Measurement report accuracy

Home Forums Help and Support Measurement report accuracy

Viewing 7 posts - 1 through 7 (of 7 total)
  • Author
    Posts
  • #514

    anonymous SourceForge
    Member
    • Offline

    Hello, I recently bought the spyder 4 colorimeter and calibrated my laptop display (Vaio S15 with IPS screen) using dispcal. The measurement report is surprisingly good and I think the screen looks good aswell. My question has to do with the accuracy of the measurement report depending on the instrument. For example, if the instrument is inaccurate, the calibration will probably be inaccurate too. But, when you do a measurement report, could it report a near perfect calibration because of its inherent inaccuracy?

    #515

    anonymous SourceForge
    Member
    • Offline

    Here are the icc profile and the measurement report if they are of any value. Thanks for your time.

    Attachments:
    You must be logged in to view attached files.
    #517

    anonymous SourceForge
    Member
    • Offline

    ICC

    Attachments:
    You must be logged in to view attached files.
    #519

    Florian Höch
    Administrator
    • Offline

    The measurement report is surprisingly good

    You can usually expect good results like that from a quick test if all went well.

    For example, if the instrument is inaccurate, the calibration will probably be inaccurate too.

    That comes down to repeatability. If repeatability isn’t good, the profile won’t be as accurate, and this would show up in a measurement report (e.g. running several of the same report in succession would likely have fluctuations in dE values, and increased dE).

    But, when you do a measurement report, could it report a near perfect calibration because of its inherent inaccuracy?

    The measurement report is a closed loop test, it won’t tell you anything about the absolute accuracy of an instrument (i.e. compared to a reference instrument).

    ICC

    Looks fine.

    #520

    anonymous SourceForge
    Member
    • Offline

    Thanks for your help. The repeatability of the measurement reports seems good, so I guess the calibration must be at least decent.

    When checking test websites like this http://www.lagom.nl/lcd-test/gamma_calibration.php , the gamma appears a little off in the green 48% test image, being near 2.0. Is that an indication that the calibration is a bit off or if the measurement reports and tone response curves of dispcal say I’m good, I shouldn’t rely on these test images?

    The other thing is banding. My desktop monitor (Dell u2311h) has near zero banding in these tests. Is it normal for some banding to be introduced after calibration? Also, I used some custom settings here and the calibration wasn’t as accurate as the laptop one. For the laptop, I used the laptop(gamma 2.2) settings, only changing colour temperature to 6500K. Is there any recommendation for the desktop monitor?

    #521

    anonymous SourceForge
    Member
    • Offline

    A couple more things I forgot. I haven’t been setting simulation profile to sRGB (I just left it blank) when I was doing measurement reports. Would that make any difference to the results?

    Finally, as per recommendation I found in wiki, I imported spyder 4 corrections and set the mode to the correct one, according to the monitor calibrated each time. However, in the correction field below I only get as options, Auto (None) and None. Sorry for the bombardment of questions and thanks again.

    #522

    Florian Höch
    Administrator
    • Offline

    if the measurement reports and tone response curves of dispcal say I’m good, I shouldn’t rely on these test images?

    Pretty much yes. The images can be useful if you don’t have a measurement instrument, to assess a display visually in uncalibrated/unprofiled state, but they have little use outside of that.

    Is it normal for some banding to be introduced after calibration?

    Typically yes, although there’s many things that can influence banding, not only the calibration curves.

    Is there any recommendation for the desktop monitor?

    For a typical desktop monitor without a focus on a specific task, the “Office & Web” preset is a good start. (Only difference to the “Laptop” and “Default” presets is the color temperature of 6500K).

    However, in the correction field below I only get as options, Auto (None) and None. Sorry for the bombardment of questions and thanks again.

    The Spyder4 corrections show up as additional measurement modes under the respective dropdown.

    • This reply was modified on 2015-05-10 14:28:48 by fhoech.
Viewing 7 posts - 1 through 7 (of 7 total)

You must be logged in to reply to this topic.

Log in or Register

Display Calibration and Characterization powered by ArgyllCMS