Luminous Landscape Forum
Raw & Post Processing, Printing => Colour Management => Topic started by: Lupin on October 23, 2013, 11:40:25 am
-
I'm using an Eizo CG241W calibrated with a ColorMunki Photo via Eizo's Color Navigator 6 software.
I 'm editing Nikon files in Photoshop CS6 & Capture NX2 and want the monitor to match my prints as closely as possible.
I've been using the ColorMunki to calibrate the Eizo to 5500K and wonder if this is the best temperature to choose.
Doses anyone have any advice?
-
I've been using the ColorMunki to calibrate the Eizo to 5500K and wonder if this is the best temperature to choose.
The best temperature value (which can vary) is that which produces a visual match.
-
The best temperature value (which can vary) is that which produces a visual match.
I haven't had them printed yet (I'll be using a lab) so I can't tell which produces the best visual match.
To my eye 6500K looks very blue on the Eizo and calibrating to that could produce prints that are far too warm. 5500K looks 'about right' but then so does 6000K.
-
do some test prints to satisfy yourself - and share it
Henrik
I use 6500k
-
I use 6500k
Do the whites on your monitor look blue(ish) at 6500K?
-
To my eye 6500K looks very blue on the Eizo
Me too. Well, not very blue, but blueish. I have settled for 6200, which seems to me to match "standard" paper color very well - in my work environment. As already said, you need to try for yourself.
-
Me too. Well, not very blue, but blueish. I have settled for 6200, which seems to me to match "standard" paper color very well - in my work environment. As already said, you need to try for yourself.
Yes, I suppose I'd better get some test prints made.
6200 still looks slightly blue to me. White looks accurate at somewhere between 5500 and 6000 on this CG241 - so 5750 is perhaps the best option. I'm keeping brightness at 100cd or below (via ColorNavigator) and the Eizo's auto brightness regulator switched off. Ambient light in the room is kept pretty constant by keeping the window blinds closed.
-
Doses anyone have any advice?
Download a "standard" test image, have it printed at whichever print shop you will be using. That will provide a reference to work with. Of course for those who do their own printing the test image provides a standard to adjust their printer against too.
With a "standard" print you can view it in the appropriate light (i.e., wherever it is your eventual product will be viewed, whether that is idea or not) and see how you will want it to appear when displayed on your monitor.
Here is the test image I've used for many years: http://digitaldog.net/files/Printer%20Test%20file.jpg
Here is another very good image, Note that it can be downloaded as a 40Mb TIFF file, down at the bottom of this page: http://www.outbackphoto.com/printinginsights/pi048/essay.html
And best of all about that image, here is a detailed essay on how to use it: http://www.outbackphoto.com/printinginsights/pi049/essay.html
Whether you use a commacial print lab and just want a reference for what they produce, or for use with your own printer, eventually the test image provides a standard against which the monitor display is configured during calibration. Brightness, gamma (contrast), and color temperature parameters are set to provide a match between the monitor's display and the paper print.
And, incidentally there are people who find anything from 80 to 160 cd/m correct for brightness (it depends mostly on the ambient light at your monitor's location), anything from 4500K to at least 6500K for color temperature, and anything from less that 1.8 to more than 2.5 for gamma. I suspect that those last two vary mostly with the color under which the print is viewed and the intensity of the light for gamma.
Another point to note is that there are people here on Lula who get pretty hot about this topic and tend to be a bit nasty about it too. They probably should be ignored on this any any other topic too...
Here is another well written authoritative discussion of this topic, with good perspective. It is written by Jim Perkins, a professor at Rochester Institute of Technology.
http://blogs.scientificamerican.com/symbiartic/2012/01/17/how-to-calibrate-your-monitor/
-
Here is another well written authoritative discussion of this topic, with good perspective. It is written by Jim Perkins, a professor at Rochester Institute of Technology.
http://blogs.scientificamerican.com/symbiartic/2012/01/17/how-to-calibrate-your-monitor/
You call it "well written"? With "good perspective"? ::) :D :-X
Quit joking and read something serious:
http://oicherman.com/Boris/Science/Boris_Oicherman_Phd_thesis.pdf
-
Many thanks for all the info Floyd.
I'll download the test images you suggested and will get the lab to print them (together with one of my own for comparison) at say 5500, 5750 and 6000K. The question now is which gamma setting to use.
