Laptop screen profiling for photo work

Home Forums Help and Support Laptop screen profiling for photo work

This topic contains 22 replies, has 3 voices, and was last updated by  Florian Höch (@fhoech) 4 years, 4 months ago.

Viewing 15 posts - 1 through 15 (of 23 total)
  • Author
    Posts
  • #423

    Greetings!

    A big thank you to Florian for the amazing work put into dispcalGUI – it helps countless people attain a higher standard in their imaging!

    I have been studying the issue of color management for some time, I have read the dispcalGUI documentation and many other sources, but when it comes to the practical side of things I am left with a series of questions as I am one of those people who fit at the periphery of two use cases. Actually I think most people who develop raw photos have laptops or typical home LCDs so I’m not alone in this boat, and once I have this stuff figured out I will document it.

    I have a high-end laptop on which I do my photo work, and of course I want to have a profiled screen and color-managed workflow. The screen does not have a wide gamut, so I use sRGB. I am using dispcalGUI-2.9.1.1 with argyllCMS-1.6.3, nvidia-drivers-340.76. I use a HueyPro colorimeter.

    13:47:07,878 Uncalibrated response:
    13:47:07,880 Black level = 0.1308 cd/m^2
    13:47:07,882 50% level = 21.90 cd/m^2
    13:47:07,884 White level = 114.21 cd/m^2
    13:47:07,891 Aprox. gamma = 2.38
    13:47:07,892 Contrast ratio = 873:1
    13:47:07,893 White chromaticity coordinates 0.2713, 0.2945
    13:47:07,893 White Correlated Color Temperature = 10680K, DE 2K to locus = 10.9
    13:47:07,894 White Correlated Daylight Temperature = 10679K, DE 2K to locus = 7.9
    13:47:07,895 White Visual Color Temperature = 9149K, DE 2K to locus = 10.2
    13:47:07,895 White Visual Daylight Temperature = 9590K, DE 2K to locus = 7.4
    13:47:07,896 Effective LUT entry depth seems to be 8 bits

    1- dispcalGUI has a setting for “Laptop” and one for “Photo”. I use a laptop on which I want to do develop raw photos with accurate color. Which one should I use? What changes should I make on top of the defaults?

    2- I read it is recommended to use 5000K for photo work (or to match printer output), but I guess this advice is for users of wide-gamut screens. Mine is not. I don’t print, I just want images to be accurate on screen and for whites to look white (not orange-ish and not blue-ish). Should I use 5000K or 6500K? Please see the my screen measurements to help you answer the question. More than just recommending one or the other, I am interested in why one or the other.

    3- Should I set a whitepoint temperature or rather the chromaticity coordinates which I get from “report on uncalibrated display device”? Am I correct in assuming that “as measured” would be the same as “chromaticity coordinates”?

    4- The white level gets detected between 113 and 117. Should I set “as measured”, or set something a few cd/m2 less, say 110? The documentation says, “NOTE that some LCD screens behave a little strangely near their absolute white point, and may therefore exhibit odd behavior at values just below white. It may be advisable in such cases to set a brightness slightly less than the maximum such a display is capable of.” How much less is “slightly less”?

    5- I use the sRGB color space for the screen but it is reported as using a gamma of around 2.38, so should I set “tone curve” to gamma 2.2, 2.38, sRGB or as measured?

    6- Both “Photo” and “Laptop” skip the ambient light level measurement, should I measure it or skip it?

    7- If I want to develop raw photos on the laptop (I use “XYZ LUT + matrix”), which measurement chart should I use: “Laptop (Gamma 2.2)”, “Photo (D50, Gamma 2.2)”, “Default test chart for LUT profiles” or “Large test chart for LUT profiles, optimized for sRGB”?

    8- Which profile type is recommended for Nvidia/KDE5 (Plasma 5) running Kolor-Manager and Oyranos? I tried “XYZ LUT + swapped matrix” and colors looked normal so I guess LUT must be working, so should I stick to “XYZ LUT + matrix”?

    9- I saw a video where the Huey stood next to the monitor and would automatically apply the necessary changes (in MS Windows) in real-time as the light in the room changed, is this possible in Linux?

    I hope you find the time to answer these, it would mean much to me. Apologies if you feel these questions have already been answered in the documentation, but I have read the whole doc page end to end at least four times by now, the first time would have been some three years ago, and I’m still unclear on these points. I must have tried over 30 different calibrations, never sure that what I set was right. I’m sure others would benefit from your answers and perhaps you can use the questions to know what issues in the documentation to expand on.

    With kind regards

    #426

    I have attached my settings which include the calibration and profile and also the measurement report.

