Retina MBP display calibration?

[millerrh]

So I downloaded ColorEyes Display Pro and did another round of monitor calibration. This software is pretty cool because it allows you to check your results afterwards

I learned two things:

1) X - Rite's software and ColorEyes Display Pro give almost identical profiles.
2) I think the reason I thought my Cinema display had a green cast was because my Retina display actually has an uncalibratable red cast!. You know how when you have orange sunglasses on and you take them off how everything else looks blue? Same thing must have been happening here. I was looking at my red Retina display for so long those whites seemed white, then when I look at the Cinema display it seems green in comparison. Your eyes automatically adjust to the color temperature so white seems white.

Attached are the calibration check results so you can see what I mean. Cinema display has colors with extremely low Delta E numbers, whereas the Red on the Retina display is not very good.

 

[JohnDoe98]

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.

So if I understand this right, the point of this stuff isn't simply to match printing with the screens, but to actually make the screens look better? I don't print anything, but I would like to get my screen to perform as best it can. When I calibrated my projector my colors and contrasts improved, I had deeper blacks and whiter whites. I was able to minimize the red push that my particular projector had. Does the same apply here with laptops? After calibrating with Spyder4Pro for instance, my movies will look better on the laptop, and the colors won't look as washed out?

 

[millerrh]

So if I understand this right, the point of this stuff isn't simply to match printing with the screens, but to actually make the screens look better? I don't print anything, but I would like to get my screen to perform as best it can. When I calibrated my projector my colors and contrasts improved, I had deeper blacks and whiter whites. I was able to minimize the red push that my particular projector had. Does the same apply here with laptops? After calibrating with Spyder4Pro for instance, my movies will look better on the laptop, and the colors won't look as washed out?

In theory yes. But there are also limits to calibration as well. Your calibration hardware will only calibrate as much as your monitor lets it. For instance if you look above I still can't seem to get rid of a reddish cast even after calibrating

 

[JohnDoe98]

In theory yes. But there are also limits to calibration as well. Your calibration hardware will only calibrate as much as your monitor lets it. For instance if you look above I still can't seem to get rid of a reddish cast even after calibrating

Oh I see. I thought you were having trouble since you were trying to match two monitors to one another, not because one of the monitors couldn't be completely fixed. I misunderstood. That's rather disappointing. Do you happen to know what screen you have, a Samsung or an LG? Is the Red push very strong?

Now I'm really unsure if its worth buying any calibrating software since it seems there isn't any guarantee it'll make things appreciably better.

Well thanks for your input.

 

[millerrh]

Oh I see. I thought you were having trouble since you were trying to match two monitors to one another, not because one of the monitors couldn't be completely fixed. I misunderstood. That's rather disappointing. Do you happen to know what screen you have, a Samsung or an LG? Is the Red push very strong?

Now I'm really unsure if its worth buying any calibrating software since it seems there isn't any guarantee it'll make things appreciably better.

Well thanks for your input.

The point of calibration is that it changes the way your monitor displays colors (and blacks and whites) so that it matches a known standard. So in theory if you calibrate one device it will look like a different calibrated device. And therefore two monitors would be calibrated to look the same. This way the way I see red or blue or a gradient is the same way you do. It's the only way to know you are looking at the same image. And if you're doing photography editing and making decisions based on color, you need to know it's correct. Otherwise I could be making a change based on the cast of my monitor and you would see it differently.

That being said, some monitors calibrate better than others. Laptop monitors historically have had limitations and nobody would really do critical color work on a laptop monitor. Somehow I thought this retina monitor would be different. And it looks different, like it has so much potential. In fact, the default calibration "Color LCD" that comes with it almost looks better (and closer to my cinema display) than what I was calibrating to. So that's why I almost thought it was my software.

I've got the LG monitor on mine which I know will just develop IR. So I'm tempted to return the thing even without IR to try to get a Samsung. Maybe it calibrates more effectively.

