Black Point “Calibration”

[Vincent]

Step 12  Blue is now at 50 (on the OSD)
Red and Green still @ 100
White x 0.3472 y 0.4992 5173K  337 Cd/m2
Black x 0.3118 y 0.3585 6406K 1.61 Cd/m2

That’s a HUGE drop to get close to D50… but maybe that’s all such display can offer.

See? I am very close to D50 chromaticities.
What should be the next step up?
Keep in mind, I can’t work at 337 Cd/m2.

So should I bring Contrast down, from 75 to…?

Good stuff you are making progress….

Brightness 75
Contrast 50 
Red 100 Green 76 Blue 59

That Contrast control setting is unnecessarily sacrificing contrast.

As I mentioned a couple posts ago, I checked the users guide for your display and noted the default setting appears to be 75. And it so happens that on the Dells I am using the default (native)  contrast control 75. I have checked this with a test pattern.

White Point : x = 0.3460  y = 0.3524  / 4968K / 92 Cd/m2
Black Point   : x = 0.2960  y = 0.2834 / 8476K / 0.6 Cd/m2

And as you can see your effective contrast is

92/0.6 = 153:1

The panel should yield ~ 1000:1

That’s sacrificing almost 3 stops of dynamic range!OSD Brightness down till it meets your needs.

On that setup:
Brightness 75 /  Contrast 50 /  Red 100 Green 76 Blue 59, which seems close to what it should be…maybe that’s all such display can offer.

MAYBE check how constrast ratio varies aiming to cooler blue like D55, if having geen and blue at higher values at that DDC/CI tool does not show that wild drop.
On older ultrasharps with native white in cyan-blue 6900K I could get >700:1 D50 (UC OFF) but Roger showed us a native WP over 10000 or so… so maybe it’s all such panel can offer.
If at D58/D55 CR is still so poor, that RGB gains seems to behave “normal” and what we see is panel limitations. Nothing can be done.

Regarding final setup, since “Brightness 75 /  Contrast 50 /  Red 100 Green 76 Blue 59”, D50 @100nit “meet your needs”, remember to do not use  blackpoint compensation in profile type, so Photoshop can do its black compensation stuff from sRGB 0nit to 0.6nit relative to 100nit white, BUT also it can simulate black in in a realistic way when you softproof, since it seems your primary goal, prepress print.
Leaving profile type (not calibration) with BPC will store RGB 0 = 0nit so when you softproof will black in simulation photoshop will believe that it has to lift blacks a lot till L3-10, white if it takes account of your actual black level it only would need to lift simulated black a little or maybe none at all.
macOS desktop does not like this kind of settimng on profiles, it will crush all out of display gamut blacks.

[Roger Breton]

Yes, I’m interested in “Print”.  Not video nor the web. And I don’t need to bust my eyes with blinding Luminance when I am on the web at midnight, looking at — ahem! — pretty images 🙂

[Roger Breton]

So I am off the hook now?

[Roger Breton]

One more thing!!

This morning, at 6:30, when I “opened” my PC, which was on “Energy saver” all night (I never shut my PC down), I had the “pleasant surprise” that the Brightness setting, SOMEHOW, reverted to 100 and Contrast to 75? Needless to say, I was welcome with 219 Cd/m2 of “photon torpedoes”, blasting @4884K…

FORTUNATELY, restoring Brightness to 75 and Contrast to 50 gave me everything back. Go figure…

• This reply was modified 9 months, 3 weeks ago by Roger Breton.

[Wire]

FORTUNATELY, restoring Brightness to 75 and Contrast to 50 gave me everything back. Go figure…

As mentioned before: On Dell the Brightness control adjusts the backlight intensity, while the Contrast control >75 expands mid tone contrast (by clipping white) or <75 reduces overall contrast (by clamping white) using the LCD aperture, not the backlight. These controls are completely independent. The proper control for setting luminance is the Brightness control.

So why do you choose to move the contrast setting away from default of 75? What purpose does this serve?

The Brightness (backlight) control should have more than enough range to hit your luminance target of 100 cdm2.

Moving the contrast control away from default  moves panel response away from native, which reduces dynamic range while offering no benefits.

