Apple iMac 27" display

[Roman]

I am somewhat new to this app and forum, so please excuse me… but I would love to see If any of you use an iMac — many studios and people do either in studio or on location — so  I wanted to pick your brains for settings advice.

Over the last few years I have been using Spyder 4 Pro for my calibration.  Good results but even in my brief exposure to DisplayCAL it appears better.

Primarily I am a photographer and I retouch all of my own images… My trouble come in that, in addition to colors being accurate, I want to use the other advanced options to create smoother gradients (for retouching/dodging-burning, etc) and better shadow detail.

Watching some tutorials…

— it seems like there are some differences between calibrating to sRGB vs. Gamma 2,2
–LUT and Curve is better than Matrix, etc. but some displays get confused by LUT.
–L-star curve would help if I am using eciRGB as my workspace in Photoshop.
…etc.

So it’s a bit confusing.

What do some of you do for this iMac display; what would you recommend to me?  ALSO, which proofing should I be using?

Thank you.

[Florian Höch]

I want to use the other advanced options to create smoother gradients (for retouching/dodging-burning, etc) and better shadow detail.

None of the calibration and profiling options will typically influence that, because in a color managed workflow, the inherent image properties define how it should look, and so it comes almost purely down to how accurate the monitor profile is and how precisely the color management is done by the respective imaging application that is being used (banding and quantization artifacts are often an effect of limited bit depth processing when not using proper dithering).

it seems like there are some differences between calibrating to sRGB vs. Gamma 2,2

Yes, but the calibration tone curve only affects non-color-managed applications (although those are the majority, atleast under Windows).

LUT and Curve is better than Matrix, etc. but some displays get confused by LUT.

It’s not the display that gets confused, some applications simply do not have color management implemented properly. Non-support of LUT display profiles is an application limitation (I would even go as far as calling it a pretty severe bug), not a limitation of the display (the display itself doesn’t know or care about ICC profiles).

L-star curve would help if I am using eciRGB as my workspace in Photoshop.

Depends. The idea is that you calibrate to L-star, and then have the same tone curve for the monitor profile, so that there are no tone-curve induced quantization artifacts when going from eciRGBv2 (which also has the L-star curve) to the display profile, using limited bitdepth processing.

In practice, there are some things to keep in mind:

• Tone curve differences are not the only thing that can introduce quantization artifacts, e.g. converting between gamuts of different sizes and shapes can have the same effect.
• Graphics cards have become more sophisticated over the years, and may allow for a higher bitdepth (usually between 10 to 12 bits integer precision) videoLUT even if the frame buffer itself is only 8-bit (e.g. AMD have been doing this for quite some time), and may employ dithering to smoothen the results.
• For there to be no quantization errors when converting between tone curves, they would have to be exactly the same. Staying in the L-star/eciRGBv2 example, this means the display profile tone curve would need to follow an “ideal”, zero black, curve, that is not in line with the actual capabilities of the display (e.g. an LCD display cannot have a zero black level unless you turn off the backlight). The result is a loss of accuracy. Wether that loss is significant or not would need to be assessed from situation to situation.

Overall, I’d consider doing the processing with sufficient precision and employing dithering to be the better solution to prevent quantization artifacts, but it depends on the situation and the equipment (software, hardware) that you have available as well as your needs.

[Roman]

Thank you for all that.

I made many calibration tests and it seems that the following settings work best:

LCD (White LED)
White/Black Level Drift Comp.
Auto-Correction

CIE 1931
6500K (or 5000K) Daylight — Blackbody from what I gather is more for LUT and monitors with internal LUT
120 cd/m
Black Level – as measured
Tone Curve – sRGB
Black Out. Offset – 100%
Ambient Light… (whatever my Spyder reading)
Black Point Corr. – Auto
Calibration Speed – depends on how precise you want DisplayCAL to be

Single Curve_Matrix setting
Profile Quality – High
Extended Matrix Testchart

My shadows and gradients are much better now.

SO HERE’s a question — if in a month I want to update this calibration, fine-tune any color drift over time… do I just click “Update Calibration”? That seems to gray-out setting options… or what is the actual process to update a calibration I like?  What goes in the test chart area?

