How to see uniformity result like this one?

Home Forums Help and Support How to see uniformity result like this one?

Viewing 12 posts - 1 through 12 (of 12 total)
  • Author
    Posts
  • #19058

    radiodus
    Participant
    • Offline

    I tested uniformity.

    But I don’t know how to see uniformity result like upper image. (ex: CCT ****K CDT ****K)

    How to see CCT / CDT in uniformity result?

    Attachments:
    You must be logged in to view attached files.
    #19063

    Florian Höch
    Administrator
    • Offline

    You can’t and you shouldn’t. CCT is not a valid measure of uniformity, because it only tells you about the red-blue color difference and ignores green. Delta E (full color difference) or Delta C (chroma component difference) is a much better measure of uniformity (combined with luminance difference).

    #19070

    Vincent
    Participant
    • Offline

    Florian is right, no use for CCT values in an uniformity report. CCT give you a hint about blue-yellow axis deviation in different places of screen, buy you get no information about the extremely common green-magenta tints across screen.

    dE (“total error”, default test) is linked in DisplayCAL to an ISO validation. A PASS/NOT PASS test which is hard to pass for common displays. All green is expected for an A grade monitor meant for graphic arts professional work.

    dC + brightness % is more like an “user driven” relaxed level of aceptable quality.
    About -15% brightness drop in corners is very probable in common displays. It does not mean that is good, just that is common for inexpensive displays (very common if display is big). It’s “expected”.
    dC is related to strong tints in display, like having a green cast in the left and a red/pink cast in the right. Most acceptable quality monitors are under 2dC in most used zones of screen and under 3dC on sides. That does not mean that a 3dC on a corner is good, just means acceptable quality for inexpensive displays not meant for color critical work.
    Most common sRGB IPS displays with reasonable price are expected to meet this relaxed common-use non-ISO criteria.

    Widegamut displays and big displays are more prone to suffer uniformity issues, specially if they are inexpensive for their size+resolution.
    For example:
    $700 is an expensive 27″ QHD sRGB monitor, but the same price means a low cost inexpensive 27″ QHD widegamut with expected very low quality control.
    $1200 is an expensive widegamut 27″ QHD (an A grade) but the same price in 32″ UHD widegamut means a low cost inexpensive display with almost no quality control.

    Your report shows extreme severity color cast, specially on right side.
    IMHO that means inmediately return for refund. I’m not saying that it is not acceptable for color critical work, I’m saying that I would not accept it even for office work.
    Of course user/owner has the last word about keeping or returning it.

    Since yours is a widegamut display, it’s expected that you wanted to use it for some kind of color critical work (otherwise you woud have choose other type of display, like a common IPS sRGB or a multimedia P3 with VA).
    Probability is against you if you expect to get an acceptable quality 32″ UHD IPS AdobeRGB display in that price range, same for 27″ UHD. It’s not impossible, it’s just highly improbable.
    A more sensible choice if yiu want a display for color critical would be an entry level reliable display in 27″ QHD like Eizo CS2730. Price should be close to what you paid for that SW320 and you won’t suffer dealing with Palette Master Elements and its bugs & limitations. Same warning goes for UP 32**Q  Dells but as I said before, you have the last word regarding money and buying decisions.

    #19071

    radiodus
    Participant
    • Offline

    Oh … you’re right.
    I mistakenly thought of the concept of tint and temperature as the opposite.

    My monitor have a strong red cast on right side, so I want to return the monitor.
    So I already sent uniformity result about Delta E and Delta C to the outsourcing customer service, But looks like they don’t understand it.
    They said to me it is normal.

    I need a way to make them easier to understand about red cast. (not Delta E and Delta C)
    Is there any good way?

    #19072

    radiodus
    Participant
    • Offline

    Red cast makes me crazy, so I will refund it and buy new monitor.

    I was considering to buy LG 32UL950 or 34WK95C, But according to your opinion, They are also likely to have a uniformity problem…

    What about ASUS PA32UC or PA32UCX? It looks like they have uniformity compensation function.

    #19073

    Vincent
    Participant
    • Offline

    And a extremely useless HW calibration software.
    Download it, install (better do it in a virtual machine), take a look. Extremely set of limited presets, no suitable corrections for i1d3. Asus software is the worst among these low cost widegamut models.

    Why do you need a widegamut display?
    Photo+printing?
    Video editing? Need P3 or just Rec709?

    For photo (AdobeRGB) get for a CS2730 or equivalent. There are no reliable low cost solutions in 27″ or 32″. That CS is the cheapest one in 27″ with good quality.

    If you need P3 UHD video there are low cost solutions with VA panels like Benq EW3270U. “Reasonable” uniformity, 3000:1, known backlight for i1d3 colorimeters & DisplayCAL.
    If you use Resolve make a (software) LUT3D and use it. For Premiere unless CC2019 have changed too much over older CCs you’ll need to use sRGB OSD mode + GPU calibration for Rec709 editing, native gamut mode + GPU calibration for P3 D65 editing.

