Can't import colorimeter corrections for i1 Display Pro

Anonymous
2013-12-09
2013-12-20
  • Anonymous - 2013-12-09

    Hello,

    I can't import colorimeter corrections for my i1 Display Pro (dipcalGUI 1.5.3.1).

    'Auto' does not work at all (i1Profiler 1.5 installed; i1Profiler 1.1 disc in drive).
    When manually selecting the the .edr files in the i1Profiler directory the error message 'No colorimeter corrections could be imported' appears.

     
  • Florian Hoech

    Florian Hoech - 2013-12-09

    Which version of Argyll CMS are you using?

     
    • Anonymous - 2013-12-09

      Thanks for the fast answer. I'm using 1.6.2.

       
      • Florian Hoech

        Florian Hoech - 2013-12-10

        There seems to be a bug in 1.6.2. Please try this (assuming you're under Windows):
        http://www.argyllcms.com/vidtools_win32.zip (32 bit)
        http://www.argyllcms.com/vidtools_win64.zip (64 bit)
        Replace the oeminst.exe in your Argyll_V1.6.2\bin directory with the one from the respective zip archive, then try again.

         
        • Anonymous - 2013-12-18

          Florian

          Using Argyll_V1.6.2_win64 as is or with either of these two vidtools_winxx.zip oeminst does not work neither in DispcalGui nor command line.
          I have removed the 64bit version and installed the 32bit version. Oeminst does not work in the delivered version but works with the above vidtools_win32.

          There is nevertheless a problem. Out of the eight edr files supplied seven are converted and one fails. Log file below. Unfortunately it is the one I need for a new Dell UP3214Q 4k monitor. Dell color calibration solution (hardware cal) uses this RG_Phosphor for that model and it is the one for several other recent RG_Led wide gamut monitors.

          Eric Vincenot

          Working directory:
          C:\ Program Files (x86)\ X-Rite\ Devices\ i1d3\ Calibrations\

          Command line:
          oeminst.exe
          -v
          -Sl
          RG_Phosphor_Family_25Jul12.edr

          oeminst: Error - Reading 'RG_Phosphor_Family_25Jul12.ccss' failed with 'Read line got symbol 'Xˆy' that's too long
          '
          Loading file 'RG_Phosphor_Family_25Jul12.edr'..done
          On entry:
          Got 'RG_Phosphor_Family_25Jul12.edr' size 59880 ftype 0xe ttype 0xf
          'RG_Phosphor_Family_25Jul12.edr' seems to be a i1d3 calibration file or .ccss
          After classification:
          Got 'RG_Phosphor_Family_25Jul12.edr' size 59880 ftype 0x8 ttype 0x4
          After de-archive:
          Got 'RG_Phosphor_Family_25Jul12.edr' size 59880 ftype 0x8 ttype 0x4
          After file extract:
          Got 'RG_Phosphor_Family_25Jul12.edr' size 59880 ftype 0x8 ttype 0x4
          Translating 'RG_Phosphor_Family_25Jul12.edr' (59880 bytes)
          Returning:
          Got 'RG_Phosphor_Family_25Jul12.ccss' size 76706 ftype 0x8 ttype 0x4

          Loading file 'C:/Argyll/Argyll_V1.6.2/ref/CRT.ccss'..done
          No colorimeter corrections could be imported.

           
          • Florian Hoech

            Florian Hoech - 2013-12-19

            Please report the issue at the Argyll CMS mailinglist. Until fixed, the Argyll 1.6.1 oeminst seems to work.

             
            • Anonymous - 2013-12-20

              Florian,
              I did not manage to find where previous versions are archived.
              Eric Vincenot

               
              • Florian Hoech

                Florian Hoech - 2013-12-20

                Just changing the version in a recent download link should work, e.g. http://argyllcms.com/Argyll_V1.6.1_win32_exe.zip

                 
                • Anonymous - 2013-12-20

                  It works. Thanks for the great support.
                  Eric Vincenot

                   
  • Anonymous - 2013-12-11

    Hello,

    first of all thank you for your support.

    For some reason only the x86 version of oeminst.exe seems to work (although I’m using Argyll CMS x64 on Win 8.1 x64). With the x64 oeminst.exe I get a crash (‘oeminst.exe stopped working’).

    I’ve one more question though:
    Before getting the Display Pro I had a Spyder 4. With the Spyder the corrections where listed in the Mode menu.
    After the import the Display Pro corrections are listed in the correction menu as spectrometer corrections (e.g. the white LED correction is named: ‘Spectral: WLED AC LG Samsung’) - the Mode menu still only lists LCD (generic) and Refresh (generic).
    Is this as intended or did something go wrong?

