ColrNavigation SW limitations

[gioallie]

Hi,

about using ColorNavigator, read here http://www.displaycalibrationtools.com/display-calibration-guides-an

Citing from their page:

1) ColorNavigator does not allow spectro offsets (or reference meter offsets), ergo you are using your colorimeter without an offset
which is amateurish considering we are calibrating Pro grade screens intended for color critical application. If you are trying to just
use your spectro as the active meter for profiling, then your low light readings will be very inaccurate.
Even when using Pro colorimeters like the Klein K10-A that have internal offset slots (that you can pre-load with offsets using
Klein's Chromasurf software) cannot circumvent this missing feature, as ColorNavigator does not allow you to select a slot and only
uses slot 0 from the K10, which is the no-compensation slot.
This missing feature alone makes ColorNavigator not suitable for color critical calibration.

2) ColorNavigator uses the internal patches of the Eizo screen - those are patches that the monitor produces, NOT (!) the graphic
card. This is a great option to calibrate the screen as an all-around monitor (being used with multiple workstations, on-set etc), but it
does not accommodate for the (possible) graphic card distortion of any of the host PC's involved - so if u're using the Eizo stationary
on a single workstation, this is not a great calibration approach.

3) The profiling process in ColorNavigator only samples a few points - the entire profiling session takes 4.30 mins. So the 3D LUT
that ColorNavigator creates is calculated from these few points only. Considering that the Eizo internally stores a 64^3 LUT (with
274,625 points) that means A LOT of these points are interpolated from just a tiny bit of data. Not very accurate.


I'm particularly surprised by #2 which cuts the video card off of the equation of the calibration chain.

What do you think ? Any alternatives available for a calibration SW that can do HW calibration of Eizo monitors ?

Bye,
Giovanni

[howardm]

my hunch is that modern digital video and cards really do not play into it much at all compared to any non-linearities or capabilities of the panel itself.

NEC Spectraview does the same thing.  It creates a linear profile for the video card where output=input.

[D Fosse]

Any alternatives available for a calibration SW that can do HW calibration of Eizo monitors ?

ColorEyes Display Pro supposedly does and should be worth a shot. But the whole DDC/CI thing is rather iffy and judging from the CEDP forum (they even have a separate Eizo sub-forum) it either works or doesn't, with no apparent pattern.

http://www.integrated-color.com/cedpro/coloreyesdisplay.html

For that matter I could give it a go myself. I have a CEDP license, but it's not installed at the moment (I use Colornavigator and am entirely happy with that).

Other than that my feelings echo Howard's. These seem to be very small-scale variations, well within the larger scale of things (individual sensor variations etc). But do report back if you find anything!

[howardm]

this might be of interest........

http://www.luminous-landscape.com/forum/index.php?topic=41580

[D Fosse]

Interesting thread, especially Jack Bingham (head of ColorEyes) stating that they never had problems getting the necessary info from Eizo. So chances are it works. Maybe I should revive my copy of ColorEyes and compare it to Colornavigator just for the fun of it (although, mind you, I'm not complaining over CN).

Anyway, to the OP - the biggest variable of all is finding the right calibration targets. Small theoretical inaccuracies are insignificant in that context.

[brntoki]

I am perhaps the worst person to comment on this, but, just maybe, the best person to comment on this. I've just recently started using the lowest end of the highest end Eizo's for color critical work, the CS-230.

You can say what you want. Color Navigator, with the Eizo supplied puck (Spyder 3 or 4), is lacking in this or that regard. Meh. At the end of the day, my prints are dead on with what my monitor and color calibrated software/workflow show me. "The eyes have it," someone once said. Theories are theoretically as good as practice, but in practice, it's different.

If prints match my screen, what is left to get a wild hair about?

[D Fosse]

Agreed. Calibration isn't a gold standard, some absolute color values carved in stone. To think so is a misunderstanding. If it was, they could just hard-wire 120/0.3/D65 and be done with it. The goal is always screen-to-output match.

It's true that CN measures relatively few patches. If you think too few, just try to run a validation, which measures a whole lot more patches, and compares the measurement with predicted values. I've done this on both the CG246 and the CX240, and in all cases delta E values are in the 0.1 - 0.5 range. Far below anything remotely visible and effectively dead on.

[deejjjaaaa]

The goal is always screen-to-output match.

and absolute majority of the people do not print, ever - does it mean that shall not try to do anything with their monitors 

[digitaldog]

and absolute majority of the people do not print, ever - does it mean that shall not try to do anything with their monitors 

So they calibrate the display for what? What are they attempting to do when they set the target calibration aim-points? I suspect in many cases, they have no idea how to answer that question. Or think there is some “standard” which does not exist, they are hitting.

