Apple iMac & Studio Displays - Display Calibration

[digitaldog]

I have been calibrating a 2017 iMac with an i1Display Pro and have been unable to see any difference between the before and after views, even after a month or more. This applies to color and brightness.

So your point is?

[David Eichler]

So your point is?

That the monitor seems very stable.

[digitaldog]

That the monitor seems very stable.

Seems; now let's see the measurement trending over a year. In dE.

[routlaw]

Curious, what would cause any given LED display to drift or loose its stability over time. I doubt the actual panels themselves would change over time, certainly voltage swings might cause drift but surely all of these displays would have internal voltage regulation, that leaves the LED backlight source itself which given their very long lifespan (normally) would indicate to me the LED’s would not drift much if at all in their lifespan. Certainly nothing like the older technology fluorescent backlight which absolutely do drift due to the phosphors burning out over time.

I don’t have a dog in this race, but the discussion does beg the question what would cause an LED display to drift?

[digitaldog]

Curious, what would cause any given LED display to drift or loose its stability over time. I doubt the actual panels themselves would change over time, certainly voltage swings might cause drift but surely all of these displays would have internal voltage regulation, that leaves the LED backlight source itself which given their very long lifespan (normally) would indicate to me the LED’s would not drift much if at all in their lifespan. Certainly nothing like the older technology fluorescent backlight which absolutely do drift due to the phosphors burning out over time.

Why not ask (as two mere examples) Eizo and NEC exactly what components in the entire system they supply, causing them to recommend re-calibration.
Here's what I'm told in my SpectraView manual:

How often do I need to calibrate my monitor?
Answer: The answer will depend on how critical a color accuracy is necessary, how long the monitor is used per day, environmental conditions and the characteristics of each individual monitor. Some users working in highly critical color applications may calibrate their displays every couple of days. Others may only need to do it once a month. The recommended and default re-calibration period is every 2 weeks. The Validate Calibration feature can always be used to verify the display’s white point and Intensity and determine if the monitor needs to be re- calibrated. Always remember to allow the monitor sufficient time to warm up and reach optimum color performance before performing any color critical work or calibrating the display.

[routlaw]

Why not ask (as two mere examples) Eizo and NEC exactly what components in the entire system they supply, causing them to recommend re-calibration.
Here's what I'm told in my SpectraView manual:

Thanks.

I suppose, but typically the customer service people one tends to connect with at technology companies do have limited engineering knowledge. One would probably need to discuss this with the actual engineers themselves. You might be able to talk with them but its highly unlikely most of the rest of use would be given the time of day with the engineers who design these monitors.

The answer provided in your Spectraview manual strikes me as being a typical canned corporate comment to cover their butt more than something truly based on science or engineering acumen, but I could be wrong too.

[digitaldog]

I suppose, but typically the customer service people one tends to connect with at technology companies do have limited engineering knowledge. One would probably need to discuss this with the actual engineers themselves.

Go for it. Not that it matters. The manufacturers provide (expensive) tools like Colorimeters and software to recalibrate their displays, some software even has reminders to do so. IF as you assume, like David, that this process is so stable, why would anyone ever need to recalibrate a display after the first time doing so?

The answer provided in your Spectraview manual strikes me as being a typical canned corporate comment to cover their butt more than something truly based on science or engineering acumen, but I could be wrong too.

It strikes you but you as yet have no data to show us why they expect recalibration as often as every day or as little as once a month. This assumption is very similar in thinking to David's whereby he feels, without any data, that his display is "stable" because he doesn't see any difference in toggling calibration on and off. And yet, there are software products that provide trending over time, using actual colorimetric measurements and dE analysis to show users exactly the stability and difference of the display over time.
Yes, you could be wrong and the recommendations from the display manufacturers could be right. I hope you don't mind that until you come up with some actual data, I'll stick with the manufacturers recommendations, broad as they may appear to some. 

Every time I look at my right foot, I assume it's exactly 12 inches long (it isn't) and every time I use it to measure something it is consistently inaccurate. There are reasons humans use measuring devices that have a known accuracy and provide consistency in measuring something (like a display).

"I often say that when you can measure what you are speaking about, and express it in numbers, you know something about it; but when you cannot measure it, when you cannot express it in numbers, your knowledge is of a meager and unsatisfactory kind."-Lord Kelvin

[routlaw]

Not trying to pick a fight on this and as previously stated I have no dog in this race. FWIW I have also been doing digital imaging since the mid 90's and as such have always calibrated my monitors (when it was possible in those days) at somewhat regular intervals. Currently I have a Spectraview PA271W, 10 years old but still ticking which IS calibrated on a regular basis.

I merely brought up the possibility of why an LED monitor might not drift and loose its stability over a period of time. Just for the record David is not the only one who believes some monitors are stable over long periods of time, and have seemingly provided some evidence to support this such as @Czornyj who also uses an NEC monitor. It is worth noting, the tech support person I discussed this with at Eizo earlier this week also supports the idea of regular calibration.

[digitaldog]

Not trying to pick a fight on this and as previously stated I have no dog in this race.

Understood. I'm simply providing the actual facts we have thus far and ignoring the assumptions.

I merely brought up the possibility of why an LED monitor might not drift and loose its stability over a period of time.

It's a possibility of a wide range of devices we call displays but nonetheless, LED display suppliers, at least two, tell us to recalibrate. Until I see measurement data that this isn't necessary, I'll stick with the manufacturer's recommendations.

It is worth noting, the tech support person I discussed this with at Eizo earlier this week also supports the idea of regular calibration.