 

[snoylekim]

The point of calibration is that it changes the way your monitor displays colors (and blacks and whites) so that it matches a known standard. So in theory if you calibrate one device it will look like a different calibrated device. And therefore two monitors would be calibrated to look the same. This way the way I see red or blue or a gradient is the same way you do. It's the only way to know you are looking at the same image. And if you're doing photography editing and making decisions based on color, you need to know it's correct. Otherwise I could be making a change based on the cast of my monitor and you would see it differently.

That being said, some monitors calibrate better than others. Laptop monitors historically have had limitations and nobody would really do critical color work on a laptop monitor. Somehow I thought this retina monitor would be different. And it looks different, like it has so much potential. In fact, the default calibration "Color LCD" that comes with it almost looks better (and closer to my cinema display) than what I was calibrating to. So that's why I almost thought it was my software.

I've got the LG monitor on mine which I know will just develop IR. So I'm tempted to return the thing even without IR to try to get a Samsung. Maybe it calibrates more effectively.

I ended up using BasicColor5 ( trial) to calibrate my rMBP w/ an LG ..gave me the best calibration .. I didn't have 'red push' or anything excessive w/ any of the software .. The report and zipped profile are in another calibration thread in this forum if you want to give that a look and possibly try that software or profile to see if it eliminates the red bias you're seeing ..gives good evaluation on the final wihite point reached, etc .. I also have a 24" Cinema which was calibrated . both came out very good and close to identical ( close enough for my uses) ..

 

[millerrh]

I ended up using BasicColor5 ( trial) to calibrate my rMBP w/ an LG ..gave me the best calibration .. I didn't have 'red push' or anything excessive w/ any of the software .. The report and zipped profile are in another calibration thread in this forum if you want to give that a look and possibly try that software or profile to see if it eliminates the red bias you're seeing ..gives good evaluation on the final wihite point reached, etc .. I also have a 24" Cinema which was calibrated . both came out very good and close to identical ( close enough for my uses) ..

Thanks I'll give BasICColor a try tonight as well as load your profile in to see if it looks any different. This color cast thing is driving me nuts.

 

[PkennethV]

After calibrating with Spyder4Pro for instance, my movies will look better on the laptop, and the colors won't look as washed out?

Will it look "better" is hard to say, because "better" is subjective/personal preference in this case, but it will definitely be more "accurate."

In the case of a projector, the projector apparently was flatter than the "standard", so calibration added contrast (among other adjustments to color, white point etc.).

But in the case of your display, your display may or may not be flatter than the standard. It may be overly contrasty (and since probably 90% non-professional, sub $1,000 display tend to be overly contrasty with its default settings, there's a good chance yours is).

So if you're display is more contrasty than the standard, the calibrated profile will actually make your screen flatter/more washed out than before...which people may or may not like...again, hard to say if you'll think it "looks better", because it's a personal preference...all that can be guaranteed, is that it may be more "accurate."


Personally, most people I know who first calibrate their displays think it looks "worse" (but they go ahead and stick to it because they need the accuracy for photography).

 

[snoylekim]

Thanks I'll give BasICColor a try tonight as well as load your profile in to see if it looks any different. This color cast thing is driving me nuts.

The profile I ended up with is the 'last' profile I attached in the thread ..Good Luck .. the report is out there as well .. I went for a D65 whie point ( which it got very close to) ..I also calibrated w/ the discrete Nvidia card selected ..I have my RMBP set for discrete always w/ gfxstatus .. figured I paid for the extra graphics card, and I'm not on battery very much ..

 

[millerrh]

The profile I ended up with is the 'last' profile I attached in the thread ..Good Luck .. the report is out there as well .. I went for a D65 whie point ( which it got very close to) ..I also calibrated w/ the discrete Nvidia card selected ..I have my RMBP set for discrete always w/ gfxstatus .. figured I paid for the extra graphics card, and I'm not on battery very much ..

Is this the posting?

https://forums.macrumors.com/showthread.php?p=16086310#post16086310

 

[snoylekim]

Is this the posting?

https://forums.macrumors.com/showthread.php?p=16086310#post16086310

yes , post #30 in this thread has my sRGB profile w/ attached report ..

