Slashdot Mirror
What is the Best Multi-Monitor Calibration Tool?
sojourndeath asks: "I am looking for a good way to calibrate multiple monitors (30-40), so that their color looks similar? It seems like everything I find is for profiling your monitor to your printer and scanner. I need to be able to have a bunch of users see the same color on any monitor? Does anyone have a good, accurate way of doing this?"
You're going to have to calibrate each monitor separately using the same calibrator.
Repeat once a month or so.
I don't envy you having to do this!
Some people are like slinkies--basically useless but they bring a smile to your face when pushed down the stairs.
I hope that they're all the same brand, or as the previous post mentioned i dont envy you. Some brands (Barco for one) have *luxury* an auto-calibration tool for some of their monitors. From observation, it seems to take around 5 minutes per monitor. I hope this helps. otherwise L.E.D.s are sensitive to wavelengths similar to that at which they radiate. Perhaps if you obtain a multimeter and measure the voltage of each led for a given (standardised) setting for each monitor, at least you should have a sort of way to quantitatively compare the brightness/colour balance. Just a thought
If all you have is a hammer, everything looks like a thumb.
> I need to be able to have a bunch of users see the same color on any monitor?
If you wanted to have the _same_ user see the same color the same way on
different monitors, that is theoretically achievable with good quality CRTs,
assuming you can put them in identical settings and so on.
But with different users, there is going to be a difference in perception.
Some people see *significantly* more color depth than others, for instance.
Also, some people's retinas are more sensitive to light than others, so they
have most of their color resolution in the darker ranges; other people have
eyes less sensitive to light and distinguish brighter colors better.
I've discovered that most of my coworkers can't tell #305050 from #294D4A,
even when they're side by side. To me, they're noticeably different in
character, and if you show me one of them by itself, I know which of the
two it is. (This is probably attributable more to the difference in
blue/green balance than the slight variation in brightness, but anyway, I
can tell.) One time I asked for a coworker's opinion on the brightness of
a certain background, and she said it was too dark, so I grabbed the V
slider (in Inkscape) and lightened it up a bit, then looked at her; she
obviously didn't realize I'd changed it at all. So I dragged the slider
over a bit more, and a bit more... after a bit I asked her how that was,
and her response clearly indicated she still didn't see a difference. I'd
changed it by probably 20 or 30 units per channel. (I quit asking for her
opinion on colors after that.) She's an extreme case, obviously, but the
basic phenomenon is universal: people don't all have the same eyes.
Cut that out, or I will ship you to Norilsk in a box.
Get a Pantone compatible monitor calibrator and software.
Like this one.
Not all monitors have the same range of brightness/contrast and colour* gamut (range of colours they can display). If you do achiveve your stated goal, it will only be by making the best monitor you have display the brightness/contrast and gamut of the worst.
Colour perception depends a lot on environmental conditions. On identical perfectly calibrated monitors, colours will not look the same if one is in a room with white walls and the other isn't... and the same goes for one being in a room with flourescent lighting, one being in the shade, one with a window behind it, or one being somewhere there is a pretty sunset happening outside the window.
Users will disagree about the extent of variations caused by environmental conditions, and will disagree about colours. If you do calibrate with the best calibration tool on earth, users will simply not believe that you've done it right, and will resent their monitors being 'wrong' (ie different to the way they were before calibration).
Monitors drift, especially cheap ones... as they warm up, as room temperature varies, and as they get old. Calibration is a neverending job.
* I'm English, from England, and I know how to spell English words. It's not my fault the founding fathers didn't take a decent dictionary to America.
A pizza of radius z and thickness a has a volume of pi z z a
This is one of those ambiguous uses of the English language.
When I read it, I took it to mean,
I want any given user out of a set of users to see the same colors on all monitors.
It appears that you took it to mean,
I want every user out of a set of users to see the same color on any single (or) on all monitors.
Since the phosphers are fairly standard from monitor to monitor in the same manufacturing run (ie, red gives of a certian wavelength across a range of monitors) then it's easily possible to make it so that any group of monitors made at the same time produce the same color regardless of the user's perception. Blue may look different for user A than for user B, but user A will see that same blue on the other monitor that they saw on the first, and user B will see the same blue they saw on the second monitor that they saw on the first.
I imagine that phospher emission wavelength is fairly standard within a narrow range, but I suspect that different models, even within the same original manufacturer, may have slight variations in emission. Therefore to get the best matching color you'd really need to purchase high quality (meant for imaging applications) monitors from the same manufacturer at the same time. You can usually tell that they're high quality and meant for imaging applications by inflated cost and included calibration device, though it's not universally true (ie, if you don't know any better, then this is about as close as you're going to get without a lot more research)
-Adam
Other posters have gone to great lengths to explain how color perception is environment and beholder dependent. So calibration is of little value. More importantly if you are generating images to be displayed on any users screen, then you have no control over their brightness, contrast, gamma or white balance or the end user's color experience.
But if you want to calibrate a monitor, I can tell you how I used to do it in broadcast.
First you set the black of the monitor: Generate a black screen image. Adjust the cut-off for each color so it just barely illuminates the phosphor. When finished, black is barely perceptable and has no pronounced color.
To calibrate white: Generate a full red screen, 100 hue and brightness, and then use a calibrated light meter to set monitor output to the color temperature of the red component of your final white. Do the same for green and blue. Display an all white screen and see if the screen is proper temperature. Check that the values of black didn't change. Get a feel for where your monitor best performs and run the monitor in a manner that doesn't cause blooming. In other words make sure that the full white value is not beyond the luminance output of your monitor.