There you go.

[jrp]

Of course all this recalibration assumes that your calibration device drifts less than your monitor.  It’s been some time since anyone compared 10 or more devices of a particular type against a reference calibration tool. 

[routlaw]

Of course all this recalibration assumes that your calibration device drifts less than your monitor.  It’s been some time since anyone compared 10 or more devices of a particular type against a reference calibration tool.

Good point, one I haven't thought of either.

It never ends though. Over the last couple of days I have calibrated my PA271W a few times and for what ever reason the each ∆ E's do not show up on the tab, stating an N/A. Never ran into this one before. Otherwise the calibration looks good.

[digitaldog]

Of course all this recalibration assumes that your calibration device drifts less than your monitor.

The "drift" of the Colorimeter?

Good point, one I haven't thought of either. 

The dE's of any manufacturer's device is well expressed and better be invisible or you got yourself a POS.
https://www.displaycalibrations.com/x-rite_i1_measurement_solutions_info.html

[routlaw]

Just copied this from the Spectraview Manual.

Note: Several factors can influence the results of the validation process and a Delta-E values of 3-5 do not always indicate that the display needs to be recalibrated. Even very slight differences in the positioning of the sensor from when it was calibrated, and the length of time the monitor has been powered on, and the basic measurement repeatability accuracy of the sensor itself, all impact the results. Also keep in mind that the Delta-E value is comprised of the luminance and color difference. Differences in color are much more critical than a difference in relative luminance.

and further down in the text:

The calibration and adjustment process has inherent variability due to many factors such as the repeatability of color sensor measurements, and changes within the display as it is adjusted. The basic accuracy and repeatability specifications for most sensors is typically +/- 0.002 or higher for CIE xy, and +/- 2% for luminance. These variances alone can give rise to Delta-E values of up to 2 or 3.

[digitaldog]

Of course all this recalibration assumes that your calibration device drifts less than your monitor.  It’s been some time since anyone compared 10 or more devices of a particular type against a reference calibration tool.

Well here's 9 in a row to start, all measured in PatchTool (36 samples) and then a dE report from ColorThink Pro.
Basic measurement repeatability accuracy of the sensor itself:

--------------------------------------------------

dE Report

Number of Samples: 36

Delta-E Formula dE2000

Overall - (36 colors)
--------------------------------------------------
Average dE:   0.10
    Max dE:   0.25
    Min dE:   0.02
 StdDev dE:   0.07

Best 90% - (32 colors)
--------------------------------------------------
Average dE:   0.08
    Max dE:   0.22
    Min dE:   0.02
 StdDev dE:   0.05

Worst 10% - (4 colors)
--------------------------------------------------
Average dE:   0.24
    Max dE:   0.25
    Min dE:   0.23
 StdDev dE:   0.01

--------------------------------------------------
--------------------------------------------------

Invisible! I1Display Pro Plus.

[routlaw]

And once again I can't seem to obtain color tracking errors, oh well.

[digitaldog]

Search some older posts here, some users had this issue in the past, none here:

[routlaw]

Yeah this has never occurred before until yesterday. No idea what is different. Will look on this forum, but an internet search turned up noting so far.

[GWGill]

Curious, what would cause any given LED display to drift or loose its stability over time.

Having investigated this in some detail, let me tell you that LEDs drift quite markedly with temperature, and to a lesser degree over time.
Monochromatic LEDs change in wavelength, amplitude and spectral shape with temperature. Phosphor based (i.e. white) LEDs vary due to the underlying monochromatic LED changes, and may have additional variation in phosphor response to compound the effect.
Over time, LEDs have a "burn in" change at the start of their lives (roughly 100 hours), and a slower change over their lifetime. The lifetime of a LED is typically quoted as the time before it reaches 50% of its initial output. Lifetime depends on how hard they are being driven, amongst other things, but are generally long, which is a good thing.
I don't have any direct experience in regard to LCD panels, but I'd be a bit surprised if there wasn't temperature and aging effects in play there as well, give the chemical/electrical nature of liquid crystal (they certainly get slower with lower temperature), and things like physical spacing changing with temperature etc. Now all this may not add up to much in day to day and less critical usage, but it is definitely there, and something to worry about if you are going to rely on a display long term.
[ If you think that making a highly stable light source is easy, then I'm sure the calibration standards people would love to hear from you! Typical light references are very expensive, and guaranteed for a quite limited number of hours. Making stable sensors has challenges as well, but seems a bit easier overall. ]

[Frans Waterlander]

Well here's 9 in a row to start, all measured in PatchTool (36 samples) and then a dE report from ColorThink Pro.
Basic measurement repeatability accuracy of the sensor itself:

Repeatability tests tell you nothing about the aging effect of the sensor. Yes, the sensor itself also ages with time.

[digitaldog]

Repeatability tests tell you nothing about the aging effect of the sensor.

Of course it does. You're lost in color space again. Try actually reading and trying to understand the topic and test ("Listen to understand instead of listening to respond." - Barack Obama)

Well here's 9 in a row to start, all measured in PatchTool (36 samples) and then a dE report from ColorThink Pro.
Basic measurement repeatability accuracy of the sensor itself:

One who understands and has the tools simply run the same test at a later date and compares the dEs. A process you can't produce nor conceive of! 
Just like (for you) the process of photography.

I'd suggest Frans, you attempt to read and understand this, but the first sentence will very likely make your head explode:
https://www.argyllcms.com/doc/WideGamutColmters.html