    If you do not need at all a widegamut display… do not buy one. There should be sRGB IPS UHD displays with reasonable QC.

    Calibrite Display Pro HL on Amazon  
    Disclosure: As an Amazon Associate I earn from qualifying purchases.

    #19824

    Xaver
    Participant
    • Offline

    Example of uniformity test  for SW270C ( DisplayCal 3.8.3 ) .  Have you seen such test for CS2730 ?

    • This reply was modified 4 years, 7 months ago by Xaver.
    Attachments:
    You must be logged in to view attached files.
    #19828

    Vincent
    Participant
    • Offline

    Yes, prad.de 5×3 brightness % and deltaC, with and without DUE active, long time ago.

    Did they provide you a handpicked unit? or will we see soon those other “common user” reports from hell like in SW2700PT before returing them to store?

    #19850

    Xaver
    Participant
    • Offline

    Handpicked! Yes, by myself 🙂
    The uniformity report is from the second piece. First one I have returned because too strong bleeding in the top left corner  and magenta, green “mold” was also too much  to me. For the first sample I have no uniformity report from DisplayCAL, only from iProfiler.  Brightness  and  color uniformity  – see files.

    Well, not all is gold that glitters.  PME calibration for SW270C sucs with iDisplay Pro. It is  too green. I guess known issue here for SW271 etc.  HW calibration is probably useless or you have to do extra on top. I have only used  calibration with iProfiler which looks better. I am new to DisplayCAL but I give a try. I have seen here .ccss flattened for SW2700. Should I use this one or any other suggestion for correction?

    Attachments:
    You must be logged in to view attached files.
    #19865

    Vincent
    Participant
    • Offline

    I’m not sure about SW270C backlight, but 99% sure that PME lacks of a correction for it (v1.3.5). My guess is QLED (like SW2700 and all this AUOptronic QLEDs) or WLED PFS (AdobeRGB flavor, Z24x).
    If you measure white obtained with PME with one of those 2 CCSS, it should measure “towards green” from daylight curve if your i1d3 is well beahaved (firmware curves very close to std observer). If it is not so well behaved (based on firmware data) I would say to test “by eye matching”, use Personal/Custom/User OSD mode and one of those CCSS and try to figure if when you set RGB gains to D65 white it is white or not. Use the “whiter” D65 CCSS.
    Further experimentation will need the help of a i1pro2-like spectrophotometer configured at 3nm (High res) in order to find if WLED PFS spikes are present in red.

    Of course, your i1d3 could be very well behaved (based on FW data, close to std observer) and the wrong  RGBLED correction used by Benq would de be good enough in that situation. In that case, WP errors after PME calibration are caused by calculation itself (wrong assumtions or oversimplifications). Dell’s DUCCS, Benq’s PME, PM, Viewsonic Colorbration and all this stuff are known to over use these simplifications, or modify brightness after modifying LUTs, etc…
    If that is happening you can use PME to emulate gamut/gamma and after that make slight WP corrections in GPU with a traditional calibration in DisplayCAL on top of your HW cal.
    Up to 6dE from your desired white should be GPU correctable in displayCAL keeping > 95% unique grey levels. As long as GPU allows high bitdepth LUTs (dithering is desireable too) it should be visually equivalent to HW cal. I’ve seeing it, it should work OK for you, as long as GPU is able to provide bandless calibration: nvidias over DP/HDMI with > 8bpc & AMDs. For intel GPUs bad news AFAIK, but i’ve not tested latest ones.

    • This reply was modified 4 years, 7 months ago by Vincent.
    #19946

    Xaver
    Participant
    • Offline

    Thank you for suggestions.

    The result with HP DreamColor Z24x G2 ccss in the attachment. I have set RGB manually via monitor OSD.

    Speed was set to High. I guess that better results can be achieved for greys with medium or low speed?

    Attachments:
    You must be logged in to view attached files.
    #19950

    Vincent
    Participant
    • Offline

    Speed was set to High. I guess that better results can be achieved for greys with medium or low speed?

    Yes, but open a gradient in a non color managed app to check if your current results are acceptable or not.
    It looks like your current grey coloration should be visually noticeable around grey 64 & 153, “smooth” gradient but “tinted”.
    If it is not smooth, then it looks like you have GPU issues related to calibration.

    AFAIK configuration is:
    high = up to 24 patches in final iteration?
    medium = up to 48 patches in final iteration?
    slow = up to 96 patches in final iteration?
    so it results in ~10, ~5 & ~3 RGB value jump between measurements in final iteration. The shorther the jump (more patches used for grey ramp), the less likely some grey tint will go unnoticed.

Viewing 12 posts - 1 through 12 (of 12 total)

You must be logged in to reply to this topic.

Log in or Register

Display Calibration and Characterization powered by ArgyllCMS