    “Measured vs. assumed target whitepoint ΔE*00” is 3.6. Is this a problem, and if so, how do I fix it?

    Attachments:
    You must be logged in to view attached files.
    #424

    Florian Höch (@fhoech)
    Administrator
    • Offline

    Hi,

    1), 2), 3) The “Laptop” preset is better suited for Laptops because those usually don’t have a way to adjust RGB balance (interactive display adjustment).

    4) Set the backlight of your Laptop to a level that allows you to use it without causing eye strain (usually that means roughly matching ambient light levels). Then use “As measured” in dispcalGUI. 110-120 cd/m2 is usually good for a not too dim and not too brightly lit environment.

    5) It is not clear to me what “I use the sRGB color space for the screen” means. Does the screen have several modes? That’s unusual for a Laptop.

    6) Skip (in fact I would strongly recommend to disable “Show advanced options” in the “Options” menu and not concern you with any of those settings).

    7) Larger charts may offer better accuracy, but I would first try the “Laptop” chart.

    8) Depends if you’re interested if a program falls back to the matrix tags, in which case you would use “XYZ LUT + swapped matrix”.

    9) It’s a gimmick. For anything color-critical, you should control the room lighting, not the other way around.

    #429

    Florian Höch (@fhoech)
    Administrator
    • Offline

    “Measured vs. assumed target whitepoint ΔE*00” is 3.6. Is this a problem, and if so, how do I fix it?

    Not necessarily, it’s probably more an indication that the Huey is not able to accurately read the whitepoint than anything (I somehow doubt it’s really in the 10K range, that would be quite blueish. Laptops I’ve seen/used were more like roughly 5000-7000K).

    • This reply was modified on 2015-04-26 19:00:56 by fhoech.
    #430

    Hi Florian,

    Could you tell me what is the best calibrate target for a laptop.

    Brightness adjustment is the only setting that laptops have, but some display drivers can adjust Brightness/Contrast/Gamma/RGB, how should I play with these setting before I start the calibration?

    Are the default values (Whitepoint as measured/2.2 Gamma) with the Laptop preset the recommended setting for laptop calibration? Is it a bad idea to calibrate a laptop to sRGB? Is it better to leave the white point to native, rather than trying to make it to D65 so that minimun work is done to the calibration curves? Can I use the display driver to adjust the whitepoint? 2.2 is set for the gamma. Should I choose a higher gamma value e.g. 2.4 for lower luminance in a dim environment e.g. 80 cd/m2? What are the threshold?

    I also want to clarity the following:
    “XYZ LUT + swapped matrix” only allows programs to use the LUT or it will display wrong colours.
    “XYZ LUT + matrix” allows programs which do not support LUT to use the matrix.
    Which kind of matrix is this? 1xGamma+MTX/3xGamma+MTX/1xCurve+MTX/3xCurve+MTX?

    Many Thanks
    Denny

    Attachments:
    You must be logged in to view attached files.
    #432

    Florian Höch (@fhoech)
    Administrator
    • Offline

    Hi Denny,

    […] some display drivers can adjust Brightness/Contrast/Gamma/RGB, how should I play with these setting before I start the calibration?

    These adjustments should not (and cannot) be used. They change the videoLUTs of the graphics chip, but dispcal needs exclusive access to them.

    Are the default values (Whitepoint as measured/2.2 Gamma) with the Laptop preset the recommended setting for laptop calibration? Is it a bad idea to calibrate a laptop to sRGB?

    There’s usually no reason to change defaults.

    Is it better to leave the white point to native, rather than trying to make it to D65 so that minimun work is done to the calibration curves?

    I’d say so, unless the perceived hue of the native whitepoint is far off.

    Can I use the display driver to adjust the whitepoint?

    No. See above the comment about videoLUTs.

    Should I choose a higher gamma value e.g. 2.4 for lower luminance in a dim environment e.g. 80 cd/m2?

    The calibration gamma only affects the appearance of non color managed content, and is usually best left at default.

    “XYZ LUT + swapped matrix” only allows programs to use the LUT or it will display wrong colours.
    “XYZ LUT + matrix” allows programs which do not support LUT to use the matrix.

    Yes.

    Which kind of matrix is this?

    3xGamma+MTX

    #433

    Hi Florian,

    Thanks very much for your reply.

    I have tested a three laptop displays and an LCD TV connected to a computer. Do you think I can calibration them with native whitepoint or should I use D65? Which mode is better for the LCD TV?