I've read Jim's article - he suggests a gamma of 1.8 for photographic print work - what do you think of using this instead of 2.2? All of the calibrations options I've produced with the ColorMunki Photo are at 2.2 - maybe I should make another set of options at 1.8 and get the lab to make six prints of each test image, i.e. 5500/5750/6000 at 2.2 and 5500/5750/6000 at 1.8.
EDIT - Actually, I can't figure out whether he's recommending 1.8 for print work (as in photography) or for print work (as in books) ???
-
Damn Czornyj,
is it summarized somewhere? ;D 281 pages! Well, not tonight honey!
thanks for sharing
Henrik
-
Chapter 5., p. 267 - Conclusions :)
-
You call it "well written"? With "good perspective"? ::) :D :-X
Quit joking and read something serious:
http://oicherman.com/Boris/Science/Boris_Oicherman_Phd_thesis.pdf
That is pretty good material, except it is a PhD thesis and does not provide even the slightest persective for the OP. In fact though, he can get exactly that! Yep, the cited article by Perkins gives a pretty good perspective on what the PhD thesis says. There is no disagreement between them.
For example, it seems that a number of people do like a more bluish looking monitor screen than measurements would suggest as correct. The thesis says, "On average, observers used more blue light to match the paint sample on computer displays than the CIE Standard Colorimetric Observer predicts." Their experiments verified the discrepancy.
-
Many thanks for all the info Floyd.
I'll download the test images you suggested and will get the lab to print them (together with one of my own for comparison) at say 5500, 5750 and 6000K. The question now is which gamma setting to use.
Keep in mind that the data in the image file is "correct". You do not want to edit or change it in anyway. Do not open it with PhotoShop and add a profile for printing before sending it off the the printer! What you want to see is what the printer (yours or the commercial lab) does with "correct" data.
I've read Jim's article - he suggests a gamma of 1.8 for photographic print work - what do you think of using this instead of 2.2? All of the calibrations options I've produced with the ColorMunki Photo are at 2.2 - maybe I should make another set of options at 1.8 and get the lab to make six prints of each test image, i.e. 5500/5750/6000 at 2.2 and 5500/5750/6000 at 1.8.
EDIT - Actually, I can't figure out whether he's recommending 1.8 for print work (as in photography) or for print work (as in books) ???
He is talking about making prints. Or actually about the monitor calibaration for editing images that will be printed. Again, the test print is an example of "correct" data. It is the "standard" to compare to.
The essay on how to use a test print, at http://www.outbackphoto.com/printinginsights/pi049/essay.html can help a great deal with how to decide brightness, color temperature, and gamma configuration parameters, both for a printer and a monitor. If you have a print driver that allows adjustments you might want to fine tune brightness, contrast, density, and saturation using a test print to determine what is "correct". Then it should be profiled with a ColorMonki to get correct colors across the entire tonal range. If you use a commercial printer, just send them that file and get a print. That is what you get from them when the image data is "correct".
The cited essay goes into more detail, but essentially you'll choose a gamma value by looking at the gray ramp. In that particular test print they have a very good ramp, with black values from 0 to 25 and with white values from 243 to 255. Ramps that show full stops are not fine enough granularity to give precise results. (Basically, everything with a value of less that 20 is going to be black, and all above 246 or so is going to be white. Not that there is not variation, but there is not enough to see any fine detail in an image. Your printer and monitor should each be configured for a gamma that provides that result.)
With either a commercial print lab or a local printer the computer monitor that will be used to edit other images needs to be first configured for brightness, gamma, and color temperature so that it closely matches the test print produced by the target printer. Then it too is profiled to provide correct colors across the tonal range.
At that point you have an "accurate" system. It is never going to be perfect, but that will be as close as is possible. What you see on the monitor will be essentially what it will look like when printed.
-
Keep in mind that the data in the image file is "correct". You do not want to edit or change it in anyway. Do not open it with PhotoShop and add a profile for printing before sending it off the the printer! What you want to see is what the printer (yours or the commercial lab) does with "correct" data.
Ok, I won't alter the downloaded image files at all. According to two Exif viewers (KUSO and Photo-Me), the 406KB file from DigitalDog is ColorMatch RGB and the 40MB file from Outback is ProPhoto RGB. My Photoshop CS6 is set up for ProPhoto RGB so it wouldn't alter the Outback image anyway. Is there any advantage to test-printing both of these or will the Outback file be sufficient?
He is talking about making prints. Or actually about the monitor calibaration for editing images that will be printed. Again, the test print is an example of "correct" data. It is the "standard" to compare to.
