After calibration, I can’t see P3 color space anymore

Home Forums Help and Support After calibration, I can’t see P3 color space anymore

Viewing 15 posts - 1 through 15 (of 18 total)
  • Author
    Posts
  • #29016

    Enrico Gugliotta
    Participant
    • Offline

    Good morning everyone.

    I noticed a strange thing after the calibration I’ve done on june of 2020.

    Before that, I used a calibration made with Spyder5 Pro and Spyder software, which seemed good, and permitted to also see the variations in P3 color space (see img Calibration 2018).

    After that, I bought a SpyderX Pro and calibrated my monitor with DisplayCAL. The colors looked good than before, but I noticed that I can’t see the P3 color space anymore (see img Calibration 2020).

    Is this normal or something went wrong in the calibrabion with DisplayCAL? My monitor is a wide gamut Asus PA329Q

    Thanks in advance

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

    SpyderX Pro on Amazon  
    Disclosure: As an Amazon Associate I earn from qualifying purchases.

    #29027

    Vincent
    Participant
    • Offline

    I’d say that your test app does not support that kind of profile type. Use single curve matrix +bpc or defaults if using MacOS

    BTW bad measureent device, bad display and backlight unsupported by all spyders. For backlight corrections for colorimeter should be close to native gamut SW2700PT user made CCSS or if forced to used bundled ones RGB-LED

    • This reply was modified 1 month, 1 week ago by Vincent.
    • This reply was modified 1 month, 1 week ago by Vincent.
    #29031

    Enrico Gugliotta
    Participant
    • Offline

    Maybe I haven’t explain me well.

    This is the website I’m talking about: https://webkit.org/blog-files/color-gamut/

    The second row of images is the one I’m talking about. When I used my calibration (2018) I could see the webkit logo in P3 space. Now after calibrating with DisplayCAL I can’t and I only see the red square, as you can see in the attachments

    #29032

    Vincent
    Participant
    • Offline

    I’d say that your test app does not support that kind of profile type. Use single curve matrix +bpc or defaults if using MacOS

    Already answered.

    #29033

    Vincent
    Participant
    • Offline

    = Use the same ultrasimplified profile type in DisplayCAL as Datacolor software uses, since some apps do not support more complex profiles

    • This reply was modified 1 month, 1 week ago by Vincent.
    #29035

    Enrico Gugliotta
    Participant
    • Offline

    Sadly I’m not understanding much.

    Are you saying that the calibration with DisplayCal is not supported by the browser, so I’m not seeing the P3 space on that site, while the calibration with Spyder software is more browser-friendly? I’m using Windows 10, not MacOs.

    #29037

    Vincent
    Participant
    • Offline

    No. I’m saying that YOUR choice of PROFILE TYPE in DisplayCAL is not friendly with some apps that only can use oversimplified profiles. This is not related to calibration at all.

    Configure DisplayCAL as explained and those apps with limited color management will work as intended.

    • This reply was modified 1 month, 1 week ago by Vincent.
    • This reply was modified 1 month, 1 week ago by Vincent.
    #29044

    Алексей Коробов
    Participant
    • Offline

    Enrico, your display is unable to show red edge of DCI-P3. It has similar gamut to my BenQ PV270 (though, panels are different), I attached gamut images for default Asus PA329 ICC. This is not new, even top Eizo displays allow to support 98-99% for AdobeRGB and DCI-P3 both, but all others with AdobeRGB full (99%) coverage don’t achieve grass-yellow-red edge of P3, having 94-96% of P3 gamut. You may check that red image in Photoshop with soft proof tools – gamut limits (set your display profile as a printer profile).

    If you saw red logo before, this only prooves that Datacolor calibration was wrong, or display ICC wasn’t used by browser for some reasons (not supported?), so display showed you Paint-like image with all gradients, but wrong colors. Or browser CMS was upgraded some time before. I see that all main browsers in Windows started to support XYZLUT+mtx ICC type in November 2020. While each one has its own CMS components, all could use some Windows library that had been changed after regular uprade. Note also that browsers aren’t precise in colors and may not support non-sRGB profiles.

    Check profiling result in DisplayCAL test tab. Beware of display settings.

    Attachments:
    You must be logged in to view attached files.
    #29052

    Vincent
    Participant
    • Offline

    No Alexei, lot of browsers do not support table-lut profiles, only matrix. When they use unsupported profiles color managements breaks as if there where no CM.

    A working CM on a browser or other app when facing a color out of display gamut will present nearest using relative or perceptual intents => a matrix profile in a GB-LED 93? % P3 will show 2nd image correctly (cutting to nearest whith chosen rendering intent). That red won’t be sRGB red, it will be nearest “P3 255 red” color it can render.

    What he is seing is unsupported profile type by browser because HE configured DisplayCAL that way. If he chooses proper profile setup image will render ok,

    (also 1st one screenshot seems bad profile)

    • This reply was modified 1 month ago by Vincent.
    #29054

    Vincent
    Participant
    • Offline

    (1st screenshot from Enrico, not from Alexei)

    #29056

    Алексей Коробов
    Participant
    • Offline

    Vincent, I disagree, but see $5.

    1. I’ve checked it at my PV270 with native gamut in Ps – red square, logo is invisible.
    2. I agree that gamut warning at Ps does not guarantee color clipping, it only tells us that color is shown wrong at display.
    3. Right, it should be visible when perceptual intent is used (perceptual conversion from image ICC to display ICC). But not colorimetric.
    4. I always set relative colorimetric intent for XYZLUT+mtx profiles and I clearly see red gradient clipping in browsers (Chrome and some others, not all) when display does not achieve image ICC gamut. Before November 2020 this had only been happened when 3curves+mtx (and single curve + mtx, of course) was used. But I also see some wrong color in browsers, low quality CM realization.
    5. I’ve recently checked Firefox, Chrome and Opera with CAL1 and corresponding native gamut ICC by Palette Master. – CM does not work. But I always see different situation (with Chrome at least) at simple displays, calibrated by DisplayCAL. Need to be investigated. What display do you use, is it HW-calibrated?

