- This topic has 74 replies, 4 voices, and was last updated 1 year, 3 months ago by
.
-
Posts
-
Will do, thank you!
Should I perform a test with it using, for instance, 79 patches, or would going with the same amount of patches as my spectrally corrected ICC (175, to be exact) be better for trial and comparison purposes?
I ran a calibration with the same settings as the spectrally corrected ICC, but with the matrix correction loaded instead, of course, and 79 patches instead of 175. The resulting ICC is attached below.
Visually, the CCMX colors are more blue than the CCSS’ (the CCSS producing whites with a very slight creamy hue), but less blue than my display’s colors uncalibrated. Also, this is not the best method for comparing, I know, but next to a family member’s more modern screen (albeit uncalibrated) and my phone’s screen as well, the CCSS colors appear more accurate than the CCMX’s do.
Additionally, if you look at the other attached image, you will see that the CCMX calibration resulted in very low coverage percentages. For instance, the CCMX reported 44.1% sRGB coverage, whereas the CCSS reported about 72% sRGB (when 175 patches were used). And even the 79-patch calibration with no correction reported better coverage at a little over 69% sRGB.
Does all of this indicate that the CCMX is no good?
Looks like I made a mistake when calculating the CCMX. Sorry about that, uploaded a corrected version.
No worries!
So, the results of the new CCMX (ICC attached below) seem much more promising. Compared to the CCSS, which has a slight yellow tint, the CCMX calibration produces whites that appear more white and greys/blacks that, likewise, appear more correct. The CCMX is cooler than the CCSS, but warmer than my screen’s uncalibrated colors. Nice middle ground I think.
And compared to my phone and my family member’s more modern screen, the CCMX colors match well for the most part, although for cool colors, the CCSS is still more accurate.
Lastly, the coverage and errors for the CCMX (shown left in the attached image) are comparable to, and in the case of some of the error values, even better than, that of the CCSS (shown right).
So, if you find that the new CCMX appears to be in order, do you think I can try a more extensive calibration with, say, 1155 patches?
I wasn’t sure if you were notified because the page glitched when I submitted my reply, but just in case, you will find the information about the corrected CCMX results in my post above.
Glad to see it helped.
I don’t expect a 1155 patch profile to be significantly different, but feel free to try.
Okay, I will give it a go, then, when I have the time. Thank you so much for all of your help with the correction.
By the way, do you mind if I ask how one can make a CCMX?
I was wondering how he made your ccmx. I imagine it was math and science and haveing a ccmx already. To make a ccmx you need a spectrometer.
I was wondering how he made your ccmx. I imagine it was math and science and haveing a ccmx already. To make a ccmx you need a spectrometer.
No, you need a reference device that gives you “reference” CIE XYZ tor R G B primaries, whatever this device is.
A CCSS needs an specrophotometer because it contains spectral data, not just CIE XYZ acting as “reference”
By the way, do you mind if I ask how one can make a CCMX?
Although ArgyllCMS may use a sligthy different algorithm, google “four color correction matrix” and you’ll see the maths. You an implement them even in excel i assume.
Anyway, a 3×3 XYZ to XYZ (CCMX or compatible) is done from a set of XYZ “reference measurements (primaries) and a set of CIE XYZ that needs to be corrected.
Hi Ben, no luck so far with creating a verification report using ArgyllCMS, if it truly does have such a feature, but I do have the results of a 1155-patch calibration (using Danij’s matrix correction) that I can share.
So, all of the calibrations shown in the attached image were made using the same settings with the following exceptions: the profile on the left was made with the CCMX, 1155 patches, and high speed; the middle profile used the CCMX, 79 patches, and medium speed; and the rightmost profile was the original, which used the CCSS, 175 patches, and medium speed. I also reran the 1155-patch calibration on medium speed (no screenshot, unfortunately), hoping for lower error and better coverage values, but while the error values improved ever so slightly, the coverage values actually ended up being a tad worse. And as you can see from the image, the high speed 1155-patch calibration already didn’t improve things much VS the 79-patch calibration. Oh well.
Also, I was unable to select the ‘Update Calibration’ option on DisplayCal’s Calibration page at any time, as it was grayed out. Would you happen to know why this might be?
Thank you, Vincent. I believe I managed to find the information to which you referred. However, since DaniJ mentioned that a 3×3 matrix correction was used, I’m not sure that the same calculations can be applied for creating another such correction (although I am somewhat confused by them in any case).
By the way, does a screen’s angle when it is measured and the amount of time said screen is given to first warm up change the accuracy from one calibration to the next very much? I have been allowing my colorimeter to rest on and warm up with my screen for at least 30 minutes before calibrating, and I am wondering if I should increase this time to an hour or more and/or be more consistent with starting the calibration immediately after the designated amount of time has passed.
DaniJ: Well, that sounds simple enough, but I’m afraid I have not yet been successful in figuring out how to replicate the values of the CCMX you provided.
In the absence of a spectrometer, are matrix corrections created from one’s display’s EDID data the next best type of correction?
I’ve only got the one display at the moment, which, thanks to you, now has a decent correction, but I was hoping to find out about the process for when I purchase a new computer and/or monitor and need a good correction. So, if it’s no trouble and EDID-based CCMXs really are ideal, could I ask for more details?
Also, if you’re interested, I have posted the results of the 1155-patch CCMX calibration in a response to Ben above.