 

[millerrh]

OK, did the test today with basICCOlor as well as loaded in your profile.

First thing I noticed was your profile made my display way warmer/more red than it already was. This makes me wonder if mine could have an already pretty red cast to it while yours was adjusting for a lack of red.

Second thing I noticed is this software also did not make a huge difference. The good news is that every software I've tried seems to work relatively the same. So I should be OK using my normal X Rite software that I can get for free.

End results/observations:

My Cinema display still seems to have a green cast to it relative to the retina monitor and/or the retina monitor has a redish cast. I cannot seem to calibrate this out.

The LUT graph of basICColor shows that maybe my Cinema display is not able to display all the colors (gets cut off at the upper right). Whereas the Retina display is pretty linear and might be the more correct out of the two of them. I am attaching both LUT graphs as well as an explanation on how to read them from their manual. So maybe it really is my Cinema Display that is green. Any help interpreting these?

I always felt that my iPad 3 had a really pure white. Sure enough, when I hold it up to both monitors, my Cinema display is greener and my Retina display is redder with the Cinema being a worse green.

My old 2010 MPB looks like crap now...not sure how I can go back to a normal laptop after looking at a retina, even when it's slightly reddish. Of course I almost always used it in clamshell mode and relied on my Cinema display. Now it is making more sense why my photos seem warmer when I view them on other displays sometimes - I must be compensating for the green cast.

 

[snoylekim]

Maybe ILukeJoseph will pop onto this thread .. he pointed me toward Basiccolor and has more experience than I do w/ calibration ( TVs and such) , and he uses the L* profile .. Since you're using the Retina in clamshell sometimes, are you fixing the graphics card w/ Gfxstatus ..wondering if maybe you'll get more consistent results w/ the nVidia ..
I haven't tried it, but it looks like BasicColor5 will let you tweak the calibration curves, so you could possibly tweak both and eliminate the red-green stuff you're seeing . Based on the reports , I'd personally want to get the grayscale much tighter .. closing in ..what fun ..

 

[millerrh]

Maybe ILukeJoseph will pop onto this thread .. he pointed me toward Basiccolor and has more experience than I do w/ calibration ( TVs and such) , and he uses the L* profile .. Since you're using the Retina in clamshell sometimes, are you fixing the graphics card w/ Gfxstatus ..wondering if maybe you'll get more consistent results w/ the nVidia ..
I haven't tried it, but it looks like BasicColor5 will let you tweak the calibration curves, so you could possibly tweak both and eliminate the red-green stuff you're seeing . Based on the reports , I'd personally want to get the grayscale much tighter .. closing in ..what fun ..

Any time the retina is hooked up to the Cinema display (clamshell or not) it always invokes the Nvidia. Since having the retina, I was actually hoping to stop using the clamshell mode and go with a dual monitor setup since the retina screen is so good, I don't just want to hide it away in clamshell mode. But this red/green shift is almost enough to make me not want to look at both screens at the same time. I was using L* for the gamma as well as that is what they recommended.

I saw where you could tweak the curves, but I don't know enough about them to do anything good by messing with it. Was hoping the hardware calibration would have accounted for this. And both monitors show they are within spec...but they are definitely not close to looking the same.

 

[thekev]

So I downloaded ColorEyes Display Pro and did another round of monitor calibration. This software is pretty cool because it allows you to check your results afterwards

I learned two things:

1) X - Rite's software and ColorEyes Display Pro give almost identical profiles.
2) I think the reason I thought my Cinema display had a green cast was because my Retina display actually has an uncalibratable red cast!. You know how when you have orange sunglasses on and you take them off how everything else looks blue? Same thing must have been happening here. I was looking at my red Retina display for so long those whites seemed white, then when I look at the Cinema display it seems green in comparison. Your eyes automatically adjust to the color temperature so white seems white.

Attached are the calibration check results so you can see what I mean. Cinema display has colors with extremely low Delta E numbers, whereas the Red on the Retina display is not very good.

