Screen calibration and color measurements

Discussion in 'Digital Photography, Home and Portable Electronics' started by qqiiipp, Jun 17, 2016.

  1. qqiiipp

    qqiiipp Guest

    Messages:
    140
    Likes Received:
    1
    GPU:
    ASUS Strix GTX 970
    Hello everyone,
    I recently bought a hardware calibration device(Spyder4) and I'm just curious to know, is there a software that I can use in combination with the hardware device to only measure the colors of my screen while showing the value offsets for each color so I can make adjustments to the monitor OSD? instead of letting the software create a custom icc profile.
    Basically what I would like is to have equal picture quality across all of the inputs in my monitor not just the one i use for the OS.
    Regards.
     
  2. eclap

    eclap Banned

    Messages:
    31,468
    Likes Received:
    4
    GPU:
    Palit GR 1080 2000/11000
    I'll pm yasamoka with a link to this thread and he'll hopefully be able to sort you out.
     
  3. qqiiipp

    qqiiipp Guest

    Messages:
    140
    Likes Received:
    1
    GPU:
    ASUS Strix GTX 970
    Thanks in advance dude, appreciated.
     
  4. yasamoka

    yasamoka Ancient Guru

    Messages:
    4,875
    Likes Received:
    259
    GPU:
    Zotac RTX 3090
    Hey there :)

    Download and install HCFR. It supports the Spyder4.

    Since you want to tweak your monitor using the OSD first, so that the color adjustments are OS-independent, you might want to do the adjustments in the following sequence.

    Perform all sensor measurements in a dark room, particularly since you're using the Spyder4. It does not keep all ambient light out.

    In HCFR, open preferences and increase the number of grayscale measurements to something high yet reasonable (20-50 is a good range). Open a test pattern window, switch to white, move the window to the center preferably, place your sensor onto the window, and perform a continuous measurement. The software will then keep reading from the sensor and will display the luminance of Red, Green, and Blue separately, along with brightness and other quantities.

    1) Maximize Contrast Ratio: Contrast ratio is the difference between the brightest white and the darkest black. Open this contrast test pattern page and place it somewhere on the monitor you are calibrating. Make sure it does not interfere with the sensor measurement area. Raise the contrast level in the OSD and observe the brightness (in cd/m2) rising. Keep doing so until the brightness stabilizes or you get clipping in light color patches on the test pattern page. Then, stop the continuous measurement and perform gamma measurement in HCFR. If you find that the gamma is unstable, particularly towards 100% gray (white), then lower contrast until gamma is stable. Low gamma towards 100% gray implies clipping, since the change in output (display) corresponding to the change in input (requested color) is low or none. You'll want to find a contrast level that does little to no clipping (preferably) while keeping the contrast ratio high. You will also find a contrast ratio measurement option in HCFR.

    Good monitors will generally stop clipping AND maximize contrast ratio at the same contrast level. Some monitors do worse and attain a maximum contrast ratio at contrast levels that involve a lot of clipping. Strike a balance much closer to limiting clipping rather than maximizing contrast ratio, or rather, be very strict in avoiding any sort of clipping. Usually, with worse monitors, you will find that clipping is a fluid matter; you will not be able to determine whether or not there is clipping, but rather, how far can you go in fixing gamma towards 100% gray (white).

    2) Adjust Gamma: You will want your gamma at 2.2 or another target, as you wish. 2.2 or sRGB is the standard for the sRGB gamut. Some monitors provide a "gamma" option or slider. Tweak and perform the gamma measurement until you get the target gamma you are looking to calibrate at.

    3) Adjust Color Gamut: Some advanced monitors ship with *working* color gamut options. You might get a choice between sRGB, AdobeRGB, etc... You might even get a slider or similar. RGB level sliders are NOT for the color gamut. They are for color temperature. Color gamut is determined by the wavelengths of the Red, Green, and Blue primaries used. RGB level sliders tweak luminance of the color channels separately. Monitors that allow gamut remapping generally tweak the triplet in tandem such that the wavelength of, say, the Red primary, is modified by adding Blue and Green components to it even when a purely Red color is requested to be displayed.