[brntoki]

and absolute majority of the people do not print, ever - does it mean that shall not try to do anything with their monitors 

But you still miss the point. If the monitor/calibration combination are able to produce output that is accurate to what is seen on screen, it means the screen is accurate. Printing (or other forms of output I suppose) is the proof, but it isn't exactly necessary to produce that proof oneself to have confidence of the accuracy of the system. It seems at the end of the day all that is being said is that, "Hey! The Eizo system has some weak links in its chain and could be better." But that doesn't tell us a single thing about how accurate the system is or isn't as it stands. If it is extremely accurate in its current form, then fretting over how much, hypothetically speaking, more accurate it can be is just whizzing in the wind. Professionals rave about its accuracy.

I don't get the point, unless the only point is that NEC, Eizo, et. al. should not rest on their laurels, but seek to improve. I'm all for improvement, but I think I'll keep working and printing and using the system as it is while waiting for that improvement. It might not be a good idea to wait for hypothetical perfection; a lot of useful work will go undone.

[D Fosse]

and absolute majority of the people do not print, ever - does it mean that shall not try to do anything with their monitors 

Perhaps they should, just once, using a good photo paper. Printed output is the reference.

Or if not, it's still possible to set the white point by visualizing paper white on screen, and then set a reasonable black point like 0.3 or 0.4

[GWGill]

Agreed. Calibration isn't a gold standard, some absolute color values carved in stone. To think so is a misunderstanding. If it was, they could just hard-wire 120/0.3/D65 and be done with it. The goal is always screen-to-output match.

That may be your goal, but don't assume that it is everyone's. It's perfectly sensible to target a specific display reproduction characteristic such as sRGB, AdobeRGB etc., as it provides a standard point of reference for viewing and interchanging electronic images. Given the wide variation in print output capability and print viewing conditions, it might be argued that it is a better reference point than print for many purposes.

There are several paths to achieving the goal of having an image displayed in a specific way, and many of them overlap with the goal of getting good display to print matching, but not all do.

[digitaldog]

It's perfectly sensible to target a specific display reproduction characteristic such as sRGB, AdobeRGB etc., as it provides a standard point of reference for viewing and interchanging electronic images.

It absolutely is but unless all other parties are calibrating the same (identically), all bets are off. And considering when and where sRGB (and lesser so Adobe RGB) are used as an output media, a large majority of users simply will not see this target calibration.

It would be real interesting to hook up say an Eizo and NEC high end unit emulating sRGB and measure how close they get to each other.

[D Fosse]

It's perfectly sensible to target a specific display reproduction characteristic such as sRGB, AdobeRGB etc.

Yes, I see the point, but - given that ambient light/viewing environment affects perception to a large degree - my point is that we need a common ground to stand on. Two displays with identical characteristics, but placed in very different environments, will not be perceived as identical. That's why print matching is still the best common ground.

The important thing isn't similar characteristics, the important thing is similar perception.

I don't print much either. My work is partly for screen viewing, partly for off-site printing.

[GWGill]

Yes, I see the point, but - given that ambient light/viewing environment affects perception to a large degree - my point is that we need a common ground to stand on. Two displays with identical characteristics, but placed in very different environments, will not be perceived as identical. That's why print matching is still the best common ground.

Print is far, far more variable than displays. Different papers, different inks,  more ink combinations, physical tradeoffs between screening artifacts, neutral robustness, sharpness & black density  - all these make it far more difficult to optimize a print or hit a particular target reproduction characteristic. Displays have 3 colorants and are mostly additive, print has more, and the interaction between the colorants is typically complex and not easily represented.
Then add in a major difference between display and prints - displays come with their own illuminant, while a prints appearance will radically change with the illuminant it is viewed in. Then add in all the all the factors you list above - print appearance will depend on the environment, just like displays.

[D Fosse]

So what would you suggest as common ground? I still maintain that D65 is not the answer. It looks different depending on panel/sensor combination.

I need quite different target settings when I calibrate my CG246 with the internal sensor vs. the i1D3. I couldn't tell you which one is more "accurate" - they just need different white point settings to give the same visual result. D65 just isn't the same between them. But once I correct for that, and just try to hit paper white (such as it is), they're indistinguishable.

(Edit - just to avoid misunderstandings, I'm not talking about dramatic differences. But enough to be taken into consideration for critical purposes).