     
    • Florian Hoech

      Florian Hoech - 2013-12-11

      This is intended. The technical reason behind this is for the Spyder 4 the additional measurement modes are stored in a special file (spyd4cal.bin) which only applies to the Spyder 4 while the entries shown under the "correction" dropdown are actual CCMX or CCSS files (a CGATS-based file format used by Argyll CMS) and can be used with any colorimeter supporting said files (e.g. both the i1 Display Pro and Spyder 4 support CCSS files).

       
  • Anonymous - 2013-12-12

    Thank you for the explanations and help.

     
  • Anonymous - 2013-12-13

    Florian, is there a way to transform Spyder4's bin format to CCSS?
    I wonder of what spectral resolution they have and compare them to i1DisplayPro (or even plot them in Excel).

     
    • Florian Hoech

      Florian Hoech - 2013-12-14

      CCSS files (the default ones) have a variable spectral bandwith of 380 up to 780 nm and a spectral resolution of 1 nm. The Spyder 4 calibration also goes from 380 to 780 nm but with a res of 10 nm. I'll post some code later (maybe tomorrow) that will allow you to convert the calibrations contained within spyd4cal.bin to CCSS files.

       
      • Florian Hoech

        Florian Hoech - 2013-12-15

        Attached is a script you can use to create CCSS files from spyd4cal.bin. You need to download and install python to use it. Then from a command prompt:

        C:\Python27\python C:\path\to\spyd4cal_to_ccss.py C:\Users\Your Username\AppData\ArgyllCMS\spyd4cal.bin

        (the above should be all in one line)

        This will write out the CCSS files to C:\Users\Your Username\AppData\ArgyllCMS.
        Each CCSS file will contain three entries (which are all the same, but Argyll CMS does not accept files with fewer entries) of the spectral values of the whitepoint of the given display technology.
        In theory using these CSS files in combination with the "LCD (generic)" measurement mode should give you the same result as using the specific measurement mode of the same name as the CCSS file, but note that this is completely up to you to evaluate since I don't actually have a Spyder 4 that I could try it with.

         
        Last edit: Florian Hoech 2013-12-15
  • Agaporni

    Agaporni - 2013-12-15

    Hello, I'm the anonymous who asked for bin->CCSS transformation.
    Thank you very much for the script and for DispcalGUI

    After plotting in Excel spectral graph I have to say that Spyder 4 spectral corrections suffer from the same inacuracies as a real mesaure from sub 2000 euro spectrophotometer.
    I'm using for comparison spectral samples from a lab grade Minolta spectrophotometer (files bundled with i1DisplayPro) and i1Pro/ColorMunkiPhoto from your database (awesome info!).

    -Blue spike in WLED is missing because the hightest intensity is somewhere arround 455nm, so Z coord will be computed with lower than reality value, as i1Pro or colormunki photo do in their optical resolution (10nm).
    -There is a missing very narrow spike in CCFL wide gamut (old spectraviews, U2410) arround green (445nm), as in an i1Pro spectra (10nm), so lower Y coord than what it's meant to be (as in a i1Pro). There is a little mess with red spikes but at least 1st red spike is captured better than in CCSS database.
    -There is no GB-LED spectral corrections (new Eizo wide gamut, new Spectraview PA242W/PA272W series, new DELL U2413/U2713, etc). From CCSS database (i1Pro/ColorMunkiPhoto) and other external plots like Prad.de Eizo CX240 review, you can see that there is the same issue like in WLED: spike is missing, computed Z coordinate is lower than it should be.
    -RGBLED (HP Dreamcolor) is missing the top of the strong red spike because of 10nm resolution.

    What I am trying to say is that bundled spectral corrections in Spyder4 software (I used Spyder4 Elite) are not very suitable for tipical color acurate enviroments: WG CCFL, WLED mid-high range monitor for web desing, GB-LED.
    An interesting question arises, as common spectrophotometers seem not suitable for color acurate enviromets (same monitor types as abobe): they are missing blue spikes in GBLED technology which leads to Z inacuracies seen in many monitor reviews. In this kind of reviews when some software used a specific colorimeter (Eizo color edge built in, i1DisplayPro for GBLED Dells) they compare to o1Pro readings. Calibrations with colorimeter and specral corrections are more or less D65, when measured with i1Pro are ussually arround 6200K (in some of them no coords were given so I do not know if were in a isoterm curve too off from daylight). Is the missing blue spike reading due to "bad" optical resolution in i1Pro/i1Pro2/ColorMunkiPhoto culprit for this missmatch?
    I've read in ArgyllCMS doc that this program "high resolution mode" with i1Pro (interpolation?) gives you 3.3nm resolution, but arises signal to noise ratio so it's unsuitable for calibration but good for white point spectrum measurements.