    Tweak and perform the saturation measurement for the color you are tweaking until the measurement falls in the triangle corresponding the gamut you are calibrating at. Generally, you will be working with sRGB for consumer content (movies, games, Internet). If you are working with another color gamut, choose from preferences.

    4) Adjust color temperature: Tweak RGB level sliders (down) until you get closest to 100% on all color channels with the continuous measurement on a white test pattern. Usually, raising RGB level sliders up will cause clipping. On some monitors, raising all from 50 to 100 will not cause any changes. Stick to going down from the default. If in doubt, perform the measurements mentioned in (1) (visually and using the sensor).

    On some monitors, tweaking sliders that should have independent effects on the display might affect other factors that do not seem to be related. In adjusting gamma, clipping might occur, for instance. Usually, changing RGB sliders is safe. Check for such changes as you're doing adjustments by performing the measurements related to the other adjustments. Once in a while, check if you're getting any clipping, if the contrast ratio is falling, if gamma is going haywire.

    If you now have a neutral grayscale, then change your monitor's brightness to a suitable level. Some monitors do not have backlight control so lowering brightness means lowering white luminance and thus contrast ratio. Watch out for that. Many monitors use PWM at visible frequencies to control backlight brightness. PWM might be used anywhere below 100% brightness, or below a specific level (e.g. 20%). It's used because it's cheaper than voltage adjustment and because lowering the voltage, depending on the backlight technology, might change the output characteristics of the backlight (changing color gamut, color temperature, and uniformity). PWM might give you headaches if you're sensitive to it, and can cause motion artifacts.
    Check for PWM here. Try other tests that are appropriate as well, such as the UFO test.

    Newer monitors are shipping less and less with PWM, which is good. Gaming monitors now avoid PWM.

    If you do not have a neutral grayscale, you might want to change the brightness to a somewhat higher level than what you use since the luminance of one or more of the color channels will be lowered after calibration (e.g. you have a green tint so after calibration a good chunk of *useless* brightness will be lost).

    Good monitors do not change their color output characteristics with brightness. Contrast ratio, gamma, and color temperature remain ~constant. For very good monitors, uniformity also remains consistent.

    5) Generate color profile: After there are no more options in the monitor's OSD to tweak, you will want to generate a color profile. Download and install dispcalGUI and configure the settings for calibration as you please. Do be warned that the Spyder4 is very slow in reading dark grays so you might want to strike a balance between the number of measurements and the time taken for calibration. It can take up to 4 hours on a Spyder4 if you go berserk with the testchart size. Optionally, choose interactive measurement just to make sure you have a neutral color temperature (6500K is the reference) before proceeding with the software calibration process.

    Standard settings are: Gamma 2.2 / sRGB, 6500K whites. Raise / lower the sliders available for high quality calibration.

    The difference between Gamma 2.2 and sRGB is that sRGB bumps up the first few dark grays such that they are more visible which helps with visibility of dark objects in games and movies. (Some? Most?) content tends to be produced with sRGB gamma in mind, so you might want to check for that and make up your mind before you calibrate. I use 2.2 but I realize that sRGB is a good choice.

    Do not get confused between the sRGB color gamut and sRGB gamma. They are defined for the same standard, which is sRGB, but here they are used interchangeably depending on the context.

    After the calibration process is done, you will have a color profile generated (and applied) for your monitor in the AppData/dispcalGUI folder. Do NOT install the color profile as that would use Windows Color Management. It does a somewhat poor job of applying high bit-depth color profiles and does not preserve color profiles if they are reset by games or other 3D applications. You might want to use my software, Color Sustainer (forum thread), or the dispcalGUI color profile loader.

    Hope this helps. If I have missed anything, I will add it to this post later.
     
