Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Retina MBP display calibration?
- Thread starter withfilm
- Start date
- Sort by reaction score
I am sorry but buying $200 worth of hardware just to get the colors on your RMBP right is just plain retarded and phony.
For your information OS X has a built in color calibrator which works fine. But thats not the point. The point is I am not going to be told by Company X what colors are the most accurate. I let my own eyes decide what looks the best. I have a phone with a sAMOLED screen and almost everyone agrees with me that it looks stunning and makes the iphone4 look so washed out. Well the purists are going to say that the colors are over-saturated.
And who says what the right colors are anyway? Company X? Don't make me laugh. Change your lightbulb from a warmer hue to a cooler hue and suddenly you have a different perception of the colors on your RMBP.
For your information OS X has a built in color calibrator which works fine. But thats not the point. The point is I am not going to be told by Company X what colors are the most accurate. I let my own eyes decide what looks the best. I have a phone with a sAMOLED screen and almost everyone agrees with me that it looks stunning and makes the iphone4 look so washed out. Well the purists are going to say that the colors are over-saturated.
And who says what the right colors are anyway? Company X? Don't make me laugh. Change your lightbulb from a warmer hue to a cooler hue and suddenly you have a different perception of the colors on your RMBP.
I really can't tell if you're a bad troll, or just too stupid to own a computer.
Last edited:
One of the main reasons to calibrate a display is to match what you see on screen to what you see on paper (print). Therefore, there definitely is an objective standard.
The point of calibration is not to make colors "look good/better" (yes, you're right, if you wanted the a screen's color to look "best" to you, you would be the best judge...but that's not why people buy hardware calibrators).
Edit:
(and that's also why you when change your work place's lighting, you should recalibrate your display...)
You are 100% correct. In example with my Samsung HDTV, I didn't "calibrate" the settings on it. I simply ste them all to look most pleasing to my eyes. On my TV I prefer rich saturated colors as opposed to being true to life (play lots of video games). On my mac I prefer to use one color profile for editing photos (just personal shots from dslr nothing "pro") and another color profile set to what my eyes think looks the most stunning lol.
I have a Spyder 3 Elite. The damn thing can't seem to calibrate the LED panels properly; my retina MBP comes up with a very pink tint by default and the Spyder 3 does very little to fix it; my CCFL LCDs come up nearly identical despite having various tint issues (Dell 30", Apple 30" and 20"). My Thunderbolt panel doesn't calibrate too badly but still more red tint than I would like.
THe issue here is I can't for the life of me calibrate the rMBP panel to something that resembles any of my other IPS panels including the Thunderbolt panel, 27" iMac, HP or Dell IPS-based panels.
I have about a week left to return my rMBP with its LG display. Would purchasing a high end calibration system solve my issues? Do you know where I could get an ICC profile from one of these systems to see roughly what I should be expecting if I forked out the cash?
A spectrophotometer such as the ColorMunki or i1Pro should give you much more consistent results across displays, especially if they are using different backlighting or are different technologies. (LCD, Plasma, DLP etc.) At least that was my experience after having several colorimeters and then buying an i1Pro.
Before you go out and buy new hardware though, try running the trial of ColorEyes Display Pro on all of your displays with the Spyder and see if that helps at all: http://www.integrated-color.com/cedpro/coloreyesdisplay.html
Make sure you enable the advanced options, turn on white point tuning, and set your white point to 0.3127, 0.3290 rather than "D65" (they should be the same, but I believe CEDP uses "6500K" for its "D65" setting) with a gamma of 2.2
While I would expect better matching between your displays when using a Spectro, calibration can only get you so far. It may not be possible to have your displays be an exact match.
Unless you are only comparing displays that have the same specificationsor even better, several of the same displays, youre probably never going to get an exact match.
And the best way to match displays is to have them all calibrated to match the lowest-common denominator between them, which is obviously less than ideal, and reduces the quality of your best display down to that of your worst.
