A little color test

[Magnus]

Conclusion: Random monitors are good enough, no calibration and color management is needed?

 

[gordo]

Conclusion: Random monitors are good enough, no calibration and color management is needed?

No, I wouldn't say that. They are not random monitors.

I happen to have a MACbook Pro and an LG TV. The displays for both are made by LG so it's not surprising to me that they would look so similar out if the box. In fact the same is true of my iPhone and iPad. And they all display the image very nearly the same.

I have a feeling that there are are very few, maybe as few as two, manufacturers of displays used by the majority of graphics people who supply art for print and the web as well as the shops that process them. Because of the way displays like LCD screens are engineered and manufactured there's probably not going to be very much difference between them as far as appearance is concerned.

My speculation is that since we are all basically looking at the same screen - we're all seeing and adapting our vision to the same thing. In that context I'm guessing that user calibration and color management applied to their individual displays provides very little benefit relative to the cost and complication involved.

I would guess there is way more variability in the observers than in the displays - and that's something that color management can't fix.

 

[Deleted member 16349]

My speculation is that since we are all basically looking at the same screen - we're all seeing and adapting our vision to the same thing. In that context I'm guessing that user calibration and color management applied to their individual displays provides very little benefit relative to the cost and complication involved.

I would guess there is way more variability in the observers than in the displays - and that's something that color management can't fix.

Again, I like to say how important it is for the output device to be consistent and predictable by design. This goes for monitors or for printing. Then software is then much easier to develop to obtain good results without too much colour management by the user.

I think the theme of this post is quite important. There is colour management done but it being done in the background of the engineered design of more consistent and predictable output devices. It still may not be done as in a general way as possible, which your comment about common display manufacturers imply but it does show it can be done.

 

[Correct Color]

Dang!

So what is a screen capture capturing if it's not capturing what is displayed on the screen?

It's capturing the RGB values that were converted from whatever RGB space the image was in to the display profile. Preview actually converts to the display profile for keeps, not just for display -- which is a reason you never, ever want to save anything in Preview. Windows, I'm not sure.

And changing your monitor settings isn't changing your profile. You'd have to reprofile your display to do that -- and if you did, unless you whacked it out so badly that it couldn't -- it would try to reinterpret that image to match the RGB values in each pixel in the new profile to the L*a*b* values of each pixel as represented by the RGB values in the image. In other words, it would still look the same. It being still the same device with the same basic capabilities.

As for this whole thing, what I'd point out is that what you have here is an image that was sent to the Internet tagged as sRGB.

And while color management on the Internet is still something of a mixed bag, what has happened in the past few years is almost all browser manufacturers recognize the sRGB tag if it's there, then convert to whatever the monitor profile is. If it's not there, then most -- Safari the last time I checked being a notable exception -- assign sRGB, and then convert to the monitor profile.

What Safari does with untagged images is just assign the monitor profile directly -- which is inordinately stupid, but hey, they're Apple, and you can't tell them anything.

Image five in the composite looks like one that had this treatment.

Also at one point, some browsers would recognize other embedded profiles besides sRGB, and convert directly from the tagged space to the monitor profile. However, that seems to have gone out of fashion, and most that recognize profiles now seem to convert first down to sRGB, and then to the monitor profile.

Of course, any one of them can change how they do it with any upgrade they send along. And it's damn tedious to figure out which one's currently doing exactly what. And I haven't felt like wading into that swamp lately, so some of my info may be out of date.

But the bottom line is that if you want a true test, all this stuff does go on in some form or fashion, and it has to be accounted for.

Also of note in this test is that almost all of the monitors are Macs of some sort. And all Mac monitors unprofiled do display somewhat similarly: Way too bright, and a good deal too blue.

Apple is not in the graphics business anymore. Pretty much now, they make their screens for watching movies.

And also note that "calibrated" does not mean profiled. "Calibration" in and of itself actually does just about nothing.


I can say that one of the things I do for a living is profile monitors. As part of the profiling process, I always do a before/after comparison. If it was true that monitors had become so similar that profiling them was unnecessary, then what I'd see on a pretty regular basis is that before and after weren't all the different.

And as of yet, that's very, very far from the case.


(Edited to add:

I have a feeling that there are are very few, maybe as few as two, manufacturers of displays used by the majority of graphics people who supply art for print and the web as well as the shops that process them. Because of the way displays like LCD screens are engineered and manufactured there's probably not going to be very much difference between them as far as appearance is concerned.

No, no... Just not true.

I have no idea how many people manufacture the hardware, but that would be something akin to saying that all Heidelbergs print the same because they were made by Heidelberg. The hardware may be identical, but what drives it is every bit as/more important.

And displays are just like any other devices that reproduce color with primaries. If you want them to reproduce color predictably, you have to calibrate them to a state, and then characterize (profile) them in that state.



Mike Adams
Correct Color