    Last edited: Jun 20, 2016

  5. qqiiipp

    qqiiipp Guest

    Messages:
    140
    Likes Received:
    1
    GPU:
    ASUS Strix GTX 970
    Hello yasamoka.
    Thanks for your time spent to reply with such a level of detail.
    I'm using the Dell U2414h which is also a PWM free monitor and it has a backlight control which will all come in handy when my Spyder4 arrive. :)
    Many thanks once again.
     
  6. yasamoka

    yasamoka Ancient Guru

    Messages:
    4,875
    Likes Received:
    259
    GPU:
    Zotac RTX 3090
    Oh, cool!

    Here's a review so you know what to expect from calibration:
    http://www.tftcentral.co.uk/reviews/dell_u2414h.htm

    Why did you order the Spyder4 in specific? If you have a higher budget, go for the Colormunki Display or the X-Rite i1 Display Pro, as these are better sensors. At the very least, isn't the Spyder5 available at affordable prices?
     
  7. qqiiipp

    qqiiipp Guest

    Messages:
    140
    Likes Received:
    1
    GPU:
    ASUS Strix GTX 970
    I found the Spyder4 Elite on ebay at a really affordable price.
    Of course I did had some knowledge based on research on the internet before buying it, is not like I jumped naked in the war or something.
    As far as I know the only difference from the Express to Pro, to Elite, is mostly the software, but still, it's a great addition.

    I though about X-rite as well, but after making some research I found equally positive and negative comments and reviews for both products(x-rite and datacolor) so in my opinion it seams like nothing but a fanbase thing, but the fact is, they are both equally good products.

    I found the Syder4 on a better offer then any x-rite available, so I got it, no other particular reason, but still a good one :)
     
  8. yasamoka

    yasamoka Ancient Guru

    Messages:
    4,875
    Likes Received:
    259
    GPU:
    Zotac RTX 3090
    The Datacolor and X-Rite software pale in comparison to the free and open-source HCFR, ArgyllCMS, and dispcalGUI.

    I have both the Spyder4Express and the X-Rite i1 Display Pro. The Spyder4 takes very long to measure dark patches and does not report black levels accurately. Also, it is much slower than the i1DP and less consistent. Sometimes, the Spyder4 used to leave slight tints and calibrations would vary slightly from one time to the next. The i1DP does not get affected by ambient light while the Spyder4 does.

    The Colormunki Display and the i1DP very likely use the exact same sensor but the i1DP is officially supported for 3rd-party software and drivers while the Colormunki Display required reverse-engineering. Both are supported in the ArgyllCMS library that dispcalGUI and HCFR use.

    Unless you got the Spyder4 at an extremely competitive price. It's still a fine sensor and does fine for most color adjustment. You'll still get results on your monitors that are much better than default.

    It's not a fanbase thing. Most color sensor reviews are subjective. They have no control group, and most do not even use multiple sensors. The proper approach is to compare the calibration results of each using a very high-end spectrophotometer, but the vast majority of reviews, obviously, cannot afford such a device ($10000+). Among the very few reviews I've found, the i1DP is faster, more consistent, and more accurate. The only sensors that score higher than the i1DP are $500+ and $1000+ and that's what makes it such a special colorimeter.
     
  9. qqiiipp

    qqiiipp Guest

    Messages:
    140
    Likes Received:
    1
    GPU:
    ASUS Strix GTX 970
    It's something with the Spyder sensor, and not that much with the sensor it self in terms of hardware, but rather with the plastic bits on the device itself that are meant to isolate the sensor from any possible leakage of ambient light.
    Which is why it's also highly recommended to calibrate in a pitch black(if possible) room.
    I know the i1D Pro is a great device, but it's also twice as expensive, and it's not an issue for me if some device takes longer or if it's much quicker then other, as long as it does the job I'll be happy with it :)
     
  10. Lavcat

    Lavcat Master Guru

    Messages:
    552
    Likes Received:
    44
    GPU:
    Radeon 7900 XTX
    Could I get help with a similar problem? I'm bringing my ten year old Nikon 9000ED film scanner back in service and I had hoped to calibrate my Samsung 275t monitor for photographic work. (I would also like to optimize the monitor for Blu-ray viewing without CMS.)

