- This topic has 8 replies, 3 voices, and was last updated 5 years, 2 months ago by
.
-
Posts
-
Hello all,
I routinely use DisplayCAL with an i1display to measure how bad the panel uniformity of LCDs are. For instance, now i have panels with a noticeable temperature shift from left to right but I need an objective data on how off it is.
I remember couple years ago DisplayCAL would also include a color for each square that would equal to not only the brightness but the temperature measured in each patch.
What did that feature go?
Is there another way to measure how accurate a panel is in terms of temperature?
Hi,
correlated color temperature (CCT) is a really bad measure of panel uniformity because it only evaluates the red-blue direction. ISO 14861:2015 criteria (using delta E 2000 color difference) offer a much more comprehensive form of evaluation.
Hi!
Thanks.
And how exactly can I make it so it shows me the delta E 2000 for colors?
Here’s what I have now: https://www.dropbox.com/s/b7ujdb07n0xssig/w5k4.html?dl=0
As you can see, it’s only brightness information.
The path I use is Tools -> Report -> Measure Display Device Uniformity
I see, so the closer the Delta is to 0, the closer the colors are to the middle patch, correct?
Also, would there be a way to plot that output into a more visual and tangible way? Like a graph or 3d shape? The green-yellow-red outlines are not enough for me.
I see, so the closer the Delta is to 0, the closer the colors are to the middle patch, correct?
Yes.
Also, would there be a way to plot that output into a more visual and tangible way? Like a graph or 3d shape? The green-yellow-red outlines are not enough for me.
Make sure you are using a modern browser (e.g. Chrome, Firefox). There is a very straightforward traffic light visualization.
I see, so the closer the Delta is to 0, the closer the colors are to the middle patch, correct?
Yes, but you must understand that deltaE00 is a color distance that takes account of brightness diferences between two colors and hue and “strength” of that colors.
If you are worried about color tints across the screen like a pink cast in the right an a green cast on the left it could be numerically more visible if you click on dE values in the report. That will shift uniformity errors to deltaC values which could be explained in a extremely simplified way as the “strength/amplitude” of the color cast you may see in your screen. This is a more relaxed test focused on the strength of color casts.
I see, so the closer the Delta is to 0, the closer the colors are to the middle patch, correct?
Yes, but you must understand that deltaE00 is a color distance that takes account of brightness diferences between two colors and hue and “strength” of that colors.
If you are worried about color tints across the screen like a pink cast in the right an a green cast on the left it could be numerically more visible if you click on dE values in the report. That will shift uniformity errors to deltaC values which could be explained in a extremely simplified way as the “strength/amplitude” of the color cast you may see in your screen. This is a more relaxed test focused on the strength of color casts.
In my case it’s not really tints, but rather temperature only. Warm/orange side vs cool/bluish side. Would deltaC work for that too?
Ideally I’d like to separate luminance from color.
I see, so the closer the Delta is to 0, the closer the colors are to the middle patch, correct?
Yes, but you must understand that deltaE00 is a color distance that takes account of brightness diferences between two colors and hue and “strength” of that colors.
If you are worried about color tints across the screen like a pink cast in the right an a green cast on the left it could be numerically more visible if you click on dE values in the report. That will shift uniformity errors to deltaC values which could be explained in a extremely simplified way as the “strength/amplitude” of the color cast you may see in your screen. This is a more relaxed test focused on the strength of color casts.
In my case it’s not really tints, but rather temperature only. Warm/orange side vs cool/bluish side. Would deltaC work for that too?
Ideally I’d like to separate luminance from color.
It’s separated, just click on dE and it changes to dC.
Temperature IS tint. Think in a 2D plane: pink-green axis and (CCT) blue-yellow axis. Think of dC as the radius from the center (white) towards actual measurement on othe of the screen cells (3×3 or 5×5) without direction info: “how severe/strong that tint is”. That is what makes such tints noticeable, not the direction (although opposite directions add color distance between screen sides)
i1Profiler-like uniformity report won’t take account of that info. +-200K in one side has not a direct translation to a mild/severe color cast. You need to use “visual units” that vary somehow in more or less perceptually uniform way close to the way humans see colors: “d” units
A different thing is that you like a 3D plot like i1Profiler, so you can rotate etc. That is “user experience”… but user experience is useless if data plotted into that graph has no direct relation with what you seek (like with CCT uniformity report).