By leaving Contrast at default 75, you get as much dynamic range as possible. Print black simulation can occur in Ps using soft proof.

[Wire]

…blasting @4884K…

Also the Contrast control likely slightly disturbs white, so that’s another reason you want to avoid it.

[Wire]

Vincent wrote:

Photoshop can do its black compensation stuff from sRGB 0nit to 0.6nit relative to 100nit white

Vincent, why do you say sRGB is 0nit?

The sRGB transfer function (which technically is an OETF) has a range limit due to its linear segment at bottom of response. This leads to a dynamic range of roughly 500:1, corresponding to a 100nit display with black level set for an office environment.

So where do you get 0nit from?

[Vincent]

Vincent wrote:

Photoshop can do its black compensation stuff from sRGB 0nit to 0.6nit relative to 100nit white

Vincent, why do you say sRGB is 0nit?

The sRGB transfer function (which technically is an OETF) has a range limit due to its linear segment at bottom of response. This leads to a dynamic range of roughly 500:1, corresponding to a 100nit display with black level set for an office environment.

So where do you get 0nit from?

All these idealized colorspaces have RGB 0 -> L* 0. A linear function Y= A*X gives you 0 at 0.
Just dump TRC curves from ref/sRGB.icm

Data_Type: curv
Count: 1024
0 0
1 5
2 10
3 15
…
1018 64809
1019 64954
1020 65099
1021 65244
1022 65389
1023 65535

So if your printer+paper combination stored in an ICC file says L 4 for black and you color manage for softproofing an image with an embeded profile RGB 0 = L* 0, and shown in a display with an ICC profile it will lift display RGB black/grey depending on display TRC.

Usually this is a non issue at 1000:1 or even 600:1 displays, it won’t vary too much regardless of L* value for RGB 0 stored in display profile.
But if you take a look at your actual display black at 150:1…. it matters.

That’s the reason all HW calibration software from Eizo ro Nec, that you say you have been using all thse years although you may have not noticed this feature, by default (for example Eizo CN7)  creates profiles with a TRC tht store RGB 0 = L* 0.
But if you truly want to simulate 250:1 on display, so there is no huge black/darks difference between color managed canvas and UI buttons, it maybe advised that resulting display calibration stores actual black level, so you configure that calibration traget to create a display ICC profile this way.
It is not enabled by defaukt bc some “simple” color management engines like desktop in macOS , not able to deal with out of gamut colors in a perceptual way (= black point compensation) and they clip. AFAIK Firefox color management does the same too although I’ve not checked since a lot of versions ago. They are “simple” bc they are meant to be fast for their desinged use.

[Vincent]

So I am off the hook now?

Brightness 75
Contrast 50
Red 100 Green 76 Blue 59

White Point : x = 0.3460  y = 0.3524  / 4968K / 92 Cd/m2
Black Point   : x = 0.2960  y = 0.2834 / 8476K / 0.6 Cd/m2

On a superficial check it seems that now, with your current configuration, you got all that display can do.

But if we take account for your previous test yesterday it show no huge brightness drop from native 350->337

Step 12 Blue is now at 50 (on the OSD)
Red and Green still @ 100
White x 0.3472 y 0.4992 5173K 337 Cd/m2
Black x 0.3118 y 0.3585 6406K 1.61 Cd/m2

but such small drop make me question about native white contrast ratio. It is advertised as 1000:1, a comon IPS display.
Your RGB gain brightness drop from RGB 100 to R=100%, G=? and B=50% does not imply that.
Where is it? Where did static CR go?

So it makes me question if
-you have some video/legal level – full level issue on your computer (display works at 0-255, but GPU sends 16-235)
-or if your Minolta CS-200 is NOT working properly in dark readings.
-or if RGB gains behave wildly on some combination
(not exclusive of each other)

Don’t all you see it too?

brightness control 100% with that near D50 RGB gain configuration, it goes from ~350 nit nominal to 337 nit, just by RGB gains. A totally normal and expected drop for Native to D50… but contrast?