    For calibration I have an X-Rite i1Display Pro and the X-Rite i1Profiler software. I also have the freeware version of CalMAN Studio and HCFR 3.4.2.

    I have a professional background in photonics but I'm not ashamed to admit I am a little lost. I read yasamoka's excellent writeup above but got stuck on: "In HCFR, open preferences and increase the number of grayscale measurements to something high yet reasonable (20-50 is a good range)." I can find no such setting in preferences or anywhere else.

    Another problem, unlike with the other tools I have, I cannot find an Adobe 1998 setting in HCFR.

    Any help would be very much appreciated.
     

  11. yasamoka

    yasamoka Ancient Guru

    Messages:
    4,875
    Likes Received:
    259
    GPU:
    Zotac RTX 3090
    Hey there!

    [​IMG]

    [​IMG]

    True, HCFR does not include an Adobe RGB setting. To get the Adobe RGB gamut as a reference in HCFR, choose "Custom" for Color Space and enter the chromaticity coordinates of Adobe RGB (from here):

    x y
    Red: 0.6400 0.3300
    Green: 0.2100 0.7100
    Blue: 0.1500 0.0600
    White: 0.3127 0.3290 (D65)

    As you may have noticed, the chromaticity coordinates of Adobe RGB differ from sRGB only in Green, when both are using a whitepoint corresponding to D65 (6500K).

    This explains why graphically:
    [​IMG]

    Good luck!

    EDIT: just noticed this reply.

    Black measurements using the Spyder4Express in a dark room did not differ much (if at all). The sensor fundamentally has trouble measuring deep blacks.

    However, as you have said, the i1DP is twice the price and as long as the main difference according to your needs is speed, and speed doesn't matter, then by all means.
     
    Last edited: Aug 15, 2016
  12. Lavcat

    Lavcat Master Guru

    Messages:
    552
    Likes Received:
    44
    GPU:
    Radeon 7900 XTX
    Thank you so much, yasamoka! You got me over the hurdle. Not all is perfect yet but I'll take up with the measurements again tomorrow.
     
  13. Lavcat

    Lavcat Master Guru

    Messages:
    552
    Likes Received:
    44
    GPU:
    Radeon 7900 XTX
    I found that in the AMD Crimson settings there is an option to set the color temperature by a slider or automatically. I assume "automatically" means from the EDID. Though I have learned to fear ever to assume.

    Toggling this setting makes the display look very, very different.

    For initially setting up the monitor I would like to disable all CMS or correction. I think I have Windows CMS disabled. What should I be doing?
     
  14. Lavcat

    Lavcat Master Guru

    Messages:
    552
    Likes Received:
    44
    GPU:
    Radeon 7900 XTX
    I learned the hard way that gamut settings in HCFR don't stick. Last time I worked with the program I had Adobe 1998 configured. Today after a lot of measurements I noticed the setting was back on REC 709.

    Probably just as well I leave the setting on REC 709 so the monitor is best adjusted for watching Blu-rays from the Blu-ray player. Then I should be able to create a Windows .icc profile set to Adobe 1998 for photo editing.

    At the moment my average gamma is 2.19, average grayscale dE is 0.90, max grayscale dE is 1.89.

    However when I measure primary and secondary colors, the average dE is 66.77. This number is so high that I expect operator error on my part. Just for fun I tried Adobe 1998. The average dE went up to 68.6. Even the white target (255, 255, 255) dE is 34.4. But when I measure white dE from the gray scale test, the same white target (255, 255, 255) dE is 1.3.

    What could I be doing wrong?


    Edit: when I use CalMAN to measure all colors the average dE is 3.6 and the max dE is 9.8.
     
    Last edited: Aug 19, 2016
  15. Lavcat

    Lavcat Master Guru

    Messages:
    552
    Likes Received:
    44
    GPU:
    Radeon 7900 XTX
    After some testing the HCFR colorspace setting seems to be properly sticking now, even after a reboot. Though I'm still reasonably certain I had had it set to custom last time I used the program and that it started off at REC 709 today.
     

Share This Page