    QUESTIONS:

    1-Florian, do you think that a Spyder4 will work with 1nm resolution CCSS obtained from i1DisplayPro edr files? These files may be copyrighted, but as xrite colorimeter drivers are avaliable for download for free, a Spyder4 owner can download them, import edr corrections to CCSS an use this increased resolution to improve his measurements.

    2-Everyone can plot CCSS files from xrite edr files, or you can plot CCSS files from DispcalGUI database and see with your own eyes.

    I think that this leads to at least ask ourselves if spectrophotometers with greater or equal than 10nm optical resolution are suitable for calibration/profiling of "well known" color acurate displays.
    They use the same ilumination and panels:
    -GBLED : all new color accurate wide gamut monitors from NEC, Eizo, Dell, Asus... all they use the same AH-IPS 8+2bit LG panels and same ilumination (but not the same color uniformity, bleeding, 0 dead pixel policies... that kind of "premium" features)
    -WGCCFL: all old color accurate wide gamut monitors from NEC, Dell, Asus... same as abobe with P-IPS/H-IPS 8+2 bit panels from LG.

    For WLED xrite edrs there are 3 spectra (LG IPS/samsung PLS): 2 are the same but one scaled down, the other has an extended red spectrum.
    From DispcalGUI database I downloaded an i1Pro spectrum from a Dell U2412M (LG IPS panel with WLED). Blue spike is missing as expected but its spectrum is like the last one (more red).


    What do you think of these questions?
    I think that i1DisplayPro/ColorMunkiDisplay colorimeters, or even Spyder4 if can use 1nm CCSS, are more acurate BY FAR* for these displays (>70% of color accurate monitors market) than sub 2000 euro spectrophotometers, and these colorimeters are updatable with CCSS files.
    Maybe people are jumping in the wrong boat when loooking for accurate measurement devices for photo edition/web design.
    xrite's spectrophotometers are seen in this graphic arts enviroment as the more acurate devices for monitor calibration. I think that they are not, an i1DisplayPro will be more acurate IF (it's a strong "if") used with the popular monitor models for photo editing.
    Well with a spectrophotomer you can profile your printer, but I'm asking what is better for these kind of "well known" displays.

    Thank for your time and that phython script

     
  • Agaporni

    Agaporni - 2013-12-15

    I forgot to attach the 1nm spectra plots from lab grade Minolta spectrophotometer, so you can compare with i1Pro spectrummeasurements (and strong inacuracies) or with CCSS from Spyder4 driver.

    WLED LG IPS:
    http://i42.tinypic.com/j63ib9.jpg

    CCFL wide gamut (old Spectraview series, U2410...)
    http://i43.tinypic.com/w8amnl.jpg

    GB-LED ( new spectraview series, Dell U2413, Eizo CX240, Asus PA249Q...)
    http://i43.tinypic.com/2lutg6v.jpg

     
  • Florian Hoech

    Florian Hoech - 2013-12-16

    1-Florian, do you think that a Spyder4 will work with 1nm resolution CCSS obtained from i1DisplayPro edr files?

    They should according to the Argyll CMS documentation (CCSS files in general).

    I think that i1DisplayPro/ColorMunkiDisplay colorimeters, or even Spyder4 if can use 1nm CCSS, are more acurate BY FAR* for these displays (>70% of color accurate monitors market) than sub 2000 euro spectrophotometers, and these colorimeters are updatable with CCSS files.

    It sounds reasonable that the better spectral resolution should lead to higher accuracy. I'm not sure how big the difference would be in reality though. Graeme has done some interesting tests comparing the i1 Pro 2 in normal and hi-res mode to a JETI specbos 1211, and if I were to draw any conclusions from that I would expect the errors between a CCSS-corrected colorimeter and a non-lab-grade spectro to be in that ballpark. There are people who have a lot more insight than me though :)
    Maybe this would be an interesting discussion for the Argyll CMS mailinglist?

     

Get latest updates about Open Source Projects, Conferences and News.

Sign up for the SourceForge newsletter:





No, thanks