- This topic has 33 replies, 5 voices, and was last updated 5 years, 1 month ago by
.
-
Posts
-
Not expecting, just hoping. I’m going to Porto next week with a group of photographers. I can’t take the external monitor with me, and the profiles produced by DisplayCAL on the MBP15 work fine so long as it is attached. If it is not, the Softproof profile does not work even when auto brightness, True Tone and Night Shift all are turned off, the way they were when the profiles were produced. The Softproof profile is critical for me. It gives me a rough idea how a photo might print. On site I can retake the picture. Back home (Switzerland) there’s not much I can do. The other profiles do work, even when the external monitor is disconnected, providing auto brightness remains off, and there’s no compelling reason to turn it on.
After writing the above paragraph, I decided to experiment some more before sending it. On the MBP15, without the external monitor, I tried connecting the i1D3 to different USB ports on the LMP USB-C mini Dock and even to an Apple USB-C to USB adapter. In all cases, I was unable to produce a Softproof profile that worked.
Finally, I decided to see what effect all these profiles had on the photo apps that I use, Apple’s iPhoto and Affinity Photo. Regarding the profiles I sent you this afternoon, the ones produced on the MBP15 while it was connected to the external display all worked as long as one of those apps was running, not only in them, but in all running apps. As soon as neither was running, the situation returned to “normal,” i. e., the Softproof profile no longer worked, and turning on auto brightness caused the others to fail as well.
So, I have a situation that I can live with for now, but I wonder how I will be able to generate such profiles without the external monitor — or whatever is attached to the Akitio Thunder2 Dock that is helping. I am planning to sell the MBP17, and the Thunderbolt Display is not ideal for photography. Is there a way of “updating” the “fine as-is” profiles without the external monitor so that they would still work, at least, in this limited fashion? I would hope, however, that the real cause of not being able to profile just the MBP15 can be found.
When time permits, I will see whether I have any better luck with bacICColor display. In the meantime, if you have any more ideas about getting DisplayCAL to work, I would appreciate them — it really is a fine product. Would you give the i1D3 a clean bill of health now, or do you suspect some subtle interaction between a failing a colorimeter and some sloppiness on Apple’s part?
Regards
Calibrite Display Pro HL on Amazon
Disclosure: As an Amazon Associate I earn from qualifying purchases.You can try a Linux Live CD (copied to USB bootable), download argyllcms and displaycal on it, run displaycal and create profile, then copy back to your mac (usb, mail, …save it the way you like most).
This way you can avoid if OSX is doing something wrong while showing ArgllCMS patches on screen, also it is unlikely that such live cd would have support for auto dimming or some propietary energy saving tweaks that may be causing your issues.Then try to validate with DisplayCAL (back in OSX). If when Displaycal asks argyllcms to show verification patches it goes wrong again because some OSX issue, try to validate with Basiccolor display (you said you own it, but it is possible that ther is a whitepoint droft caused by bassicolor software lacking of proper spectral corection for that display)
PS: Make sure you choose single curve +matrix profile, OSX desktop color management is not ready to use more compelx profiles.
PS2: P3 screen mackbooks should use Apple’s P3 CCSS corections in community correction database or Xrite’s “Panasonic VVX17P051J00” in DisplayCAL i1d3 bundle.
I am pretty sure that I have found the cause of the problem. Both the MBP17 and the MBP15 have both on-board and external graphic processors. In System Preferences > Energy Saver “Automatic graphics switching” is usually checked, allowing the machines to switch between the two for better battery life. This switching might have been the cause of the random behavior that Florian observed. When an external monitor is attached, the external graphics processor is always used, and I imagine that programs like iPhoto and Affinity Photo also use it. That explains why (a) the profiles for the MBP15 are good when an external monitor is attached, and not when not (i. e., because the MBP15 wasn’t switching back and forth between graphics processors during calibration and profiling), and why (b) when no external monitor is connected some of the profiles generated when one was do not work (i. e., because the machine is using the built-in processor) unless iPhoto or Affinity Photo is running (in which case the external graphics processor is running).
So, the solution is to uncheck “Automatic graphics switching” before calibrating and profiling, and if later a profile doesn’t seem to work, uncheck it. The safest would be, leave it unchecked and check it only when battery life is important.
I suppose DisplayCAL could sniff this option and ask permission to change it, so that switching back and forth does not interfere with calibration and profiling. I don’t know whether it is possible to stop the switching so that only the internal graphics processor runs. Furthermore, I don’t know whether it would be possible to switch profiles when the processors are switched.
The Softproof profile produced using Apple’s P3 CCSS corrections seems a bit brighter than the one produced without it.
Regards,
The Softproof profile produced using Apple’s P3 CCSS corrections seems a bit brighter than the one produced without it.
Without one of those 2 corrections (comunity or bundle) you are using “colorimeter observer”. You are taking as granted that it matches exactly CIE 1931 2º observer. Although i1d3 are very close to that, they are not exactly equals this observer.