Ok.
The essay on how to use a test print, at http://www.outbackphoto.com/printinginsights/pi049/essay.html can help a great deal with how to decide brightness, color temperature, and gamma configuration parameters, both for a printer and a monitor. If you have a print driver that allows adjustments you might want to fine tune brightness, contrast, density, and saturation using a test print to determine what is "correct". Then it should be profiled with a ColorMonki to get correct colors across the entire tonal range. If you use a commercial printer, just send them that file and get a print. That is what you get from them when the image data is "correct".
The cited essay goes into more detail, but essentially you'll choose a gamma value by looking at the gray ramp. In that particular test print they have a very good ramp, with black values from 0 to 25 and with white values from 243 to 255. Ramps that show full stops are not fine enough granularity to give precise results. (Basically, everything with a value of less that 20 is going to be black, and all above 246 or so is going to be white. Not that there is not variation, but there is not enough to see any fine detail in an image. Your printer and monitor should each be configured for a gamma that provides that result.)
With either a commercial print lab or a local printer the computer monitor that will be used to edit other images needs to be first configured for brightness, gamma, and color temperature so that it closely matches the test print produced by the target printer. Then it too is profiled to provide correct colors across the tonal range.
At that point you have an "accurate" system. It is never going to be perfect, but that will be as close as is possible. What you see on the monitor will be essentially what it will look like when printed.
I'm in the middle of reading the 'printinginsights #049' essay. One question springs to mind - I don't have my own printer so I'll be using a commercial lab. My own image files already have the printer profile supplied by the lab, so no problem there - but what about the Outback 40MB test image? Should the lab use their own profile on this too - or should I tell them to use ProPhoto RGB?
-
Is there any advantage to test-printing both of these or will the Outback file be sufficient?
Oh, I expect that just the Outback file would be just fine.
Should the lab use their own profile on this too - or should I tell them to use ProPhoto RGB?
Good question. I'm not positive, as I don't use a commercial print lab.
I'd try both, but I expect that using their profile is the right way to do it.
-
On the contrary - it makes it clear that the problem is extremely complicated, to some degree all observer models fail, and even in ideal condition there's no chance to find the magic numbers for calibration target that will produce perfect visual match.
Furthermore, there are other variables like spectra of the light, remission curve of paper, amount of OBA, precision of the sensor, viewing conditions and so on...
That is pretty good material, except it is a PhD thesis and does not provide even the slightest persective for the OP. In fact though, he can get exactly that! Yep, the cited article by Perkins gives a pretty good perspective on what the PhD thesis says. There is no disagreement between them.
-
On the contrary - it makes it clear that the problem is extremely complicated, to some degree all observer models fail, and even in ideal condition there's no chance to find the magic numbers for calibration target that will produce perfect visual match.
Furthermore, there are other variables like spectra of the light, remission curve of paper, amount of OBA, precision of the sensor, viewing conditions and so on...
Wonderful, but between you and the thesis, there is nothing there for the OP.
-
Ok, I won't alter the downloaded image files at all....
ColorMatch RGB......
ProPhoto RGB........
I'll be using a commercial lab.......
Have you asked the lab what colourspaces they can handle ?
Do they expect images in a specific colourspace ?
My own image files already have the printer profile supplied by the lab, so no problem there.....
What do you mean by that ?
Have you just soft proofed to the printer profile ? or assigned a printer profile ? or converted to a printer profile ? the last two almost certainly would be a wrong approach.
Details like this are where people get things wrong.
-
the computer monitor that will be used to edit other images needs to be first configured for brightness, gamma, and color temperature
This thread has deteriorated somewhat into noise, but just to keep one thing clear: In a color managed environment, monitor gamma is "invisible". Source profile gamma is remapped into monitor gamma. So all this discussion about gamma is moot. What you aim for by setting monitor gamma is simply to make the monitor perform well, and that's usually accomplished by staying close to native.
Outside color management is a different matter. And that's where the Scientific American article repeatedly referred to belongs. Authoritative, maybe, but without a trace of modern color management.
-
This thread has deteriorated somewhat into noise, but just to keep one thing clear: In a color managed environment, monitor gamma is "invisible". Source profile gamma is remapped into monitor gamma. So all this discussion about gamma is moot. What you aim for by setting monitor gamma is simply to make the monitor perform well, and that's usually accomplished by staying close to native.