(And here’s a little side question:  I tried going away from a single curve to Curves — why, for my display, is this providing worse results at the same settings… blacks crushed and banding… I would thing individual curve adjustments, rather than a single (average) curve, would be better?)

Thank you.

[Roman]

Oh one more thing — The display has a NVIDIA driver.
What is a better option for DisplayCAL, calibration and proofing, etc?  OS X Display Driver or NVIDIA Web Driver?

[Florian Höch]

White/Black Level Drift Comp.

That does nothing but increase the measurement times unnecessarily (but shouldn’t hurt otherwise). Black level drift compensation is not needed for colorimeters generally, and white level drift compensation is only useful if the display doesn’t have stable light output (e.g. it makes no sense for LCD monitors because they have a static backlight).

Blackbody from what I gather is more for LUT and monitors with internal LUT

The difference between the daylight and blackbody targets is that they are based on slightly different spectra. Daylight is the more common target and should be used unless you have special needs.

Ambient Light… (whatever my Spyder reading)

I would strongly recommend against using ambient light level adjustment unless you know exactly what you’re doing. On Mac OS X, this setting doesn’t affect the calibration result anyway because the whole desktop is color managed.

if in a month I want to update this calibration, fine-tune any color drift over time… do I just click “Update Calibration”?

Yes.

I tried going away from a single curve to Curves — why, for my display, is this providing worse results at the same settings… blacks crushed and banding…

This depends on the application(s) you’re using. All the Apple apps (Photos, Preview) seem to do processing with limited precision, no or insufficient dithering, and have bugs when it comes to certain types of profiles. I would highly recommend against relying on Apple’s picture apps for those reasons.

[Roman]

thank you.

So in using my Spyder to evaluate and create an ambient light metering — this would be less optimal? Leave it as un-checked?

Also, for updating the current calibration… once I click “update calibration” some of the options are grayed out; nothing needs to be adjusted, changed or corrected at this point? AND what should be in the Testchart section, which chart or file type for updating the calibration?

thank you,

Roman

[Florian Höch]

So in using my Spyder to evaluate and create an ambient light metering — this would be less optimal? Leave it as un-checked?

In essence, yes.

once I click “update calibration” some of the options are grayed out; nothing needs to be adjusted, changed or corrected at this point?

Correct.

what should be in the Testchart section, which chart or file type for updating the calibration?

Leave it as-is.

[postmodernboom]

Hello and happy new year!

I have a question about updating the calibration.  So I got the alert to update my calibration… I loaded the icc profile and the settings that I had initially created through calibration… checked off “Update Calibration”.  It did its thing and it created a drastic change.  I did it twice with equally bad/drastic results — once by selecting the current color profile as my good, recently calibrated profile in System Preferences AND once by selecting the default iMac profile.

What is the proper way to do this?

If what I did was correct, what could be the issue?

Thanks so much!!!!

[Florian Höch]

Hello and happy new year!

Thank you, likewise!

So I got the alert to update my calibration…

Hmm. Where did that come from? DisplayCAL doesn’t alert about the age of the current calibration and profile, so there may be another software installed that is possibly intefering.

[postmodernboom]

Oh yes, sorry — it was my Spyder software keeping track of time.  But what about what I asked?

“I loaded the icc profile and the settings that I had initially created through calibration… checked off “Update Calibration”.  It did its thing and it created a drastic change.  I did it twice with equally bad/drastic results — once by selecting the current color profile as my good, recently calibrated profile in System Preferences AND once by selecting the default iMac profile.

What is the proper way to do this?

If what I did was correct, what could be the issue?”

Would appreciate the step-by-step to help.

Thank you.

[Florian Höch]

I’m not sure what else the Spyder software may be doing under Mac OS that may be interfering with updating the calibration using DisplayCAL. I’d suggest removing the Spyder software completely and then re-trying to update the calibration.

[postmodernboom]

Okay – I will remove it.

But what is the proper, step-by-step process to update cal?  (I asked that a few times before, just really would like to know).

Thanks.

[Florian Höch]

But what is the proper, step-by-step process to update cal?

There’s not really any special steps you need to take. You enable the “Update calibration” checkbox and click “Calibrate”. Although I would create a new calibration & profile instead of updating the existing one if it is (say) already several months old.

[Author]

Posts