This is the reason Xrite saves in firmware (supposed) actual observer response of that (batch?) colorimeter. With actual observer (firmware) and spectral sample (CCSS files) you can measure to a high degree of precision. And this is just software! Upgrades for future displays like yours of better graphic arts backlight do not need Xrite intervention at all.This stuff (“software” corrections) and inter instrument agreement if what give i1d3 family its fame.
I appreciate your explanation. I’m sure that such corrections improve the accuracy of the results, for otherwise there would be no need for the software to make them and contributors to provide them. My observation was not intended as an argument against using them.
Indeed, I was rather arguing with myself. “If you’re so sure you’ve found the problem, why do the two profiles that were made using the external graphic processor nevertheless produce slightly different results?” I would attribute that to differences in the lighting conditions under which the two were generated. I’m leaving for a trip to Porto with a group of photographers on Tuesday. I should be packing things, but instead I’m devoting a lot of time — and, if not for you and Florian, it would be even more — to color profiling a new laptop. I was on the verge of giving up and just taking the old MBP17. Imagine how p****d off I would have been in Porto when it first booted up and manifested the same problem. So my observation about the differences in the profiles was a kind of reality check. Are the two graphic processors the whole explanation, or is there something else going on? The “something else,” I’m convinced, are the different lighting conditions.
Is there a way of “updating” the “fine as-is” profiles without the external monitor so that they would still work
The problem is not the profiles, they are fine.
Hi,
I’m having similiar issue with MBP (2018) + xrite i1 Display 2 and seems like it’s because of the GPU switching, as noted above. If you calibrate when “Automatic graphics switching” is off (or connected to an external monitor) and then change to use Intel graphics colors go all crazy. For some reason GPU made profiles don’t work with Intel graphics, but the other way around yes.
Any ideas how to fix this?
The profiles that “go crazy” seem to be the ones for which the brightness is specified. The others seem only to turn some shades of gray to cyan. I have found no solution other than to use the builtin Apple profile whenever possible. When I need a profile produced by DisplayCAL using the external graphics processor, either I turn that processor on manually, or I trust that the software that needs it will do that.
BasICColor, for whose basICColor display software I have a license, has informed me about what to do to get their software to work, but I have not been able to figure out the implications for DisplayCAL and have not had the time to try out their recommendations. They say the say that the root of the problem are the changes to the way High Sierra handles brightness. I don’t know whether I am at liberty to post the information here for others with more knowledge in these matters .
Oh please share your findings, after all we’re here to try to solve these issues and get our dear viewing devices to do the job, right. I also have a license of basICColor Display.
I guess one way would be somehow momentarily disable the Radeon GPU, since Intel graphics made profiles work ok in both cases.
Yes, I will definitely share my findings; however, up to now I have not had the time to try the recommendations that I received from BasICColor.
Regarding your contemplating momentarily disabling the Radeon GPU, it is my understanding that DisplayCAL begins running on the onboard GPU, but during the calibration and profiling something causes the external GPU to take over. So I guess if one could stop that, at least nothing would interfere with calibration and profiling. There’s an app, gfxCardStatus, that might support that, “might” because, although I have the app and see options which look as if they allow one to use only the one or other GPU or to switch automatically, I have not had the time to experiment.
Please keep us informed should you decide to go that route.
Yup, using integrated GPU only seems to work! Use gSwitch (updated version of gfxCardStatus) and you should be fine.
https://codyschrank.github.io/gSwitch/Thanks for doing the “grunt work.” Let me get it straight. So, you switched to using just the integrated GPU, calibrated and profiled with DisplayCAL, then switched to the external GPU — or to automatic switching and used some software that uses the external GPU –, and the profiles were still OK?
Yes, that was the procedure and so far everything seems to work ok, with Intel and Radeon GPU.
Thou, I must say the other day I was able to calibrate on Radeon GPU a profile that worked also on Intel…so there’ definitely something random there. Also, the bug I used to experience was “cyan in the dark tones”. Anyways, try with gSwitch and see how it goes.
“Cyan in the dark tones” is what I was experiencing when I switched to the preset Softproof profile while the builtin GPU was running. The other preset profiles did not do that. I chalked it up to the fact that Softproof specifies the brightness.
But glad to hear that it’s possible to generate profiles using only the onboard GPU that then work with the external GPU as well. Thanks.
I just tried to reproduce what I understood you to have done. Here’s what I did:
- Set gSwitch to “Integrated Only”
- In DisplayCAL
- in Settings selected “Softproof (5800K, 160 cd/m², L*)
- in Corrections selected “Spectral: LCD RG Phosphor IPS (Apple 15″ MacBook Pro with Thunderbolt 3 and Touch ID (Mid 2018) (ColorMunki)”
- Clicked “Calibrate & profile”
- Used MacBook Pro’s Brightness slider (System Preferences > Displays) to set 160 cd/m².
Already in the last phase of Calibrate & profile cyan appeared in the dark tones. The profile was unusable, regardless of the GPU selected in gSwitch.
I tried correction that I found for the 2018 MBP, but the results were the same.
Other than what I have described above, I left all other settings in DisplayCAL in their default states. What did you do differently than I?