The monitor's hardware adjustment is simply to make the monitor perform well, and that is not what has been discussed. The discussion is about the target for the calibration process, and hence what the monitor shows when there is color management.
Outside color management is a different matter. And that's where the Scientific American article repeatedly referred to belongs. Authoritative, maybe, but without a trace of modern color management.
The article specifically discusses calibration of a monitor for a color managed system.
-
The article specifically discusses calibration of a monitor for a color managed system.
...in 90's.
Take two displays, calibrate one of them to gamma 2,2 and second to gamma 1,8, create display profiles. Then open same test image on both monitors in a colour managed application - they'll look exactly the same.
-
I'm using an Eizo CG241W calibrated with a ColorMunki Photo via Eizo's Color Navigator 6 software.
I 'm editing Nikon files in Photoshop CS6 & Capture NX2 and want the monitor to match my prints as closely as possible.
I've been using the ColorMunki to calibrate the Eizo to 5500K and wonder if this is the best temperature to choose.
Doses anyone have any advice?
Send a file to your commercial printshop. Then take thar print and put it under a daylight lamp. Then compare with image on your screen. Do the same with a white paper from yohr printshop. Match the luminace amd white point of your monitor to get a match or as close as possible. Also read andrew's article why are my prints too dark. That is all.
-
The article specifically discusses calibration of a monitor for a color managed system.
No, it doesn't. You, and the good professor, both seem to miss the distinction between calibration (a basic device modification), and monitor profile (a description of the device in its current state).
A color managed application like Lightroom or Photoshop doesn't care whether the monitor is calibrated or not. It's irrelevant for the purpose. All it cares about is the profile, the description. A standard color management chain converts from a source profile to a destination profile, and that's exactly what happens here. The document profile is converted directly to the monitor profile and that is what goes to the display. IOW, the calibration is not part of the color management chain.
From a color management perspective, a reasonably good monitor doesn't need to be calibrated beyond setting the white point (which then becomes simply "white"). It just needs to be profiled.
(You might want to set the black point as well, but leave that out for now).
I'm sure professor Jim Perkins is a great scholar, but he's very out of date regarding color management. Everything he's saying in that article relates to using non-color managed software. That's the extent of its "authority".
-
No, it doesn't. You, and the good professor, both seem to miss the distinction between calibration (a basic device modification), and monitor profile (a description of the device in its current state).
The "good professor" in fact discussed calibration using various hardware devices, listing several from Pantone, Datacolor and X-Rite in Part One. The purpose of using these devices, which is what the article discusses, is to generate a profile of the monitor.
It appears that you may be assuming the article is about the hardware configuration of the monitor, which it is not.
A color managed application like Lightroom or Photoshop doesn't care whether the monitor is calibrated or not. It's irrelevant for the purpose. All it cares about is the profile, the description.
The purpose of calibration is to produce the profile. Absent calibration there is no profile and therefore no real color managed system at all.
A standard color management chain converts from a source profile to a destination profile, and that's exactly what happens here. The document profile is converted directly to the monitor profile and that is what goes to the display. IOW, the calibration is not part of the color management chain.
A "document profile" cannot be converted directly to a "monitor profile" absent a profile of the display device having been made with a calibrator. IOW, the display device characteristics that are profiled play a very important part in a color managed system.
From a color management perspective, a reasonably good monitor doesn't need to be calibrated beyond setting the white point (which then becomes simply "white"). It just needs to be profiled.
Calibration is the method by which a profile is produced.
(You might want to set the black point as well, but leave that out for now).
I'm sure professor Jim Perkins is a great scholar, but he's very out of date regarding color management. Everything he's saying in that article relates to using non-color managed software. That's the extent of its "authority".
He knows that calibrating a monitor produces a profile of the monitor.
"The sensor measures a series of colors on your screen and creates a “profile” that brings your display to a reference state." Datacolor's description of calibrating a monitor at <http://spyder.datacolor.com/portfolio-view/spyder4elite/>.
"monitor profiles are prepared when conducting calibration" Eizo's description of calibrating a monitor at <http://www.eizo.com/global/library/management/calibration/>.
-
Have you asked the lab what colourspaces they can handle ?
Do they expect images in a specific colourspace ?What do you mean by that ?
Have you just soft proofed to the printer profile ? or assigned a printer profile ? or converted to a printer profile ? the last two almost certainly would be a wrong approach.
The lab advises sending files either in sRGB or with one of their profiles applied (the profile used depends on the paper chosen by the client).