This is even a problem when you get into really high-end displays and calibration gear such as Sonys $20,000 OLED monitors. Most calibration hardware shows them to be a match with other displays, but to the human eye, they are clearly different. http://forums.creativecow.net/thread/277/13743
This is extremely dated information. Beyond that the ColorMunki spectrophotometer is not a good choice at all. It falls below any recent colorimeter generation for more money. The i1 Pro solution was something common due to their use by many people for measuring printer output already and a lack of suitable colorimeters for the most recent display generation. Check out
http://www.drycreekphoto.com/Learn/Calibration/MonitorCalibrationHardware.html
Note that the spectrophotometer version of the colormunki is terrible, absolutely terrible. I'd never suggest anyone try to use that on their display. It's asking them to drop $450 for a headache. A lot of people then try to work things out with support and go past the return limit. I'd suggest they don't do this in the first place. The spectrolino being a solid device isn't anything new. They're just old and no longer supported, which is bad.
Really if you're going to look at the i1 pro, at least look at the new one.
http://www.luminous-landscape.com/reviews/accessories/eye_one_pro_ii.shtml
I get why you're saying to use a spectrophotometer (larger field of view, not filtered for a limited spectrum), but it's not a perfect solution, and those were just bad choices of hardware. The i1 Pro was never that great. It was just what was available in its price range. Keep in mind the older colorimeters like the DTP-94, Spyder 2 (and 3), and i1 display 2, were all based on much older algorithms. They were still sticking it out from the late crt era.
Okay last point would be that a match isn't guaranteed regardless. Displays are different, and software packages include emulation features + "tweak" adjustments to compensate for this. Think of how some of the super expensive soft proofing packages worked. They'd certify very specific hardware used under very specific conditions. I don't think trying to match your iphone or plasma to your rMBP is practical. They all have different native hardware characteristics, so it almost comes down to the largest shared set of points, and even then the data is interpolated to a large degree.
Bleck I wanted to quote this too. This isn't even a good idea much of the time as you're addressing the data in a very awkward high level adjustment kind of way. The display on the rMBP and the thunderbolt display have no internal addressable LUT, so you can crush shadow detail or cause banding if you take it too far. Personally given the way Apple's hardware is set up, unless you're trying to match a specific standard or other hardware target, I'd just hit white point native and let it run a profile. I'd set the luminance to match. If it's noticeably far apart, I'd go from there. My main concern was the two spectrophotometers you suggested as the colormunki sucks and the i1 pro will eventually be de-supported. I give it 3 years, but if you're going to spend that much there's no reason not to get the newest one.
Last edited:
Thats definitely an interesting test, but none of the measurement data is provided, they are only testing LCD displays, and they do not specify which displays were tested.
There was only max/mine dE2000 values posted for the meters, and no mention of what the measurements taken were. It looks like they are probably only greyscale measurements, where colorimeters will usually do OK with.
An accurate test would have a minimum of:
- 11-point greyscale data using dEuv
- RGBCMY data using dE94 or 2000.
In tests I have seen performed across various display technologies and LCD backlighting types, the i1Pro and ColorMunki came out on top above all colorimeters once you started taking gamut measurements from the displays. These tests included high-end colorimeters such as the Klein K-10. (reference meter was an ORB optronix SP-100 Spectroradiometer)
The thing with a colorimeter is that it will perform very wellabove any consumer-grade spectroif the target displays gamut and spectral characteristics happen to be a good match for its optical filters, and you are testing a new meter rather one that has been in use for a while. (colorimeters degrade over time whereas most spectro designs are relatively stable)
The problem is that without a reference meter to hand, you have no way of knowing whether the target display is indeed a good match.
The i1Pro 2 hardware revision primarily addresses measurement stability over time. Accuracy is on-par with the original i1Pro hardware. You will actually find that the i1Pro 2 hardware labelled as an "i1Pro Rev.E" device. (the last "i1Pro 1" was an "i1Pro Rev.D")
The colorimeters I recommended actually came second in the test you posted. I would still prefer a spectro over these.
Right, I dont think it is a good idea to be profiling to the lowest common denominator just to get a closer match, that was my point.