First those Delta E numbers are meaningless. For some reason these software developers spin them to look good. If they were spot on, you wouldn't be able to spot even a minute difference between them the colors on the screen and a reference grade color chart under lighting of appropriate brightness of the same color temperature. Most of it is that they're picking easy targets, and accounting for a certain amount of colorimeter hardware variation.

There shouldn't be a significant problem like this. I'd have to look it up, but I'm pretty sure some of those programs have a "match" mode where it attempts to match behavior as well as possible. You're still likely stuck with the matrix profile behavior of attempting to manipulate curves at an instructional level. I've stated this before. It's not perfect, but a decent program will have some ability to make minor tweaks. It's just that curve form tends to be a bit too touchy.

I saw where you could tweak the curves, but I don't know enough about them to do anything good by messing with it. Was hoping the hardware calibration would have accounted for this. And both monitors show they are within spec...but they are definitely not close to looking the same.

Tweaking curves manually is a terrible idea. It's too easy to break profiles, even if you know what you're doing. Try doing it on a photo and see how quickly you can break the behavior. With a profile you have significantly less latitude, and I somehow doubt those offer enough control for fine tuning a match.

What do you mean hardware calibration? Some more expensive displays have internal LUT systems. Apple has never used such a system, so checking hardware calibration will be meaningless. Even with those, the levels are carefully set at the factory and the primaries are measured.

The point of calibration is that it changes the way your monitor displays colors (and blacks and whites) so that it matches a known standard. So in theory if you calibrate one device it will look like a different calibrated device. And therefore two monitors would be calibrated to look the same. This way the way I see red or blue or a gradient is the same way you do. It's the only way to know you are looking at the same image. And if you're doing photography editing and making decisions based on color, you need to know it's correct. Otherwise I could be making a change based on the cast of my monitor and you would see it differently.

That's the theory of it. It's also supposed to provide a way to track display behavior/gamut over time. With photography it's typical to proof check. Color grading and other things that don't involve hard copies have other methods. I should caution you that even two displays of the same brand, type, and batch can look different due to aging. This is why some software will have "match" modes. You can still validate the correlated versions against standard targets that are supported (although not always certified) by the software package. NEC and Eizo (maybe Quato) are the only ones I've found to be exceptionally easy in terms of getting reasonably good consistency. The last thing I should ask is if you're warming both displays up prior to running the software.

 

[millerrh]

Hi, thanks for all your input.

What I meant by "hardware calibration" is that I'm using a spectrophotometer to actually measure the colors to create the profile. You'd think that would take all the variability out of it. So any differences I'm seeing are likely due to inherent differences in the displays that go beyond what the calibration adjustment can account for.

At least that's my impression now.

I have not seen a match mode on any of the software I have been trying this out on, but it very well may be there. I'd ideally like to stick with X Rite's i1 Profiler as I already have a license to that by virtue of my i1 Pro device.

 

[thekev]

Hi, thanks for all your input.

What I meant by "hardware calibration" is that I'm using a spectrophotometer to actually measure the colors to create the profile. You'd think that would take all the variability out of it. So any differences I'm seeing are likely due to inherent differences in the displays that go beyond what the calibration adjustment can account for.

At least that's my impression now.

I have not seen a match mode on any of the software I have been trying this out on, but it very well may be there. I'd ideally like to stick with X Rite's i1 Profiler as I already have a license to that by virtue of my i1 Pro device.

Blarg I don't mean to hijack the thread, but I want to respond to this. First the newest colorimeters work well enough, even with LED lighting, although the i1 Pro is a solid device. I haven't kept up with all of them. I have i1 Profiler, Color Navigator, and Spectraview on this system. Color Navigator has multiple display matching. I thought i1 Profiler had a similar feature for some reason, but I can't locate it at the moment. I only use it to profile the macbook pro display. Regarding Basiccolor, they do have a mode called "hardware calibration" meaning much of the information is stored in the display. It can provide better results than the archaic method of using gpu output instructions within the profile data.

