Apple Explains How to Custom Calibrate the New MacBook Pro Displays

[David Abrams]

In theory, you can use any meter to measure the display, the question is, 'how accurate will that measurement be'? The MacBook Pro XDR uses a PFS Phosphor based LCD display which has spikes in the red (see attached spectral power distribution). Because of this spike, it is important to use a spectral radiometer with enough spectral bandwidth to properly categorize the display's performance. A meter with lower bandwidth may not have the resolution to accurately measure the spike. Of course, this isn't only on the XDR Display but any display that uses PFS Phosphor. If a colorimeter is properly profiled with the MacBook Pro Display XDR, it should have fairly good accuracy, but a generic profile may have larger error than desired. When profiling, or calibrating a colorimeter, the most common method is to use a matrix calibration, which measures WRGB, this takes into account the filters and gamut of the display and without the company disclosing what they used to calibrate that profile it is difficult to be certain of the accuracy of the profile; of course, you could test it, but if you have a spectral radiometer why waste the time!

While not on the Pro Display XDR, we have had some colleagues test PFS Phosphor displays and the error between spectral radiometers that may occur. They found that the difference between a 5nm spectral bandwidth meter and a 2nm spectral bandwidth meter can have an error (difference) of +/- .006 on the red y. On the sample MacBook Pro XDR display, the error we measured out-of-the-box was approximately .004 on the y-- in this case using an improper meter could add more error than out-of-the-box.

Many meters don't require Windows, but the software we happened to be using to plot and chart the data is Windows based. We could have easily used Patterns to put up the white, measure with the meter, enter the Fine-Tune, and called it a day, but it wouldn't easily allow for a validation of the result and the charts we were able to provide on the thread.

 

[Klickst]

But, to be honest, who can afford a spectroradiometer? I'm speaking now for the use of the MBP as a photographer, not video production or anything like that. Even as a professional photographer, that's an expense that's out of proportion.

 

[Klickst]

I found a way to do with X-Rite device. Woraking on a video guide. Will share here when it is done

Great! ?

 

[James211]

But, to be honest, who can afford a spectroradiometer? I'm speaking now for the use of the MBP as a photographer, not video production or anything like that. Even as a professional photographer, that's an expense that's out of proportion.

@David Abrams correct me if I’m wrong, but my assumption here is manufacturers of spectrophotometers will design a unit that is specific to PFS displays. Ages ago NEC made a line of monitors that came with its own version spectrophotometer, I’m guessing for a similar reason.

 

[David Abrams]

Yes, there are many options for calibration without the need to purchase a spectral radiometer. Our business is primarily a video system calibration business that supports end users, high-end home theater, and production/post-production clientele. For those that either don't have the time, the expertise, or don't want to spend on the equipment, they'll hire us to perform the calibration.

Of course, there are companies that will offer a colorimeter with custom calibration matrices built-in specifically for their monitors. As you mentioned, NEC is one of those companies, but so is LaCie and HP. These meters have custom matrices loaded in specifically for the monitors they are designed to calibrate.

 

[badlydrawnboy]

Ok, I'm totally lost. I know enough about this stuff to be dangerous, but I'm clearly over my head in this thread.

I use a Calibrite ColorChecker iDisplay Pro and DisplayCal software to calibrate my iMac Pro 5k retina display. Can I use this same combination to calibrate my 14" Macbook Pro display? If so, do I need to do anything differently?

Thanks.

 

[badlydrawnboy]

I found a way to do with X-Rite device. Woraking on a video guide. Will share here when it is done

Can't wait!

 

[Matck06]

Can't wait!

ditto I can't wait I ordered the i1 display pro

 

[luxnova]

Looking forward to seeing that. Hoping it does not require Windows as an intermediary.

It does not. Video going live on a my YouTube channel in a few hours. I'll post the link when it is ready.

 

[luxnova]

ditto I can't wait I ordered the i1 display pro

Yes but I have caveat about it. I'll share more of my thoughts in the video.

 

[Matck06]

Yes but I have caveat about it. I'll share more of my thoughts in the video.

Ah ok, the video comes out when I can't wait to see it.

