- This topic has 16 replies, 2 voices, and was last updated 7 years, 3 months ago by
.
-
Posts
-
Hi!
I’m wondering about the logic of resetting to linear curves before calibration. I think I have seen it is good practice to reset via “Reset video card gamma table” before calibrating. Why is that?
My hypothesis is that if we calibrate over an existing calibration, we are getting LUT values relative to that calibration and the OS will think that the values inside the LUT are with regards to linear correction curves. Am I right?
Thus, should it not be necessary to always reset before calibrating? Or does choosing “Update Calibration” writes a new LUT table considering the previous one?
For my second question, suppose I want to compare the calibration profile coming from the Colormunki Display software and the calibration profile from DisplayCal: How can I see the quantitative difference in performance and which is superior? Is it through the measurement report under the verification tab? If so, how should I analyze the data? Moreover, I am having trouble loading the profile into DisplayCal. Colormunki Display gave me a .icm file but how should I load it so that I can test it with the Measurement report?
Third, when I try to compare calibrations using the “Settings” field in the main menu and pressing “Install profile”, I am given a window in which I can check “Preview calibration”: is this Preview Calibration relative to linear curves?
Fourth, suppose I want to calibrate over an existing calibration, how can I ensure DisplayCal uses that calibration? It seems to me that when DisplayCal loads, it resets the calibration. Having Googled the issue, I see that the solution is to select “Current” for settings but doesn’t that only set the values for the Calibration and Profiling tabs? Is the correct way to do this to “Load calibration curves from calibration file or profile…”? How about “Load calibration curves from current display device profile”? Does this load the current default calibration profile for the display?
Finally, what is the correct way of only using Windows to load the calibration profile given by DisplayCal? Is it to locate the produced .icc and in Color Management add that file? Why is this method less reliable than using DisplayCal’s loader?
Thanks!
Calibrite Display SL on Amazon
Disclosure: As an Amazon Associate I earn from qualifying purchases.Also, is it possible to specify a target whitepoint value when doing the measurement report under the verification tab?
Hi,
I’m wondering about the logic of resetting to linear curves before calibration. I think I have seen it is good practice to reset via “Reset video card gamma table” before calibrating. Why is that?
Doing it manually is unnecessary because it’s automatically done at the start of calibration.
My hypothesis is that if we calibrate over an existing calibration, we are getting LUT values relative to that calibration and the OS will think that the values inside the LUT are with regards to linear correction curves. Am I right?
No.
How can I see the quantitative difference in performance and which is superior? Is it through the measurement report under the verification tab?
Correct.
If so, how should I analyze the data?
Lower delta E (especially average) = more accurate. Use a higher number of patches than default to get a better view (e.g. large verification testchart).
Moreover, I am having trouble loading the profile into DisplayCal. Colormunki Display gave me a .icm file but how should I load it so that I can test it with the Measurement report?
Assign the ColorMunki profile to the display (i.e. using the profile loader or Windows color management settings), then choose “<Current>” under “Settings” in DisplayCAL.
Third, when I try to compare calibrations using the “Settings” field in the main menu and pressing “Install profile”, I am given a window in which I can check “Preview calibration”: is this Preview Calibration relative to linear curves?
Yes (note that the automatic dialog right after calibration&profiling previews relative to the previous calibration).
Fourth, suppose I want to calibrate over an existing calibration, how can I ensure DisplayCal uses that calibration?
Youn can’t calibrate over an existing calibration unless you’re updating a calibration that has been previously created by DisplayCAL. But you can profile an existing calibration by setting whitepoint and tone curve on the calibration tab to “As measured”, and disabling interactive display adjustment.
Finally, what is the correct way of only using Windows to load the calibration profile given by DisplayCal?
Open Windows color management settings, go to advanced tab, click “Change system defaults”, on advanced tab enable “Use Windows display calibration”.
Why is this method less reliable than using DisplayCal’s loader?
Because Microsoft doesn’t care, and their loader always quantizes to 8 bit (which can introduce banding) and scales wrongly.
Also, is it possible to specify a target whitepoint value when doing the measurement report under the verification tab?
That wouldn’t make sense – the assumed target whitepoint is always the closest daylight (or blackbody) locus. You can also check the distance to the profile whitepoint in the report.
Yes (note that the automatic dialog right after calibration&profiling previews relative to the previous calibration).
Is there a way to compare two calibration profiles visually? E.g. the comparison given by the automatic dialog right after the calibration.
Assign the ColorMunki profile to the display (i.e. using the profile loader or Windows color management settings), then choose “<Current>” under “Settings” in DisplayCAL.