For editing - I start with a NEF, make the initial adjustments in NX2 and save as a ProPhoto RGB TIFF. Then I do the rest of the edit in CS6 and save as a PSD file.
For printing - to prepare the file for the lab I'll open the PSD in CS6, apply the lab's paper profile using the 'convert to profile' option, then save as a differently-named TIFF. This is the file that will go to the printer.
-
The purpose of calibration is to produce the profile.
I give up. I don't have time for this. This is totally confused, and not mere semantics.
Anyone else?
-
The lab advises sending files either in sRGB or with one of their profiles applied...............
Sounds like you're across the issue pretty well.
Andrew's advice (Digitaldog) reply #1 is the best to follow. Get a known image printed by them, then adjust CT to provide the best match in the expected viewing conditions.
-
Andrew's advice (Digitaldog) reply #1 is the best to follow. Get a known image printed by them, then adjust CT to provide the best match in the expected viewing conditions.
Yes, I'll send them the 40MB test image (Outback's) and get them to print on the chosen paper.
The Eizo is currently on 5500K and looks a tad warm - so I suspect the print will be slightly blue in comparison. I reckon changing to 5750 or 6000K (or somewhere between the two) will give the most accurate match to the test print.
-
I give up. I don't have time for this. This is totally confused, and not mere semantics.
Anyone else?
I suspect that it actually is sematics. All of these words for parts of the process are used in sometimes narrow ways, and at other times are overloaded wtih broader meanings.
Generating a profile of the monitor's characteristics is part of "calibration". But calibration also includes configuration of the monitor itself and it also includes software configuration of the video card (look up tables, for example) and possibly software configuration of either the OS or other system software (video drivers, graphic display systems, etc).
The point is, and has all along been, that Jim Perkins absolutely was speaking of "calibration" using hardware calibrators and he absolutely was discussing the entire range of configurations including generation of a monitor profile.
And his discussion was clearly intended for a system using color managed applications.
-
Let me just say one thing, and then I'll leave it:
Without a clear linguistic distinction between calibration and profiling, the underlying mechanics and processes are obscured and any meaningful discussion impossible. If it's all indiscriminately lumped under the heading "calibration", entropy goes up and understanding down.
Precise language is essential. Calibration is one thing. A monitor profile is something else. Yes, the calibration LUT is often embedded in the profile, simply because it's a convenient place to store it, and I suspect this is the basis for much of this confusion. But they're still separate with separate functions.
Again: calibration is not part of the color management chain. Color management relies solely on the profile, which is made after the calibration, not during it, and is entirely independent from it.
-
Again: calibration is not part of the color management chain. Color management relies solely on the profile, which is made after the calibration, not during it, and is entirely independent from it.
A profile is generated using the hardware calibration equipment and the measurements made with it that are used for other parts of the calibration process. The profile is not independent from the rest of the process, it is merely the end of the entire process.
-
I'm using an Eizo CG243W in 10 bit mode connected to display port with the following calibration settings (DTP 94 puck):
| Luminosity: | 100 cd/m² |
| Color Temperature: | 5500 K |
| Gamma: | L* |
Luminosity above 100 gave me dark prints, 6500 K gave too warm prints, Gamma values messed my shadows toning.
With the values above I'm quite happy, but of course thats all workplace and exhibition place dependent.
E.G: my PC workplace doesn't have a white painting but some apricot-orange like walls - pros might sigh at this.
But much of the walls is covered by bookshelves with white and mixed colors from the books.
I usually have the room light quite dim here, so 100 cd/m² is bright for this situation and the walls don't mess up my color vision too much.
After printing I turn up the light (halogenide indirect light against the white ceiling) to judge color.
Exhibition places vary, but are mostly a mixed daylight/halogenide Tungsten situation.
Printer is an Epson 7890, printing on Moab Summerset and Ilford Gold Fibre Silk papers
Hope that helped a little
Cheers
~Chris.
Addendum:
If you like to go really deep into that color stuff (beyond the questions of mere calibrating or profiling) I recommend this book by
Mark D. Fairchild: "Color Appearance Models"
website: http://www.cis.rit.edu/fairchild/CAM.html
-
Luminosity above 100 gave me dark prints, 6500 K gave too warm prints, Gamma values messed my shadows toning.
To be pedantic. Those values caused you edit your images wrongly, they didn't change the image data at all.
-
To be pedantic. Those values caused you edit your images wrongly, they didn't change the image data at all.