Where did it go? Did you have ~0.3nit at Brightes=100% & RGB gains =100% so it behaves as advertised? Does your CS200 measure a close value too?
-If your measurement device reads 0.5-0.3nit … then RGB gains are behaving wildly on black point as you move to you target whitepoint. It may be worth checking when it goes wild.
-If your measurement device reads abour 3 nit : it is not measuring properly blacks or you have some issue in your computer (typical one seen here is 16-235 vs 0-255) or some other issue to be found.

Anyway that Dells is like a common 43″ TV without TV tunner, not really meant to be pushed to these warm whites, although you and push it to this zone at 337/1.61 expense, but check the suggested issues too because device behavior seems wrong.

• This reply was modified 9 months, 3 weeks ago by Vincent.
• This reply was modified 9 months, 3 weeks ago by Vincent.

[DaniJ]

I also find the very high black readings suspicious. Could you attach a spectrum of the native black reading and the warmer one?

[Roger Breton]

For now, all measurements are in integrated XYZ tristimulus form. The CS-200 does take spectral readings, internally, but only makes the integrated values available.

Keep in mind that, Black measurements take 60 seconds to complete. Whereas White measurements take not even 1 second. The instrument adjusts integration time automatically.

At 0.60 Cd/m2, my last reading, what do you find “suspicious”? Not low enough?

[Roger Breton]

Vincent wrote :

Native white contrast ratio. It is advertised as 1000:1, a comon IPS display.

After monitor reset, full Brightness / Contrast / Gains, I get :
White 349 cd/m2
Black 1.61 cd/m2

CR = 349 / 1.61 = 216

Why are you so “hung up” on contrast ratio?

[Vincent]

Vincent wrote :

Native white contrast ratio. It is advertised as 1000:1, a comon IPS display.

After monitor reset, full Brightness / Contrast / Gains, I get :
White 349 cd/m2
Black 1.61 cd/m2

CR = 349 / 1.61 = 216

Why are you so “hung up” on contrast ratio?

Because it is showing a defect somewhere. Maybe on your CR200, or in RGB gains behavior of that dell (past some value) or a misconfiguration in GPU on maybe a defective monitor.
The reason is written above. It only falls from 350bit @100% brightness to 337nit… but black level is ~1nit.
Something is fishy, hence the importance of measuring black level at 100% brightness RGB gains 100%, etc.

You did it, but i did not read it before till now, msg lost in all these crossed messages, my fault:

For my curiosity, I decided to Reset the monitor…
See attached OSD screen capture :
– Brightness and Contrast = 75% each
– Gains @ RGB = 255 255 255

White (As measured)
Chromaticity  x 0.2668 y 0.2770
CCT 12952 K
Luminance 270 Cd/m2

Black (As measured)
Chromaticity x 0.3114 y 0.3352
CCT 6547 K
Luminance 2.46 Cd/m2

Here we can see the problem

So:
-That Ultrasharp 43″ unit (that one) is defective
-You have legal/full levels misconfiguration on you computer
-That CR200 is not useful anymore, at least for dark readings.
or a mix of them or even some unlisted issue.

Try to find somebody with an i1d3 or even a SpyderX or newer, or even an i1Pro2/3 and read BW after factory reset.
If these see the same black ~1-2nit and associates 100-200:1, the issue is in GPU configuration (typical video level issues, check OS & GPU driver settings) or monitor itself (defective, so return or use warranty).
If these devices read a black closer to 700-1000:1 static CR… bad news, your CR is not very reliable, at least for dark readings. Start planing its renewal.

• This reply was modified 9 months, 3 weeks ago by Vincent.

SpyderX Pro on Amazon  
Disclosure: As an Amazon Associate I earn from qualifying purchases.

[Roger Breton]

Vincent wrote:

-That CR200 is not useful anymore, at least for dark readings.

It is a CS-200 and at $14,000 US a pop, I tend to trust it more than any of my other instrument.

Could be I have a defective monitor.
Could be my GeForce GTX3080 Ti is fucked.

[DaniJ]

There’s a reason a pro kit includes both a spectrometer and a fast & sensitive colorimeter: spectrometers were never designed for measuring dark values.

[Author]

Posts