I'd agree with you there. The site is a bit fragmented in some ways. They go into displays they've tested in other portions, but it isn't completely unified. I noticed Chromix is mentioned there, and they do some of their own testing, but it's presented in more of a blog fashion.
I've noticed the same thing. The i1 display pro/ colormunki display (note the colorimeter version, the name is used for different devices) do perform well with some of the Adobe RGB displays. I agree that it's a generic approach done for cost reasons, but once you're up to NEC/Eizo there is some level of testing. NEC offers oem packages with tested colorimeters for their displays. Eizo uses lookup tables in their software for any colorimeter they support. It's not the same thing as having a device 100% matched to the characteristics of the display, but none of these solutions are completely infallible. Displays aren't even 100% uniform and you're generally basing the calculations off the center (although some do have options to test this). If you scroll down the page on that link, they compare against sRGB and "wide gamut" as in greater in volumetric representation than sRGB . In the ones labeled wide gamut, the colormunki scores poorly against their reference radiometer. In the case of Apple displays, they are within the sRGB realm.
It still scored much worse than the colormunki display which is a colorimeter. Perhaps that's the one you were suggesting came in second? That's basically an i1 display pro, but you can't use it with non X-rite software. The spectrophotometer version is labeled "photo/design". It did not come in second, and the first chart is just unit to unit variability.
When were they performed? If they were older tests, more than a year ago really, I'm not surprised. Even then the colormunki was a poor choice, but none of the colorimeters made at that time were really designed for some of the recent display generations.
They were testing against a reference grade radiometer. It mentions a Photo Research PR-730 was used for the tests.
The colormunki photo/design (spectrophotometer) measured
"standard gamut" mean error 4.9 White, 10.6 black.
"wide gamut" mean error 6.1 white, 15.9 black
These are basically terrible. You'd be better off leaving it at the out of the box profile in most situations compared to such results. The validations are likely to come in cleaner assuming that colors measured are within reproducible gamut even accounting for manufacturing variations in the displays.
The colormunki display which is a colorimeter scored 1.1, 2.0, 1.7, 2.8.
I think you misread the chart somewhere, but this is why I've stated that X-rite's naming convention is very poor.
I don't understand why people share calibration profiles. Sort of defeats the point of calibration if you're using one from another display!
You just spent $2000-4000 on a computer; if you need accurate colors usually means you spend a ton more on camera equipment, printers, etc; spend the $100 or so on a good calibrator.
I should point out the point of calibration isn't what you think looks good. It's so WYSIWYG works. You want the picture you took to match the colors you need, and it needs to look that way on the computer, and when you print it, it needs to look right as well.
You just spent $2000-4000 on a computer; if you need accurate colors usually means you spend a ton more on camera equipment, printers, etc; spend the $100 or so on a good calibrator.
I should point out the point of calibration isn't what you think looks good. It's so WYSIWYG works. You want the picture you took to match the colors you need, and it needs to look that way on the computer, and when you print it, it needs to look right as well.
I guess this thread might be the right one to ask. Does anyone find that blues tend to be a lot more saturated (or something, not a graphics pro here) compared to pretty much any other IPS display? I've compared to 2 NEC 2490wuxis, an NEC 20WMGX2, a Dell 2209WA, and an IPS Dell 2007FP. No matter how much I calibrate it, I can't eliminate the difference
----------
I understand the sentiment, but can't printers render colors a bit off, and can't cameras capture the colors a bit wrong? Most people here won't be able to ensure color parity between all their devices anyway, so they may as well go for what looks good to their eyes.
Also if someone primarily designs computer graphics, I'm not sure calibrating with anything more than the eye is essential. Most consumer displays are off compared to the "ideal" calibration points anyway. And as seen as in this thread, many actually like how the inaccurate pictures looks more. So when your end results are meant to reach that audience, it might actually be better to use an out of the box calibration
----------
I understand the sentiment, but can't printers render colors a bit off, and can't cameras capture the colors a bit wrong? Most people here won't be able to ensure color parity between all their devices anyway, so they may as well go for what looks good to their eyes.