The differences you're seeing could represent a couple things. To the best of my understanding they use a transformation matrix to manipulate the behavior of output curves. Chromix also provides a small description in their wiki. There are several potentially limiting factors. One would be the target. Another would be the device, although with the possible exception of extreme shadow values, an i1 pro should do a very nice job. The last is just the nature of matrix based profiles. They tend to be limiting. i1 profiler has an LUT based mode, but I'm not sure how well it works for generating a standard profile. I'll look at it again later.

 

[iLukeJoseph]

Hi, thanks for all your input.

What settings exactly are you using for BasicColor5? The measured contrast ratio looks very low compared to what I measured (look my graphs in my Sig. rMBP Samung). Problem could be that the i1 Pro is not good for reading black levels.

Speaking of the i1 Pro, when was the last time it was Nist Certified? I believe they are supposed to be done once a year, that is really only if you use it for business purposes though. But it very well could be out of spec. Oh and what revision is it? Rev. D is the "good" one, and it is not the UV Cut model is it?

 

[Moshe1010]

Can somebody post here ICC profiles for Samsung display? I know that every display is different, but still, please.

 

[millerrh]

What settings exactly are you using for BasicColor5? The measured contrast ratio looks very low compared to what I measured (look my graphs in my Sig. rMBP Samung). Problem could be that the i1 Pro is not good for reading black levels.

Speaking of the i1 Pro, when was the last time it was Nist Certified? I believe they are supposed to be done once a year, that is really only if you use it for business purposes though. But it very well could be out of spec. Oh and what revision is it? Rev. D is the "good" one, and it is not the UV Cut model is it?

For BasicColor5
Color Temp: D65
Tonal Response: L*
Luminance: 120 cd/m^2
These are also the values I was using with i1 Profiler, except gamma of 2.2

Eye-one Pro
Rev B
My father gave it to me years ago...as far as I know it has not been calibrated. Very well could be the issue.
Was doing a little research on this today and ran across this review:
http://www.luminous-landscape.com/reviews/accessories/i1display_pro.shtml

In it, they say:

If you have a i1Pro or other spectrophotometer you may ask yourself – why bother getting an i1Display Pro colorimeter? A spectrophotometer is a more sophisticated instrument, is it not? Well, yes and no.

In speaking to some experts in the field what I have discovered is that a spectro is not as sensitive at lower luminance levels as is a colorimeter. This means less accurate profiling of quarter tones and image shadow areas on-screen. So, for those that have a monitor with decent dark tone performance, a colorimeter like the i1Display Pro makes a lot of sense, even if you already own an i1Pro device.

Makes me wonder if spending money on an i1 Display Pro might be a better use of money as opposed to getting this calibrated.

Sanity check
I have a low-end Pantone Huey Pro I use at work that I brought home today to try on my home screens. The result was the same thing, but even worse. Check out the picture. I drug a window so it covered both screens (Cinema display on top, Retina display on bottom.) This is what I'm dealing with. Now a Huey is not a good colorimeter, but it does sort of do a sanity check to check the operation of the i1.

I also hooked back up my old MBP and while it's not perfect, it definitely matches my Cinema Display a lot better. It was profiled with the i1 Profiler and the Eye-One Pro. Not sure what this means. Either the retina really is actually more red or the Cinema display and the 2010 MBP display calibrate similarly.

Very frustrated.

 

[millerrh]

Hey guys, check this out. I took a photo of my rMBP next to my 2010 MPB from behind. Never noticed this before, but the apple logo is glowing pink on my rMPB. No wonder I have a reddish cast, the backlight is straight up pink/magenta.

I have the LG monitor and was contemplating replacing just due to the potential IR issues, but I think this is going to make this thing go back sooner.

Maybe I'll just reformat my hard drive and live with my 2010 until rev 2 of the rMBP gets released and QC is a little better.

 

[iLukeJoseph]

Wow that is odd that your apple logo is pinkish. This whole time I was thinking bad meter, but now that you showed that I am not sure.