 

[luxnova]

Here's a link to the video on how to set up a custom reference mode, fine tune calibration and then do a full calibration along with reasons why you might not want to do that yet

it is longer so use the timestamp to skep to calibration section

 

[badlydrawnboy]

Here's a link to the video on how to set up a custom reference mode, fine tune calibration and then do a full calibration along with reasons why you might not want to do that yet

it is longer so use the timestamp to skep to calibration section

Thanks so much! Does this cover using an external device like the iDisplay Pro?

 

[luxnova]

Thanks so much! Does this cover using an external device like the iDisplay Pro?

Yes for both calibration fine-tuning and also for full custom calibration to create an icc profile, but for many reasons as I showed and explained in the video toward the end, we might want to hold off on that part. Just doing the calibration fine-tune will do wonders. These displays are extremely close to spot on from the factory!

 

[paalb]

Here's a link to the video on how to set up a custom reference mode, fine tune calibration and then do a full calibration along with reasons why you might not want to do that yet

it is longer so use the timestamp to skep to calibration section

Thanks! I have not got my new mac yet, but now I am prepared. By the way I was already subscribed and left a thumbs up. I recommend that all following this thread do the same.

 

[luxnova]

In theory, you can use any meter to measure the display, the question is, 'how accurate will that measurement be'? The MacBook Pro XDR uses a PFS Phosphor based LCD display which has spikes in the red (see attached spectral power distribution). Because of this spike, it is important to use a spectral radiometer with enough spectral bandwidth to properly categorize the display's performance. A meter with lower bandwidth may not have the resolution to accurately measure the spike. Of course, this isn't only on the XDR Display but any display that uses PFS Phosphor. If a colorimeter is properly profiled with the MacBook Pro Display XDR, it should have fairly good accuracy, but a generic profile may have larger error than desired. When profiling, or calibrating a colorimeter, the most common method is to use a matrix calibration, which measures WRGB, this takes into account the filters and gamut of the display and without the company disclosing what they used to calibrate that profile it is difficult to be certain of the accuracy of the profile; of course, you could test it, but if you have a spectral radiometer why waste the time!

While not on the Pro Display XDR, we have had some colleagues test PFS Phosphor displays and the error between spectral radiometers that may occur. They found that the difference between a 5nm spectral bandwidth meter and a 2nm spectral bandwidth meter can have an error (difference) of +/- .006 on the red y. On the sample MacBook Pro XDR display, the error we measured out-of-the-box was approximately .004 on the y-- in this case using an improper meter could add more error than out-of-the-box.

Many meters don't require Windows, but the software we happened to be using to plot and chart the data is Windows based. We could have easily used Patterns to put up the white, measure with the meter, enter the Fine-Tune, and called it a day, but it wouldn't easily allow for a validation of the result and the charts we were able to provide on the thread.

David great information, thank you! I appreciate this and I am sure many on this thread do as well. I wanted to add based on my extensive testing as of now, the matrix data used in the X-Rite devices are not adequate to fully profile these displays and get the best result at this time. The gamut that result from these displays conforms to Display P3 but are much smaller than that of a MacBook Pro, or Air that comes before it. The other issue with using X-Rite colorimeter right now is that even though the display backlight is PFS Phosphor like you and many other have tested. That yield a terrible calibration result and the work around for now is to use White LED, which is the necessary the correct backlight source but it gives a much better and accurate result with the whitepoint and ∆e outcome.

If anyone uses any X-Rite color spectrophotometer, which are great devices, but they are better for prints than display. The measurement bandwidth is about 10 nm which does not work well for displays in general but is much worse for these mini LED display. Hopefully they will work on a update to their backlight matrix and driver soon.

 

[David Abrams]

David great information, thank you! I appreciate this and I am sure many on this thread do as well. I wanted to add based on my extensive testing as of now, the matrix data used in the X-Rite devices are not adequate to fully profile these displays and get the best result at this time. The gamut that result from these displays conforms to Display P3 but are much smaller than that of a MacBook Pro, or Air that comes before it. The other issue with using X-Rite colorimeter right now is that even though the display backlight is PFS Phosphor like you and many other have tested. That yield a terrible calibration result and the work around for now is to use White LED, which is the necessary the correct backlight source but it gives a much better and accurate result with the whitepoint and ∆e outcome.