Also if someone primarily designs computer graphics, I'm not sure calibrating with anything more than the eye is essential. Most consumer displays are off compared to the "ideal" calibration points anyway. And as seen as in this thread, many actually like how the inaccurate pictures looks more. So when your end results are meant to reach that audience, it might actually be better to use an out of the box calibration
One that works even moderately well with LED backlighting will run you closer to the $200 realm
. Colorimeters are a bit limited by their nature in a lot of ways, and only the really really recent ones work truly well with any of the recent display models. You won't eliminate the color difference. There's a reason I dislike the term calibration. It's misleading. I've tried to explain this. The things take a profile of the information, make adjustments based on that to the output fed to the display, and that is what you get. You can find articles on matrix and LUT systems. Many of these fall in the former category. They provide a basic description of gamut using a number of test points. It's not the same thing as really setting every level in the display, and beyond that displays have different gamuts and characteristics. They will look different. In some software you have modes designed for emulation or tweaking adjustments. These still aren't perfect. Even with emulation, you're essentially going for lowest common denominator, and it's not an exact science (again just clipping outer boundaries for the most part). Assuming a good profile, it will be a somewhat accurate description of the device, so adjustments made in color managed applications will at least be predictable.
Regarding Apple, they optimize their profiles in a weird way. On my 2011 macbook pro, the out curves suggested that most of the adjustment was to blue and green to match an oddly high red gamma on its out of the box profile.
ICC calibration does not adjust display gamuts. It profiles the display, so that colour management aware software can display colours as accurate as possible on the device. If your display has a wider gamut with blue, it should display colours in-gamut the same as other profiled displays, but may also be capable of displaying colours outside the range of the other screens.
This is why you also need to profile your printer, and use soft-proofing to emulate the printer output on your display. It still wont be perfect, but you can usually get very good results first try when doing this.
Cameras can also be profiled by photographing an XRite ColorChecker under the same lighting conditions as your colour-critical shot. (you can also make general calibrations that should suit most situations if you are using Adobe software)
You should always design on a display that is as accurate as possible. Over time, display accuracy has been gradually improving, to the point where even (relatively) cheap devices like the iPad 3 display sRGB imagery with very high accuracy.
You can never assume anything about the kind of displays your audience is going to be viewing your content on.
Osx is generally color aware is it not? And I know photoshop is. In both cases the colors on the blue end are dissimilar enough that I cannot really trust it. At first I thought it was just my comparing to an aging monitor, but even after comparing to new displays (even other apple displays), I started to notice the blues look different compared to mine. I know it's not a white point issue either since actual white images look very similar. As far as I'm aware the rMBP screen is supposed to be 99% sRGB. So shouldn't it look pretty close to other sRGB screens when calibrated?
How many people here do you think actually calibrate their printers and cameras - more importantly, how many of them do you think actually need to? My point was that do many people here stress over improper calibrations and chide others for using inaccurate methods... But at the end of the day even a calibrated display won't help most consumers. Most consumer printers and cameras cannot be calibrated anyway. So what's the point if all of their devices still won't match?
And even with IPS panels starting to proliferate, I would argue TN panels are still far more common. I'm sure one could conduct a survey to figure out what their audience generally has. And I think it is important to take into account what kind of display your audience will have, because ultimately it is their experience that matters. In the days of CRTs, one was able to count on his/her customer base having similar views. Today, not so much. In fact if i were a designer, I would probably get an example of an IPS, TN, and MVA/PVA and preview the content on each to see if there are any noteworthy issues.
Last edited:
What device are you using for calibration, and what are the display gamuts like?
All displays and devices should be calibrated if you actually care about accurate colouror as accurate as the devices allow.
You should be able to profile just about any printer, and Adobe software can create a profile for any camera that can shoot RAW.
Panel type has no significant bearing on display gamut or colour accuracy.
If you design for a standard reference (typically sRGB for the Web, for example) then you stand the best chance of having your content look good.
If a display is deficient in blue for example, then everything will be equally deficient in blue if it were designed for a standard target, so your content would match.
If you designed content tailored to that one display, then it will look wrong on every other display out there.