BUT what still strikes me as a bit odd, is that even if your monitor was putting of a "red-cast", your i1 pro should be able to detect that. And your calibrations should fail (or at least show a high red reading), but according to both BasicColor and Coloreyes your profiles are good.

The only way I see this as not being a hardware issue (rMBP hardware), is if your meters are out of spec. Let's say Red is out of spec on the meter, and it is reading it to low (which is generally the first color when a colorimeter drifts out of spec, I am not 100% sure if that applies to the i1Pro), the software will read as red is to low and start bumping it up. And you will end up with over saturation of red in the grayscale.

I don't know if I would get rid of the i1pro honestly. But also I wouldn't personally use it till I got it certified, because if it was 2nd hand, I would have no idea how the previous owner took care of it. I couldn't be 100% sure that the meter was accurate.

BUT, for what your doing, I would recommend an i1 Display Pro. They are freaking awesome. Very fast, accurate, reads much further into black, compare my chart's contrast ratio to yours. And for me the best part about them is that they are sealed. So they stay accurate for much longer. I have purchased brand new i1d2's that were so far out of spec it was insane. The problem with colorimeters with their sensors exposed, is even sitting on the self at the store they are drifting out of spec.

If you do end up replacing your laptop and still are having issues with calibrating let me know. We will figure it out.

 

[millerrh]

Well here's what I decided to do:

Return my rMBP...not worth it to try to keep the LG screen that glows pink.

Reading about all these issues with these panels has me wanting to wait until next gen.

I just placed an order for an i1 Display Pro and a Samsung SSD and I'm going to make my 2010 model as good as it can be in both speed and color. I think adding an SSD will make it feel like a new computer. I may still run out of RAM and won't have the fastest thing around, but its served me OK since 2010 and my photography equipment hasn't changed all that much since then.

I'll see what happens next time around. These screens just have too many quality issues for me right now.

 

[millerrh]

Got an update for you guys. I returned my rMBP and bought a new i1 Display Pro. Used it on my 2010 MBP. When I compare profiles made by the i1 Display Pro to the ones made by the old eye one Pro there is a huge difference.

I see now that the i1 Profiler software lets me choose backlighting technology. I don't think this was an option with the eye one Pro hardware. I assume for the 27" LED Apple Cinema display I should choose White LED as the backlight. Is this right?

This setting is much warmer than the old eye one Pro profile. Now I am wondering if the eye one Pro was incorrectly calibrating my screens and the rMBP might not have had as bad of a color cast as I originally thought. (if you remember I thought the rMBP had a red color cast and had a pink apple logo). But my new i1 Display Pro profile is also a bit redder and warmer than what the eye one was telling me.

Strange how two monitor calibration devices run by the same software can yield such different results. I think I should be trusting the newer i1 Display Pro. I found the calibration certificate for the eye one Pro - 2007 - so it is quite a few years old for calibration. Plus the new i1 Display Pro is newer and designed with newer LED backlighting in mind. I'm just amazed at how different things look when I thought the whole time I was looking at a calibrated display.

 

[evilzardoz]

Channeling my inner thread necromancer here...

I've bought the 2013 system and have noticed on the newer LG displays, the pink colour cast is less pronounced. I also had my old 2012 display replaced with a Samsung display which is very warm, a lot less pink and is far more preferable to my eyes. Blues are not as oversaturated/purple and it's probably the best display I've ever seen in a 2012 system.

Unfortunately I'm told the display modules are not compatible, so swapping the 2012 and 2013 displays is not something I can do. My 2013 system also has a Samsung panel (my first 2013 had the LG) - oddly enough, when I calibrated the LG 2013 panel I had and the Samsung in the 2012, I ended up with a very similar result with a Spyder 3 Elite... in fact, the pink returned to the display with the calibration profile enabled. A side note - the Samsung displays have a better viewing angle in terms of backlight light falloff.

This leads me to believe that if I get a calibration device that's able to read the display properly, then I'll end up with the right result. The new 2013 iMacs should be an interesting test as these things seem to come out of the factory with a whitepoint that's almost spot on compared to the earlier models (again, pink!).

Has anyone had any similar experiences?