Agreed - its pedantic ... ;)
-
I'm using an Eizo CG243W in 10 bit mode connected to display port with the following calibration settings (DTP 94 puck):
| Luminosity: | 100 cd/m² |
| Color Temperature: | 5500 K |
| Gamma: | L* |
Luminosity above 100 gave me dark prints, 6500 K gave too warm prints, Gamma values messed my shadows toning.
With the values above I'm quite happy, but of course thats all workplace and exhibition place dependent.
I'm not sure what L* is - I've seen it at the right end of ColorNavigator's gamma adjustment slider but don't I know why it's called L* instead of a number (such as 2.7). What does it mean?
-
I'm not sure what L* is - I've seen it at the right end of ColorNavigator's gamma adjustment slider but don't I know why it's called L* instead of a number (such as 2.7). What does it mean?
It's tonal response curve of L*a*b colour space, it doesn't match any gamma curve
-
It's tonal response curve of L*a*b colour space, it doesn't match any gamma curve
What is the advantage of using it?
-
What is the advantage of using it?
Gamma basically is a relict from the times of cathod ray tube monitors.
My impression is that I get better results from L* with modern LCD screens.
I won't go into the depths of L*a*b here, just so far, that its a color space designed to roughly linearize the numbers representing colors with the subjective perception of colors in a way that you'll feel something being double as bright when the L value is doubled, and same for color in an analogue way.
There have been perception tests made in the past with individuals judging color to achieve that.
But L*a*b is not perfect and you can get nasty color casts when using it in the wrong way - but this is only when using it in Photoshop or elsewhere as a working colorspace - that problem has nothing to do with that monitor calibration thing.
But thats another story.
I'd suggest just try it and see how the tones of your screen and your prints match when using it, especially the shadows.
Cheers
~Chris
-
.......
I'd suggest just try it and see how the tones of your screen and your prints match when using it, especially the shadows.
Cheers
~Chris
Ok, I'll create a profile using L* with 5500K and see what it looks like. My CG241 won't do 10 bit though (unlike your CG243) - I don't think my HD6700 graphics card will either :'(
-
Ok, I'll create a profile using L* with 5500K and see what it looks like. My CG241 won't do 10 bit though (unlike your CG243) - I don't think my HD6700 graphics card will either :'(
It works as well with 8 bit color - plays totally no role with your problem ...
Experimenting with the settings will in the end give you good results.
Good luck!
~Chris
-
What is the advantage of using it?
Better feeling ;)
-
What is the advantage of using it?
Depending on who you ask, either nothing or it's the greatest target for calibration of this aim point that exists. Still waiting on some useful peer review evidence that using Lstar for display calibration is useful.
In terms of language, Calibration is placing a device into a known, repeatable and if possible ideal condition. A profile just reflects that condition. Yes, you can profile without calibrating (a modern Epson printer with the Epson driver is such an example). You don't have to calibrate a display but if you want a decent print to screen match, you'll probably have to do so. The profile describes this for the ICC aware app's.
-
.... In terms of language, Calibration is placing a device into a known, repeatable and if possible ideal condition. A profile just reflects that condition. Yes, you can profile without calibrating (a modern Epson printer with the Epson driver is such an example). You don't have to calibrate a display but if you want a decent print to screen match, you'll probably have to do so. The profile describes this for the ICC aware app's.
Yes, I should have said: I'll calibrate to L* with 5500K and see what it looks like.
-
It works as well with 8 bit color - plays totally no role with your problem ...
Damn, I was hoping you'd say it was vital - then I'd have an excuse to go out and buy a CG276! :D
-
Depending on who you ask, either nothing or it's the greatest target for calibration of this aim point that exists. Still waiting on some useful peer review evidence that using Lstar for display calibration is useful.
In terms of language, Calibration is placing a device into a known, repeatable and if possible ideal condition. A profile just reflects that condition. Yes, you can profile without calibrating (a modern Epson printer with the Epson driver is such an example). You don't have to calibrate a display but if you want a decent print to screen match, you'll probably have to do so. The profile describes this for the ICC aware app's.
For me it was simply following an advice from my local camera dealer and it worked greatly.
I'm totally not interested too much in the academic reasoning behind that, since I know the whole color calibration thing is b0rk3d - because our vision is b0rk3d.
I mean our eyes and film and digital sensors - they translate wavelength distributions into a set of 3 or 4 values (RGB values and neural response respectively).