    Display A
    Black level = 0.3229 cd/m^2
    50% level = 15.94 cd/m^2
    White level = 86.93 cd/m^2
    Aprox. gamma = 2.45
    Contrast ratio = 269:1
    White chromaticity coordinates 0.3119, 0.3269
    White Correlated Color Temperature = 6567K, DE 2K to locus = 3.8
    White Correlated Daylight Temperature = 6568K, DE 2K to locus = 1.0
    White Visual Color Temperature = 6430K, DE 2K to locus = 3.6
    White Visual Daylight Temperature = 6603K, DE 2K to locus = 0.9
    Effective LUT entry depth seems to be 7 bits

    Display B
    Black level = 0.0876 cd/m^2
    50% level = 12.18 cd/m^2
    White level = 75.28 cd/m^2
    Aprox. gamma = 2.63
    Contrast ratio = 859:1
    White chromaticity coordinates 0.3101, 0.3162
    White Correlated Color Temperature = 6770K, DE 2K to locus = 3.2
    White Correlated Daylight Temperature = 6777K, DE 2K to locus = 7.2
    White Visual Color Temperature = 6916K, DE 2K to locus = 3.1
    White Visual Daylight Temperature = 7155K, DE 2K to locus = 6.9
    Effective LUT entry depth seems to be 8 bits

    Display C
    Black level = 0.1093 cd/m^2
    50% level = 18.23 cd/m^2
    White level = 78.81 cd/m^2
    Aprox. gamma = 2.11
    Contrast ratio = 721:1
    White chromaticity coordinates 0.3117, 0.3447
    White Correlated Color Temperature = 6446K, DE 2K to locus = 13.0
    White Correlated Daylight Temperature = 6439K, DE 2K to locus = 10.3
    White Visual Color Temperature = 5964K, DE 2K to locus = 12.7
    White Visual Daylight Temperature = 6087K, DE 2K to locus = 10.0
    Effective LUT entry depth seems to be 8 bits

    LCD TV with CCT mode 1
    Black level = 0.2300 cd/m^2
    50% level = 16.53 cd/m^2
    White level = 76.91 cd/m^2
    Aprox. gamma = 2.22
    Contrast ratio = 334:1
    White chromaticity coordinates 0.3123, 0.3402
    White Correlated Color Temperature = 6448K, DE 2K to locus = 11.0
    White Correlated Daylight Temperature = 6443K, DE 2K to locus = 7.8
    White Visual Color Temperature = 6049K, DE 2K to locus = 10.6
    White Visual Daylight Temperature = 6181K, DE 2K to locus = 7.5
    Effective LUT entry depth seems to be 10 bits

    LCD TV with CCT mode 2
    Black level = 0.2305 cd/m^2
    50% level = 17.08 cd/m^2
    White level = 78.35 cd/m^2
    Aprox. gamma = 2.20
    Contrast ratio = 340:1
    White chromaticity coordinates 0.3022, 0.3316
    White Correlated Color Temperature = 7063K, DE 2K to locus = 11.7
    White Correlated Daylight Temperature = 7055K, DE 2K to locus = 8.7
    White Visual Color Temperature = 6525K, DE 2K to locus = 11.3
    White Visual Daylight Temperature = 6677K, DE 2K to locus = 8.4
    Effective LUT entry depth seems to be 10 bits

    Many thanks
    Denny

    #434

    Florian Höch (@fhoech)
    Administrator
    • Offline

    Do you think I can calibration them with native whitepoint or should I use D65?

    Use native. I wouldn’t trust the Huey’s accuracy.

    Which mode is better for the LCD TV?

    There are no obvious criteria by which I could decide this. Mode 2 seems to be closer in white color temperature to the others for what it’s worth.

    #435

    Hi Florian,

    Thanks for reply. I am using i1 Display Pro. In that case, is it better to use D65 over native?

    Thanks
    Denny

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

    #436

    Florian Höch (@fhoech)
    Administrator
    • Offline

    Thanks for reply. I am using i1 Display Pro. In that case, is it better to use D65 over native?

    The i1 Display Pro is accurate enough that you can set a target other than native.

    #437

    Hi Florian,

    I have done two calibrations (LUT+MTX) with the i1D3.

    Native:
    avg dE 0.26 max dE 0.62 RMS 0.29 sRGB coverage & volume 64.7 66.6

    Daylight 6504K:
    avg dE 0.13 max dE 0.78 RMS 0.18 sRGB coverage & volume 67.5 70.3

    The 6504K result has lower dE and bigger sRGB coverage & volume and a better whitepoint, but the maximum levels for the calibration curves has to be limited. The only good thing about the native result is the R, G, B calibration curves remain at 100% (256 levels).