    #29058

    Enrico Gugliotta
    Participant
    • Offline

     

    What he is seing is unsupported profile type by browser because HE configured DisplayCAL that way. If he chooses proper profile setup image will render ok,

    (also 1st one screenshot seems bad profile)

    My calibration profile in DisplayCAL is WLED PFS / RGB LED, which I think it’s the correct profile for my monitor. But probably I’m wrong, I don’t know.

    #29060

    Vincent
    Participant
    • Offline

    Vincent, I disagree, but see $5.

    1. I’ve checked it at my PV270 with native gamut in Ps – red square, logo is invisible.

    There are 3 images. 1st is full 255 sRGB red. 2nd one is P3 imagewhich softproofed in PS to sRGB leaves you full 255 sRGB red.

    3. Right, it should be visible when perceptual intent is used (perceptual conversion from image ICC to display ICC). But not colorimetric.

    No. It will be visible with relative, no doubts about that. By definition on a GB-LED display it will render (clipped) as the closest in a native gamut red wich will be more saturated than sRGB.
    Download P3 image (click P3 thumbm then download bigger. You’ll see by yourseft.

    5. I’ve recently checked Firefox, Chrome and Opera with CAL1 and corresponding native gamut ICC by Palette Master. – CM does not work. But I always see different situation (with Chrome at least) at simple displays, calibrated by DisplayCAL. Need to be investigated. What display do you use, is it HW-calibrated?

    Profies from PM/PME are broken for those browsers. Same for i1Profiler derived works. Not supported. PME’s lack of CHAD and the other ones seems D50 PCS white with CHAD and most of them do not like it. I’ve not seen implementation details but If had to make a guess I woult point at it as the culprit.
    Reprofile with matrix+1TRC with no GPU calibration on CAL1/CAL2 and they should work. I’m seeing that test on Firefox and Edge with no issue (Edge seems to have a lag after loanding until CM runs IDNK why)

     

    What he is seing is unsupported profile type by browser because HE configured DisplayCAL that way. If he chooses proper profile setup image will render ok,

    (also 1st one screenshot seems bad profile)

    My calibration profile in DisplayCAL is WLED PFS / RGB LED, which I think it’s the correct profile for my monitor. But probably I’m wrong, I don’t know.

    That is not profile type. That is backlight type. It is different with no relation. Do as instructed.

    Regarding backlight correction for colorimeter white may look not white after calibration (slight green or magenta cast), that’s all. It is not related to your browser issues. Make a profile with profile type single curve matrix with balck point compensation, calibration speed medium/slow and thats all. Close all browsers, make sure your new profile is loaded and set as default profile, open browser on that test and all will look as intended.

    #29061

    Алексей Коробов
    Participant
    • Offline

    Vincent, I disagree, but see $5.

    1. I’ve checked it at my PV270 with native gamut in Ps – red square, logo is invisible.

    There are 3 images. 1st is full 255 sRGB red. 2nd one is P3 imagewhich softproofed in PS to sRGB leaves you full 255 sRGB red.

    3. Right, it should be visible when perceptual intent is used (perceptual conversion from image ICC to display ICC). But not colorimetric.

    No. It will be visible with relative, no doubts about that. By definition on a GB-LED display it will render (clipped) as the closest in a native gamut red wich will be more saturated than sRGB.
    Download P3 image (click P3 thumbm then download bigger. You’ll see by yourseft.

    I made tests in Ps with that image, P3 ICC, I understand it. And I get similar results in almost all cases. You see that Asus color gamut does not achieve P3 red at any brightness level, so R=241…255 may be clipped at the same point, this depends on conversion realization (does ICC standard  determine out of gamut color conversion for colorimetric intents?). I usually sign AdobeRGB ICC to sRGB test images to show the effect on office screens. I know, that out of gamut is not equal to posterized clipping. By example, 45% NTSC  notebook panels show a part of red gradient before clipping appears, but all pure reds are wrong cause panels gamut don’t achieve sRGB gamut ceiling (in 3D L*a*b).

    #29064

    Enrico Gugliotta
    Participant
    • Offline

     

    My calibration profile in DisplayCAL is WLED PFS / RGB LED, which I think it’s the correct profile for my monitor. But probably I’m wrong, I don’t know.

    That is not profile type. That is backlight type. It is different with no relation. Do as instructed.

    Regarding backlight correction for colorimeter white may look not white after calibration (slight green or magenta cast), that’s all. It is not related to your browser issues. Make a profile with profile type single curve matrix with balck point compensation, calibration speed medium/slow and thats all. Close all browsers, make sure your new profile is loaded and set as default profile, open browser on that test and all will look as intended.

    Ok thanks Vincent, sorry for my ignorance in these setting. Just wanted to say that my work is totally Lightroom\Photoshop + prints for my clients. I don’t care if on that particular website I can’t see the P3 color space.

    So, keeping this in my mind, if I do what you say I’m going to be ok with my workflow “and” also seeing the P3 color space on browser?

    Settings are attached

    Attachments:
    You must be logged in to view attached files.
Viewing 15 posts - 1 through 15 (of 18 total)

You must be logged in to reply to this topic.

Log in or Register

Display Calibration and Characterization powered by ArgyllCMS