(4 in the case of our eyes - since we have an additional luminosity channel based on mostly green wavelengths).
So - all this reasoning is sort of messed from ground up - despite all the brilliant science behind it.
After all you simply need to find a way to work which allows you to get the look you want to have for your images.
And that is highly subjective, artistic, esoteric, (fill in whatever you like).
Apart from that, for all people seriously interested I strongly recommend that book of Mark D. Fairchild mentioned above.
Cheers
~Chris
-
A profile is generated using the hardware calibration equipment and the measurements made with it that are used for other parts of the calibration process. The profile is not independent from the rest of the process, it is merely the end of the entire process.
Floyd, let me ask you this question. Suppose you place two identical monitors side-by-side. Then you "calibrate" the first monitor to gamma=1.8 and then "profile" that monitor. If you do not like the semantic separation of calibrate/profile, consider it just one combined process. Now you repeat that combined process on the second monitor, but with gamma=2.2. Finally, you open a reference image with an embedded profile in a color managed application on both monitors. Will the two images look different or the same? (The correct answer is that they will look the same.)
-
From a color management perspective, a reasonably good monitor doesn't need to be calibrated beyond setting the white point (which then becomes simply "white"). It just needs to be profiled.
How does the white point setting interact with the profile and the embedded color space of an image? After I set or calibrate the monitor to a given color temperature and luminosity, is that the "white point" from now on? Does profiling change the display of white (255,255,255) based on the color space embedded in the image? How about neutral gray (128,128,128)? I realize the color space gamma will change the displayed brightness, but will the color temperature also have an effect, or is the original calibration color temperature preserved?
-
Floyd, let me ask you this question. Suppose you place two identical monitors side-by-side. Then you "calibrate" the first monitor to gamma=1.8 and then "profile" that monitor. If you do not like the semantic separation of calibrate/profile, consider it just one combined process. Now you repeat that combined process on the second monitor, but with gamma=2.2. Finally, you open a reference image with an embedded profile in a color managed application on both monitors. Will the two images look different or the same? (The correct answer is that they will look the same.)
Yep, but now let me ask you this... what if you are downloading data from the web using a browser that does not have color management? What if you are embedding images into any of a variety of other systems, such as a webpage, a computer program, a text editor, or an PDF file, and what about using even color managed applications to view images that do not have an embedded profile?
I personally have always objected to the way most computer systems use color management these days because the above situations are not handled well. I prefer to let the OS deal with color management rather than each individual application. I don't mean to say that it is a perfect solution, but for my purposes it is vastly better. In that direction my own workstation has a customized configuration menu for monitor calibration, and color management to overrides it is generally disabled. With two clicks of a mouse I can have a large menu of different monitor calibration profiles to choose from.
-
With two clicks of a mouse I can have a large menu of different monitor calibration profiles to choose from.
You really don't understand colour management at all.
-
+1
For non colormanaged situations i do not care at all.
For producing my images i do care and the info provided by bruce, andrew and jeff proved to work for me the past 5 years. And when i publish to the non colormanaged world, i just check on regular non colormanaged computer displays or hdtv's. In caze of doubt on a ipad display.
-
You really don't understand colour management at all.
+2. That's the impression you get. But I've been in discussions like this before and they usually go nowhere (except spiralling round in circles).
I don't know why some people insist on overcomplicating this, constructing all these personal theories. Yes, finding a workable set of calibration targets can take some effort, but the underlying processes are simple.
One example:
what about using even color managed applications to view images that do not have an embedded profile?
What's the big issue here? Two things can happen: 1) the Photoshop model, assigning the working space. Problem solved, the color management chain humming nicely along and ending up in the monitor profile. Firefox mode 1 uses this. 2) the Safari/default Firefox model, leaving it unmanaged, sending the source data straight to the display.
You just need to figure out which one applies. You don't need a conspiracy theory to deal with it.
And so on.
-
I don't know why some people insist on overcomplicating this, constructing all these personal theories.
Because they don't understand the fundamentals.
-
Because they don't understand the fundamentals.
… they don't WANT to understand the fundamentals, enjoy argument for argument's sake - commonly referred to as trolling.
-
..... Will the two images look different or the same? (The correct answer is that they will look the same.)
How can they look the same? One monitor is set to 1.8 and the other to 2.2 - if everything else remains equal they should look something like this: http://imageshack.us/a/img407/6302/gammadifference.jpg.