Hoping you guys can help me with something. I just calibrated a rMBP as well as a 27" Cinema Display using an i1 Pro and the X-Rite i1 Profiler software.
But the colors don't really match. It's subtle, but the Cinema display seems to have a slightly warmer cast, maybe even a touch of green compared to the rMBP.
I'm a bit confused by this as I would have assumed the calibration should have evened out the color variations between them. Perhaps I'm just up against a physical limit of one or the other.
It's a bit frustrating because I was planning on using a dual monitor setup, but the color variation between them is a bit distracting.
White balance for laptop reads 6509 K
White balance for Cinema display reads 6503 K
So at least in theory the whites should be very close, but I can't shake this subtle green cast on my Cinema display.
So are any of you running a rMPB next to a Cinema or Thunderbolt display? Are you able to get colors to match exactly? It really is pretty subtle, but I'd like to fix it if it's possible.
But the colors don't really match. It's subtle, but the Cinema display seems to have a slightly warmer cast, maybe even a touch of green compared to the rMBP.
I'm a bit confused by this as I would have assumed the calibration should have evened out the color variations between them. Perhaps I'm just up against a physical limit of one or the other.
It's a bit frustrating because I was planning on using a dual monitor setup, but the color variation between them is a bit distracting.
White balance for laptop reads 6509 K
White balance for Cinema display reads 6503 K
So at least in theory the whites should be very close, but I can't shake this subtle green cast on my Cinema display.
So are any of you running a rMPB next to a Cinema or Thunderbolt display? Are you able to get colors to match exactly? It really is pretty subtle, but I'd like to fix it if it's possible.
Did you calibrate the two displays at roughly the same time, in the same room/same lighting conditions? Even though it's not "supposed" to be the case, I've heard that the variance in ambient light during calibration can have a slight effect on the final profile.
Yep. I even did it twice...the second time I did it in absolute darkness thinking ambient light might have been creeping in under the glass facade of the Cinema display. Also did the second one with the highest number of color squares they offer - didn't seem to matter.
Always thought my Cinema display calibrated nicely, but now I feel like the rMBP actually looks more natural (skin tones, etc.).
Has anyone successfully gotten the rMBP and a Cinema Display to look really close to each other? I wonder if using a different software like ColorEyes Display Pro would be any better.
Is there any sense buying these programs and calibration devices if I'm not a photographer or video editor? I also read reviews that state how their display brightness went very low. I really like my 7000K display, is that going to change or I can manually configure it without loosing the colors?
I think so. I was introduced to calibrating my display because of photography, but now when I look at uncalibrated displays (those that are really off), I am irritated by them. It's almost like when you go to looking at a regular display after owning a retina display. You never knew what you were missing before. But once you see a properly calibrated display, those that are not will stick out like a sore thumb.
Absolute darkness isn't a bad idea. I should note that it's normal to leave the colorimeter plugged in a few minutes prior to running the software. It's also a good idea to let both displays warm up a minimum of 30 minutes. In some cases notebooks have given strange results. You can usually find bug reports if you check the developer's support forum meaning X-rite or Datacolor, depending on the colorimeter you purchased. Feeling natural is a difficult thing to determine. I could show you tricks that would make black and white imagery appear to be in full color. So much of it is suggestive. The typical method for photographers is compare to printed values. Colorists have a different method of verifying the hardware.
Well the hardware always drifts with time. Even colorimeters eventually drift as they age. I wouldn't worry if it's measured 6500 or 7000 at a given range. I mean being able to hit D65 is obviously ideal, but you do drop some colors when attempting to remap via profiling. It's basically divided into two parts. Part of the profile describes the hardware behavior in terms of reference values to facilitate color management. The other part assigns a set of instructions to modify the final output. If you're leaving everything native, it can still measure behavior to a "reasonably" accurate degree assuming good uniformity from center to edge. If you're not that picky and it looks fine to you, I wouldn't worry about it. The one scenario I find really silly is when someone picks up a used display that looks like crud and expects profiling it to fix everything. That doesn't work. Think of this as something that helps better match a consistent target if it's already most of the way there.