    Which one is consider to be a better result? Since laptops do not have RGB gain control, limiting the maximum levels for calibration curves would reduce the number of colors and the gradient would be less smooth, right? For the native whitepoint, although the eyes will be adapted to the native whitepoint, would the edited photos still have a color cast compare to the photos edited on a D65 display?

    Hope you can help me to clarify this. By the way, what are the differences between the four temperature below and which I should I look at?

    This is the uncalibrated results
    Black level = 0.1093 cd/m^2
    50% level = 18.23 cd/m^2
    White level = 78.81 cd/m^2
    Aprox. gamma = 2.11
    Contrast ratio = 721:1
    White chromaticity coordinates 0.3117, 0.3447
    White Correlated Color Temperature = 6446K, DE 2K to locus = 13.0
    White Correlated Daylight Temperature = 6439K, DE 2K to locus = 10.3
    White Visual Color Temperature = 5964K, DE 2K to locus = 12.7
    White Visual Daylight Temperature = 6087K, DE 2K to locus = 10.0
    Effective LUT entry depth seems to be 8 bits

    I have attached the two icc profiles.

    Many thanks
    Denny

    Attachments:
    You must be logged in to view attached files.
    #440

    Florian Höch (@fhoech)
    Administrator
    • Offline

    Hi Denny,

    Which one is consider to be a better result?

    The 6504K profile definitely looks slightly better.

    Since laptops do not have RGB gain control, limiting the maximum levels for calibration curves would reduce the number of colors and the gradient would be less smooth, right?

    In practice both profiles are not very different in that regard. Both lose around 40 levels out of 256 (in 8 bit), but with a bit of luck you’ll have a recent graphics chip which applies dithering.

    For the native whitepoint, although the eyes will be adapted to the native whitepoint, would the edited photos still have a color cast compare to the photos edited on a D65 display?

    If you view them side by side, probably yes. Viewed in isolation, less likely.

    #441

    Hi Florian,

    Why would you say both profiles lose around 40 levels out of 256? The native whitepoint profile do have 256 levels, right?

    How do I know if a graphics card is capable of dithering?

    Thanks
    Denny

    #442

    Florian Höch (@fhoech)
    Administrator
    • Offline

    Hi Denny,

    Why would you say both profiles lose around 40 levels out of 256? The native whitepoint profile do have 256 levels, right?

    You can look at the calibration curves of both profiles in the curve viewer. You’ll see that both do strong corrections across the neutral axis.

    How do I know if a graphics card is capable of dithering?

    Looking at a smooth grayscale gradient in an image viewer with color management turned off(!) is usually a reasonable test. I.e. first look at it with the videoLUT set to linear (no calibration) to make sure banding is not in the image itself or introduced by the display (some panels are less than 8 bit and use dither internally to generate intermediate steps). Then load your calibration. If it still looks smooth, the graphic card either dithers, or ouputs a higher bit depth (both should suffice to reduce or eliminate calibration-related banding artifacts).

    #443

    Hi Florian,

    I’ve got huge delta E differences on calibration results on my LCD TV (connected to an AMD 6850 display card) between the new and old version. I was using the same setting and did the two calibrations with an i1D3 one after the other.

    Argyll 1.6.3/DispcalGUI 2.6
    delta E avg 0.6 max 5.1 rms 0.9
    Luminance reported in profile is 82.92 (close to white level from report on calibrated display)
    Current calibration response:
    Black level = 0.2050 cd/m^2
    50% level = 19.39 cd/m^2
    White level = 83.23 cd/m^2
    Aprox. gamma = 2.10
    Contrast ratio = 406:1
    White chromaticity coordinates 0.3121, 0.3286

    Argyll 1.7/DispcalGUI 3.0
    delta E avg 9.9 max 19.6 rms 10.7
    Luminance reported in profile is 60.59 cd/m2 (Is this wrong? White level from report on calibrated display device is 82.71 cd/m2 )
    19:23:21,849 Current calibration response:
    19:23:21,855 Black level = 0.2030 cd/m^2
    19:23:21,857 50% level = 19.20 cd/m^2
    19:23:21,858 White level = 82.71 cd/m^2
    19:23:21,860 Aprox. gamma = 2.11
    19:23:21,861 Contrast ratio = 407:1
    19:23:21,863 White chromaticity coordinates 0.3127, 0.3290

    Which one is correct? The suffix of the attached files shows the version number.

    Many Thanks
    Denny

    Attachments:
    You must be logged in to view attached files.
Viewing 15 posts - 1 through 15 (of 23 total)

You must be logged in to reply to this topic.

Log in or Register

Display Calibration and Characterization powered by ArgyllCMS