If anyone uses any X-Rite color spectrophotometer, which are great devices, but they are better for prints than display. The measurement bandwidth is about 10 nm which does not work well for displays in general but is much worse for these mini LED display. Hopefully they will work on a update to their backlight matrix and driver soon.

Hi Luxnova,

I don't understand, "The display conforms to Display P3 but are much smaller than that of a MacBook Pro, or Air that comes before it." Our measurements illustrate the display is capable of hitting P3, the addition of the luminance component for HDR puts the display's color volume way higher than that of the previous generation of MacBook Pro or MacBook Air.

Perhaps I'm not quite understanding what you are referring to?

 

[luxnova]

Hi Luxnova,

I don't understand exactly what you mean by, "The display conforms to Display P3 but are much smaller than that of a MacBook Pro, or Air that comes before it." Our measurements illustrate that the display is capable of hitting P3 without issue, the additional of the luminance component for HDR would put the display's overall color volume way higher than that of the previous generation of MacBook Pro or MacBook Air.

Perhaps I'm not quite understanding what you are referring to?

Let me clarify, I am not disputing your results at all especially with the use of a proper calibration instrument. I was referring to any one who uses X-Rite / Calibrite solutions off the market right now. The icc profile generated has the shape of a Display P3 color gamut but the coverage of the color gamut is smaller. This is even smaller than other Apple Laptop display of the previous generation and it is because these devices rely on a matrix data that needs to be update before these X-Rite device can reach it full calibration potential. For any one who does not have a color radiometer and want to use X-Rite devices, it is now a waiting game until the software and driver catches up.

Attached image, outlined is Display P3, solid color gamut is the icc profile generated using X-Rite i1Display Pro Plus with i1Profiler, and even when I used a spectrophotometer i1Pro 3 Plus the gamut is about the same and may be even a bit smaller.

 

[David Abrams]

Hi Luxnova,

I understand where you're referring to now, thank you. It appears that the primaries that were measured for the color profile aren't quite hitting the desired P3 targets, which is causing that differentiation. It could be a factor of the meter error, or it could be a factor of the pattern measured. ?

My results were after the 'Fine-Tune' but I'll see if we can create a custom ICC and how that correlates to your results. I've been tied up with some client projects that have taken up the past weekend and will extend into this weekend, but I'll try to carve out some time as this is quite exciting!

Best,

David

 

[luxnova]

Hi Luxnova,

I understand where you're referring to now, thank you. It appears that the primaries that were measured for the color profile aren't quite hitting the desired P3 targets, which is causing that differentiation. It could be a factor of the meter error, or it could be a factor of the pattern measured. ?

My results were after the 'Fine-Tune' but I'll see if we can create a custom ICC and how that correlates to your results. I've been tied up with some client projects that have taken up the past weekend and will extend into this weekend, but I'll try to carve out some time as this is quite exciting!

Best,

David

Hi David,

I appreciate that. Thank you! And yes this is fun and exciting!

Art

 

[Matck06]

@luxnova Thanks for the tutorial I used displaycal instead I post my results below

 

[luxnova]

@luxnova Thanks for the tutorial I used displaycal instead I post my results below

@luxnova Thanks for the tutorial I used displaycal instead I post my results below

Thank you for sharing, yes DisplayCal is one solution right now and I am fully aware of that update from DisplayCal.

 

[sydlow]

Here's a link to the video on how to set up a custom reference mode, fine tune calibration and then do a full calibration along with reasons why you might not want to do that yet...

Thanks for taking the time to publish this.
BTW, the Delta-E on my 16" MBP came in at 0.8 with very low standard deviation.

 

[James211]

Thanks for the video, well done! One issue I’m having, even after an extensive warmup, when I set SDR to my chosen 120cdm, the actual measurement comes in at around 110cdm. I tried 80 as well and that came in at 75. Any thoughts?

 

[luxnova]

Thanks for the video, well done! One issue I’m having, even after an extensive warmup, when I set SDR to my chosen 120cdm, the actual measurement comes in at around 110cdm. I tried 80 as well and that came in at 75. Any thoughts?

That is about the standard deviation. This is why we do the fine-tune.