Once black and white are correct, display a black and white stairstep signal. If all the channel gammas are correct, the steps should appear even to the eye and all the steps should be grey. If not, the trick is to adjust gamma of the color you don't see to correct the problem. Gamma correction can be very gross and you might not be able to make every step grey.
These steps correct the color balance of a monitor, but you still need to check purity, pin cushioning, convergence, horizontal and vertical linearity before you can be sure that the image on one monitor is the same as the image on another calibrated monitor. I can't image why you would go through this type of trouble.
Of note: The same calibration issues can be applied to audio. Years ago I wired up a new audio system at a recording studio. The studio had done several gold albums including one by the Rolling Stones. All the mics were adjusted to remove bandwidth irregularities. The engineers recorded and set levels for all sessions by listening to the audio from huge JBL speakers set-up with perfectly flat amplifiers. However, when they went to generate the final mix, the did it by listening to the audio through cheap 5 inch speakers. In this manner, they could provide the best listening experience for the majority of users.
machinator omnis sine licentia
after wading through a couple dozen posts of hopelessly useless pedantic crap, i figured i'd offer a reasonable suggestion: check out the colorvision spyder 2 calibration tool. it's relatively inexpensive, supports windows and mac, and is widely used throughout the industry for photo manipulation and graphic design workstations. combine with a print scanner, and you can get full start-to-finish calibration of your workflow process. here's a review of the previous model.
as some others have noted you can plan on recalibrating at least once a month, particularly with new monitors. if color accuracy is less important than precision (that is, it doesn't matter if the color is correct as long as it looks the same everywhere), make sure you are using the same model of monitor on each desktop as each phosphor combination used in a given model of tube produces a different color gamut. in all events, stay away from lcd - the gamut is crap and they don't hold calibration well.
I am looking for a good way to calibrate multiple monitors (30-40), so that their color looks similar? It seems like everything I find is for profiling your monitor to your printer and scanner.
"Calibrating" a monitor means adjusting its gun controls until the color output matches a standard for voltage vs. color. "Profiling" a monitor means telling it to output color numbers (r,g,b) and recording the difference between the actual and expected colors for a given color space. The profile also records the range of possible colors that the device can output (the gamut).
When the computer has a profile for the monitor, it "knows" how the monitor behaves. It can internally adjust the color numbers so that a document can be displayed faithfully. It may attempt to map each color in the document to the correct output color, or it may transform the gamut so that the document *appears* similar to the original (perceptual matching).
How does it know what color is in the document? The document has it's own profile. Same with the printer and scanner. And they all have different gamuts.
A profile is a mapping between color numbers (e.g., RGB) and actual absolute colors (e.g., L*a*b), along with gamut information. So saying "profiling your monitor to your printer and scanner" doesn't mean all that much.. each device has a profile that translates from color numbers to absolute colors and hopefully after all that translating back and forth you get the results you want. Often the marketing materials will talk about "matching scanners to prints to the screen" but that's not exactly what you're doing when you create a color-corrected workflow.
The best thing to do, after learning a little about how it works of course, is to get an inexpensive color meter like the ColorVision Spyder or the GretagMacbeth Eye-One and using it to calibrate and profile each monitor. You won't be able to "calibrate" it fully of course unless you can individually tune each gun, like on the Sony Artisan.
I have a Spyder, it's "ok". The software for the Mac is garbage that looked like it was ported from Windows by a 2-year-old, but it's "good enough". The Eye-One is a good choice.
You will find that this is a *very* subjective activity. If each monitor is in the *exact same* environment, you may achieve your goal of color uniformity. But if one monitor is in a sunny office, the other in a darkened room with only 5000K ambient lighting, people will *perceive* different color on each. Your users will just have to learn how to mentally adjust, just like a recording engineer learns his room and speakers, etc, etc.
If you can put all 40 monitors in a big rack and see them all at once, then at least you can get them all close to the same baseline. You'll have to repeat this every 2-3 months, preferably every month.
Once upon a time we used to adjust 20+ monitors in a television control room manually by following these steps:
;)
1. Send colorbars from the same source (if possible) to all the monitors;
2. Kick each monitor into blue-only mode, which turns off the green and red guns;
3. Adjust the contrast, brightness, tint and color (saturation) so that all the bars look the same;
(You see, color bars are set up so that, when viewed on the blue gun only, adjusting the tint adjusts the brightness of two bars in opposite directions, the brightness another two bars, and so on. To adjust the contrast, you twiddle the contrast knob and look at the two associated bars -- one gets brighter, one gets dimmer. You set it such that the two bars appear to be the same brightness. Repeat for the other controls.)
4. Pop out of blue mode, and all the monitors look essentially the same. Piece of cake.
Of course, computer monitors don't come with a blue-only mode, and I believe even component monitors pull the sync signal off of green, so you couldn't just unplug the red and green.
So perhaps this advice isn't helpful. But if anyone out there is trying to calibrate TV monitors...well, glad I could help.
Different people really do have different amount of the red green and blue sensitive cones in their eyes, in fact, 1/1,000,000 has a mutation where they have 2x as many green receptors as any other color receptor. There's an exhibit at the Exploratorium where there's an orange dot that's really an orange wavelength of light, and then surrounding it there are differently proportioned red/green light mixes, and different people see that central orange dot to match WILDLY different surrounding mixed color dots. The exhibit's in the seeing section, for those that care. It's striking to take a group to that exhibit and see the variety of responses! So based on that exhibit, and the nature of monitors (being an amalgam of red/green/blue to approximate continuous color), I believe it to be impossible to calibrate a set of monitors to look proper to everyone. One might be able to get them so they all look the same to everyone, but they could all look very wrong to some of the people, etc...