-
How can they look the same? One monitor is set to 1.8 and the other to 2.2 - if everything else remains equal they should look like this: http://imageshack.us/a/img407/6302/gammadifference.jpg.
The key to the question is that it says view an image with "an embedded profile in a color managed application" which means the monitor profile is used to determine what standard the monitor is set for so the viewing application can change the image data to match the requirements of the embedded profile. Since in either case it is the embedded profile which determines the final appearance, the usefulness of the two different monitor profiles is as an accurate starting point when adjusting the image data. You don't end up viewing image data from the original file, and instead see adjusted data edited to match the embedded profile to the selected monitor.
Of course if the embedded profile does not produce an image that matches your printer output (for example because it was constructed for use with a different paper), then the displayed image will not be what you get when the image is printed.
That is one reason that I do not prefer to have color managed by each application, and prefer to ignore embedded profiles by default. Those do work for some people's workflow, but is awkward for the work that I typically do. I usually want to see everything in one partricular manner regardless of the viewing program or the embedded profile. For example when working on images for web display everything looks approximately correct for an sRGB monitor set for 6500K or higher (many default to above 9000K), with gamma 2.2 and with slightly excessive brightness for the environment. That means it won't be far off for most web viewers. But when working on print jobs, one of several other monitor profiles are more appropriate.
I don't work to the lowest common denominator of color management, and instead fine tune color management to the specific workflow being used, which is optimized for more flexibility in areas of more significance for time or other efficiency.
-
I still don't get it - ??? - maybe I'm missing the point behind Eric's question.
He asked if an image with an embedded profile would look the same on two identical monitors (one at 1.8, one at 2.2) when opened in a colour managed application.
I only have one monitor so let's suppose I simulate having two by using this (very extreme) scenario:
Create ColorNavigator Target A - 100cd / 5500K / Gamma 1.0
Create ColorNavigator Target B - 100cd / 5500K / Gamma 2.6
Open the 40MB Outback ProPhoto RGB test image in Photoshop CS6 (with working space: ProPhoto RGB)
If I do nothing else but switch the CG241 from Target A to Target B (simulating monitor A and monitor B), the appearance of the test image changes dramatically - it certainly does not look the same. If Target A is 1.8 and Target B is 2.2, the change would be far less dramatic but the test image would still look different.
-
How can they look the same?
Welcome to color management. This is a standard profile conversion, and source gamma encoding (document profile) is remapped into destination gamma encoding (monitor profile). This, remember, is a precise description of the monitor's actual behavior, whatever gamma it is calibrated to. The net result is linear. It doesn't matter what the respective gamma values are on either side of this conversion.
Exactly the same thing happens when you convert from Adobe RGB (gamma 2.2) to ProPhoto (gamma 1.8 ).
Without color management you don't have this translation and a 2.2 encoded file must be viewed on a 2.2 display (or any other equal relationship).
-
When you switch from calibration target A to target B, you also need to relaunch Photoshop (or whatever CM'ed app you use for viewing), so that it can pick up the corresponding profile and make the proper conversion for the display.
-
When you switch from calibration target A to target B, you also need to relaunch Photoshop (or whatever CM'ed app you use for viewing), so that it can pick up the corresponding profile and make the proper conversion for the display.
Ah, ok - that answers it.
-
How can they look the same? One monitor is set to 1.8 and the other to 2.2 - if everything else remains equal they should look something like this: http://imageshack.us/a/img407/6302/gammadifference.jpg.
They look the same due to the Display Using Monitor Compensation architecture in ICC aware applications which looks at the display profile AND the working space of the document and builds a unique on the fly preview. The gamma settings within the display and working space profiles have no direct role here in color managed applications outside DUMC in terms of the preview. IOW, ColorMatch RGB with a 1.8 TRC gamma when converted to sRGB with a 2.2 TRC gamma appear the same. In color managed app's. Outside color managed app's they do not. Or sRGB on a display calibrated to 1.8 looks the same on a display calibrated to 2.2 within color managed app's. Outside such app's they do not. Non color managed app's don't know what a display or working space profile are. They have absolutely no idea what any of this means. The condition of the display gamma now has an effect on the preview.
-
Yep, I've got it now. These three lines are the key:
1. sRGB on a display calibrated to 1.8 looks the same on a display calibrated to 2.2 within color managed app's.
2. Outside such app's they do not.
3. Non color managed app's don't know what a display or working space profile are ..... the condition of the display gamma now has an effect on the preview.