Why can’t I do a Measurement Report when choosing another “Settings”, for example sRGB?
Because Microsoft doesn’t care, and their loader always quantizes to 8 bit (which can introduce banding) and scales wrongly.
- What do you mean by Microsoft doesn’t care?
- Is there a way to check if our LUT is 8 bits?
- What about the profile given by DisplayCal? Is it deeper than 8 bits?
- Do you know if the profile given by Colormunki Display’s software has 8 bits values?
Thanks 🙂
Is there a way to compare two calibration profiles visually? E.g. the comparison given by the automatic dialog right after the calibration.
Sure, switch between them in the profile loader to see the calibration differences (if any).
Why can’t I do a Measurement Report when choosing another “Settings”, for example sRGB?
It would be pointless to use one of the presets to verify because they’re just presets, not based on real measurements.
What do you mean by Microsoft doesn’t care?
That bug has been around since Windows Vista (XP didn’t have a native way to load calibration). It’s still around in Windows 10. Ten years.
Is there a way to check if our LUT is 8 bits?
The problem is not the videoLUT.
What about the profile given by DisplayCal? Is it deeper than 8 bits?
Do you know if the profile given by Colormunki Display’s software has 8 bits values?
16 bits. No profiler that I know of (except DataColor’s software, which encodes the video card gamma table tag in 8-bits) uses 8-bit encoding anywhere.
Sure, switch between them in the profile loader to see the calibration differences (if any).
Is there a way to switch quickly between them like at the check of a box as in the “Preview calibration” case?
16 bits. No profiler that I know of (except DataColor’s software, which encodes the video card gamma table tag in 8-bits) uses 8-bit encoding anywhere.
So if I’m loading my Colormunki’s calibration with the DisplayCal Profile loader, I should get better colours?
Is there a way to switch quickly between them like at the check of a box as in the “Preview calibration” case?
Yes. Right-click profile loader icon, choose “Profile associations”. Double-click the profile you want to assign, which will instantly also load the calibration from the profile.
So if I’m loading my Colormunki’s calibration with the DisplayCal Profile loader, I should get better colours?
Not “better colors”, but no quantization artifacts introduced through the bugs that are in Windows native loader.
Not “better colors”, but no quantization artifacts introduced through the bugs that are in Windows native loader.
As long as I don’t see the artifacts under Windows’ native loader, do I have the same color accuracy as if I’m loading with DisplayCal’s loader?
As long as I don’t see the artifacts under Windows’ native loader
On things like a grayscale ramp, they’re probably going to be visible.
do I have the same color accuracy as if I’m loading with DisplayCal’s loader?
About the same, possibly very slightly lower due to the wrong scaling.
How can I re-enable DisplayCal’s profile loader if I had done:
Open Windows color management settings, go to advanced tab, click “Change system defaults”, on advanced tab enable “Use Windows display calibration”.
without trying to install a profile?
Also, there are entries in the Profile associations window which I cannot set because they don’t exist anymore. For the same reason, I cannot remove them as it says “The specified color profile is not associated with the specified device.” How can I remove those entries?I found the solution: I just unchecked and checked the box “Use my settings for this display device.”Also: is there a way to measure how much better the DisplayCal loader is better vs Windows’ native loader? For example by running a Measurement report on the same profile loaded once with DisplayCal and another time with Windows’ native loader? It seems to me that it is always using DisplayCal’s loader when using DisplayCal’s functions.
Also: is there a way to measure how much better the DisplayCal loader is better vs Windows’ native loader?
As the main difference is the lack of additional quantization artifacts when using the DisplayCAL loader, any method that samples data only very sparsely (like the measurement based verification) is not a very suitable means to check this.
It seems to me that it is always using DisplayCal’s loader when using DisplayCal’s functions.
It is not using the loader, but the same underlying (high precision) way to set the videoLUTs.
It is not using the loader, but the same underlying (high precision) way to set the videoLUTs.
Does it mean that everytime I do a measurement report, it’s loading the profile with high precision (the same unerlying way as the loader)?
Also, how do I re-enable the use of the profile loader after I set Windows to be the default manager?
Does it mean that everytime I do a measurement report, it’s loading the profile with high precision (the same unerlying way as the loader)?
Yes (although to be technically correct, what gets loaded into the videoLUT is only the 1D calibration part of the profile).
Also, how do I re-enable the use of the profile loader after I set Windows to be the default manager?
Remove the checkbox from “Use Windows display calibration” under Windows color management settings -> advanced -> change system defaults -> advanced.
