and black respectively, in Cd / m2

    // Check according to the gamma of each point for a range of values (0-255): Out=(255*(Count/255)^GammaOut), GammaOut: = IF (1> N / 255; IF (N / 255> 0; LOG (Out(f_rec.1886) / 65535; N / 255); 1); 1), where Out (0 … 65535), a N = ( Count) (0 … 255)

    This is how I checked the accuracy of the gamma function achieved …

    • This reply was modified 6 months, 1 week ago by  Alex Sikorsky. Reason: literal error
    Attachments:
    You must be logged in to view attached files.
    #15720

    Florian Höch (@fhoech)
    Administrator
    • Offline

    I suppose that a large number of interractions and operating time contribute to the overheating of the devices, which makes the shadows bluish at spectrophotometers

    Yes. For that reason, for prolonged measurement runs, a spectrometer needs to be regularly dark-calibrated (i.e. taken off the screen and put on the reference tile). Argyll normally asks for this in around 30 minute intervals I think, but you can disable it in the advanced options (not recommended!). Usually it is better to use a colorimeter because they are more stable over time.

    the accuracy of curve construction is disturbed even at colorimeters in the shadows …

    The picture you attached shows the difference in RGB separation. This is just informational and not suitable for visual evaluation. For that, you can look at the delta C component in the overview.

    This is how I checked the accuracy of the gamma function achieved …

    I can’t easily verify the spreadsheet. If you attach the profile and related files (“create compressed archive…”, next to settings with profile selected), I can take a look.

    #15722

    Alex Sikorsky (@alex-5)
    Participant
    • Offline

    This is how I checked the accuracy of the gamma function achieved …

    I can’t easily verify the spreadsheet. If you attach the profile and related files (“create compressed archive…”, next to settings with profile selected), I can take a look.

    profile *.icm

    Attachments:
    You must be logged in to view attached files.
    #15724

    Alex Sikorsky (@alex-5)
    Participant
    • Offline

    This is how the curve (1) looks like, which is built during calibration, and that should be, if corrected (2)

    Attachments:
    You must be logged in to view attached files.
    #15727

    Florian Höch (@fhoech)
    Administrator
    • Offline

    This is how the curve (1) looks like, which is built during calibration, and that should be, if corrected (2)

    Hmm no, that doesn’t seem to make sense. The “should be” curve looks wrong (darkening).

    The profile follows pretty close BT.1886 with the given parameters (black level = 0.2924 cd/m2, normalized to 0..100). See attached tone curve plots (top = your profile, bottom = idealized Bt.1886 profile with same black and white level parameters). Note that both profiles scale black to zero in the tone response curves.

    Attachments:
    You must be logged in to view attached files.
Viewing 15 posts - 1 through 15 (of 16 total)

You must be logged in to reply to this topic.

Log in or Register

Display Calibration and Characterization powered by ArgyllCMS