- This topic has 21 replies, 3 voices, and was last updated 2 years, 3 months ago by
.
-
Posts
-
It seems I am a lucky guy who managed to get this gorgeous display which suppose to arrive tomorrow. I am just a hobbyist, and did zero commercial work. This monitor is an overkill for me, but being technical background I tend to pull most of each device I have, maybe that’s some kind of OSD.
Anyhow, I quickly realized this monitor only allows WB adjustments and my iDisplay Pro wouldn’t work with it. No way I can afford spending another $1k just to buy a probe, yet I still want to calibrate the monitor.
The way to do it is by using Resolve’s built in option for 3D LUT (I own mini monitor 4k card) and displayCal.I wondered if there are any users who had used displayCal to calibrate these OLED panels, and if so with what success and if they could give me some tips.
So far I managed to get following:
Display & Instrument:
Display: Resolve
White level drift compensation -> checked (What about the black level drift compensation?).
Not sure what to do with other options, especially with Full field pattern insertion and Override minimum display update delay.
Output level: Full Range
Correction: LED OLEDCalibration:
Observer CIE 1931 with White points: x .307, y .318 (Sonny recommended).
Not sure what to put for White and Black level.
Then I chose curve Rec.709, but this just opened more possibilitiesProfiling tab is pretty much clear.
3D LUT
Source colorspace Rec.709
Tone curve: Rec.1886 – Gama 2.4 – Absolute (as panel is OLED and can “show” black)Anyhow, I attached all the screenshots for easier check. T
Thanks a lot!!!black level drift compensation?
Only for spectrometers.
Not sure what to do with other options
For Resolve, always enable override display update delay 1000ms (should be automatically enabled when choosing Resolve under display).
Not sure what to put for White and Black level.
Leave “as measured”, but adjust white to 100 cd/m2 (SDR) during interactive adjustment.
Then I chose curve Rec.709, but this just opened more possibilities
When creating a 3D LUT, you can skip 1D LUT calibration without sacrificing accuracy (set calibration tone curve to “as measured”).
black level drift coplay update delay 1000ms (should be automatically enabled when choosing Resolve under display).
Thank you very much for responding, I appreciate it.
What about the settle time multiplier? Only concerned not to do something bad to the screen.Also, I just noticed Correction for OLED is WRGB, but Sony PVM is using RGB panel. Wouldn’t this give me wrong measurements?
Cheers!!!
BTW Is there a way to use displaycal for measurement only? I see there’s Verification tab, but I am not sure how to start it.
What about the settle time multiplier?
Leave as-is.
Only concerned not to do something bad to the screen.
Nothing bad can happen to the screen, these options are for Resolve which requires additional delay.
Also, I just noticed Correction for OLED is WRGB
No, “LED OLED (created from primaries W, R, G ,B)” is for RGB OLED.
BTW Is there a way to use displaycal for measurement only? I see there’s Verification tab, but I am not sure how to start it.
Yes, but you need to create a profile first. You can then use the verification tab to compare to any target (simulation profile with “use as display profile” checked) in self check report mode without running additional measurements.
Hi there. I must be doing something wrong. Attached are two Measurement reports. One done immediately after the calibration, and the other after 3D LUT was loaded to Davinci.
I had set DisplayCal as per attached 1-4 jpgs. Could you check if I missed anything here, please?Monitor:
Basically, I started in User 1 mode and copied D65 profile as a starting point on the monitor.
When the calibration started, the luminance was around 88-90 Nits, so I had to dial in the Contrast to 116 from 100.
RGB default settings were 128, 127, 140, and I only had to go +1 on G to 128, 128, 140, and the columns were perfectly aligned. It seems Sony already accounted for Judd offset.
Color space was left as ITU-R BT.709. Should have I set Native here instead of Rec.709?
Gamma 2.4One strange thing as well. I have started Lightspace, and set step to 100%, pulled the RGB bars all the way up, and it measured 121.6 Nits. I am not sure where the difference of 20 Nits is coming from between DisplayCal and Lightspace.
3D LUT encoding needs to be full range, not TV RGB, same for output levels in DisplayCAL.
Color space was left as ITU-R BT.709. Should have I set Native here instead of Rec.709?
Native is preferable because it will not unnecessarily restrict the gamut.
Native is preferable because it will not unnecessarily restrict the gamut.
Thank you again for a quick response. I will try to redo it tonight.
There is one more thing in monitor settings. Should I also set the input range there to Full Range when doing the calibration? The default Monitor’s input value is set to Limited which is what I plan on using.
So, should I set it to Full Range for calibration, and return it to Limited after or…?There is one more thing in monitor settings. Should I also set the input range there to Full Range when doing the calibration?
Leave it as limited. The chain is as follows: DisplayCAL -> full range -> Resolve (converts to limited range) -> limited range -> monitor. From this it should be clear why DisplayCAL needs to be set to full range (or the default, auto), because Resolve will take the signal and convert it to limited range. So an already limited range signal coming into Resolve would be converted again, thus leading to reduced contrast (raised blacks/lowered whites) like in your earlier case.
So an already limited range signal coming into Resolve would be converted again
Ok, that makes sense.
Could it be that the fact I had input, output encoding and output levels in the software held at 16-235 caused Luminance at the beginning of testing (initial RGB, Luminance adjustemnt before calibration) to be wrong as Lightspace was showing 121 Nits for what DisplayCal was showing 100 Nits?
The only difference in measuring Luminance was that in Lighstspace I used White area (all bars pulled all the way up), and Display cal some percentage grey area, but that shouldn’t matter, correct?Could it be […]
Yes, naturally.
The only difference in measuring Luminance was that in Lighstspace I used White area (all bars pulled all the way up), and Display cal some percentage grey area, but that shouldn’t matter, correct?
Well, of course that matters. Any level of gray that is not full white will have lower luminance than the latter. As I said, this is easily fixed by setting output levels and encoding in DisplayCAL correctly (and a reason I recommend using the presets, as they will set everything correctly for you).
It seems I always find something else which makes me doubt.
What about Tone Curve Absolute vs Relative. I had this initially set to Relative, but per description is should actually be Absolute as OLED can show “real” black?So my setting are:
Source colorspace Rec.709
Tone Curve: Custom Gama: 2.4 Absolute (When I click on Absolute it actually moves Tone Curve to Rec.1886, and I need to select Custom again).
Black output offset – 0% because I set the Curve to Absolute, otherwise it should be 100% ?So I did both gamma mapping scenarios. I still get NOK for measured vs assumed target white point, but I guess this is ok because the measured one is what I want.
Gamma is all over the place at higher percentages, but I am not sure if I am reading this well or my probe is not working well or…Is there anything else you’d suggest?
Attached are the results and settings from both sessions.What about Tone Curve Absolute vs Relative. I had this initially set to Relative, but per description is should actually be Absolute as OLED can show “real” black?
Black output offset – 0% because I set the Curve to Absolute, otherwise it should be 100% ?
On devices with zero black level, it doesn’t make a difference.
Custom Gama: 2.4 Absolute (When I click on Absolute it actually moves Tone Curve to Rec.1886, and I need to select Custom again)
Absolute with 0% black output offset = Rec. 1886, but see above.
I still get NOK for measured vs assumed target white point, but I guess this is ok because the measured one is what I want.
Correct.
Attached are the results and settings from both sessions.
Results seem plausible, but I wonder how stable the display is. Did you use the smallest possible patch size during measurements? Black background or default (APL background)? Attach the profile please.
The “no LUT applied” reports also have the LUT applied (device link profile), Resolve doesn’t have a way to apply LUTs via the pattern generator.
