Home › Forums › Help and Support › Blacks appear more bright on OLED notebook
- This topic has 13 replies, 3 voices, and was last updated 1 year, 4 months ago by Prapan Chulapinyo.
-
AuthorPosts
-
2022-11-05 at 7:54 #37594
I’ve calibrate Asus Vivobook Pro 15 OLED. And it work well, color is good match according to verify. Only blacks appear more bright compare to Dell IPS display that calibrate from Displaycal.
Single-curve + LUT. Almost default only change is white 150cd, patch around 2000, calibration speed medium. What can cause this? How to get black level appear near the IPS panel?
2022-11-05 at 11:10 #37595Try XYZ LUT + matrix, which is the most accurate type of color profile.
2022-11-05 at 11:56 #37596Try XYZ LUT + matrix, which is the most accurate type of color profile.
Yes, that is what I’ve done.
Black level: As measured
Tone curve: Gamma2.2 , black output offset 100% > (that cause the problem?)
Black point correction: no check Auto, 0%, Rate 4.00
- This reply was modified 1 year, 5 months ago by Prapan Chulapinyo.
2022-11-05 at 12:35 #37599Verify = ICC profile matches display behavior
Verify =/= display matches your expected behaviorgrey calibration priorizes grey neutrality and grey levels separation over target TRC (“gamma”). You can see this in veryfy report.
Why? because as long as actual TRC is recorded on profile color managed apps will work as expected. Windows desktop is not color managed and a lot of video players are not color managed hence there is no futher conversion to target gamma for non color managed apps.
You can use DMWLUT if you actually want that nominal target gamma, but near dark greys will be less hard to separate visually.2022-11-06 at 10:50 #37611Verify = ICC profile matches display behavior
Verify =/= display matches your expected behaviorOk I’ve just don’t sure if I do something wrong with setting of OLED profiling or not. I just do the same with IPS panel except I enable white level drift compensation. Both: single curve + matrix, XYZ LUT + matrix.
It OLED panel made by Samsung (ATNA56YX03-0), AM-OLED. I search from google and I use RGB OLED family (sony, samsung, lenovo) correction with Xrite i1display pro.
Windows desktop is not color managed and a lot of video players are not color managed hence there is no futher conversion to target gamma for non color managed apps.
OK, I not see it myself but it my friend which I calibrate his Dell IPS display before. When He use this OLED display with his photo edit on Photoshop he told me that their Dell IPS on black area look more black, less detail but compare to this OLED after done calibrate/profile it on the dark area more light and more detail than Dell one. He tend to believe on Dell IPS one than OLED.
Again, I not see it myself compare yet. I don’t have verify report now. Once I go to my friend I’ve post it here. I need to go and see it myself some other day.
That why I come to ask is this normal behaviors or not? Or it just profile match the display already, only nature of OLED that more contrast and show the dark area better than IPS panel? If that the case should I keep using it and accept it as normal?
Or should I put calibration black level of OLED equal to black level of IPS panel to have same black visual level to their eyes? It will help – put same black level? Or not much because how high contrast of OLED panel? Black level of this OLED is 0.0000 cd/m^2 and Dell IPS is around 0.1-0.15 cd/m^2 (not sure exact).
- This reply was modified 1 year, 5 months ago by Prapan Chulapinyo.
Calibrite Display Pro HL on Amazon
Disclosure: As an Amazon Associate I earn from qualifying purchases.2022-11-06 at 13:02 #37614Verify = ICC profile matches display behavior
Verify =/= display matches your expected behaviorOk I’ve just don’t sure if I do something wrong with setting of OLED profiling or not. I just do the same with IPS panel except I enable white level drift compensation. Both: single curve + matrix, XYZ LUT + matrix.
It OLED panel made by Samsung (ATNA56YX03-0), AM-OLED. I search from google and I use RGB OLED family (sony, samsung, lenovo) correction with Xrite i1display pro.
That’s nothing to do with profiling. I had explained it to you on previous message: on color managed apps (ICC aware) it does not matter actual gamma value (your raised dark grays) as long as profile records actual gamma. A gamma 1.8 dsplay and gamma 2.4 display should render same sRGB image in a color managed app as long as their profiles matches each display behavior.
So ArgyllCMS gray calibration aims to minimize grey color error and grey brightness difference rather to aim to “exactly” your desired constant gamma across all greyscale.On non color managed apps this display TRC mismatch vs your desired target gamma may show that issues, but if you need to use an RGB OLED widegamut for non color managed apps, you’ll have color issues too so using a LUT3D and aiming to desired TRC will solve oversaturation and TRC.
Windows desktop is not color managed and a lot of video players are not color managed hence there is no futher conversion to target gamma for non color managed apps.
OK, I not see it myself but it my friend which I calibrate his Dell IPS display before. When He use this OLED display with his photo edit on Photoshop he told me that their Dell IPS on black area look more black, less detail but compare to this OLED after done calibrate/profile it on the dark area more light and more detail than Dell one. He tend to believe on Dell IPS one than OLED.
Read doc and see what BPC does to profiles. It’s fake infinite contrast to avoid some software oversimplifications.
Again, I not see it myself compare yet. I don’t have verify report now. Once I go to my friend I’ve post it here. I need to go and see it myself some other day.
That why I come to ask is this normal behaviors or not? Or it just profile match the display already, only nature of OLED that more contrast and show the dark area better than IPS panel? If that the case should I keep using it and accept it as normal?
Or should I put calibration black level of OLED equal to black level of IPS panel to have same black visual level to their eyes? It will help – put same black level? Or not much because how high contrast of OLED panel? Black level of this OLED is 0.0000 cd/m^2 and Dell IPS is around 0.1-0.15 cd/m^2 (not sure exact).
You’ll need to see if those displays match their profile and actual TRC, in a report. Once you have it we can see what is actually happening.
Report with all simulation options off, just test if profile matches each display. That report will also show actual TRC.
For example, IPS display may have an inaccurate profile that darkens dark greys in color managed apps because it profile TRC says its TRC is higher than actual value.
Whatever wrong behavior you see, TEST, measure. Always.
2022-11-22 at 9:47 #37767You’ll need to see if those displays match their profile and actual TRC, in a report. Once you have it we can see what is actually happening.
I have report files already. And from what I see OLED laptop screen in compare with U2410 and it quite normal in dark tone now. I have more to ask:
- Display Dell U2410 profile made 3 month ago show shift in white point but averageE and maxE still in recommend and nominal range. It does mean screen is deteriorate or age? Do you think this display still good condition?
- For OLED Asus notebook I don’t know why uncalibrate report show white ~6254K-64488K normally notebook I’ve calibrate is around 7000-8000K , there is MyAsus app which I set to native, notmal is this cause it?
- Pixel shift should be turn off during profiling? I do turn off. And good to turn on when not calibrate to safe screen, or am I miss something? For ABL I’ve not found any setting menu however I do turn on white level drift compensation. For screen safer at the time of profiling kick-in sometime I don’t know how to turn off (now I know there is in MyAsus app) so keep move mouse during profiling which is fine.
- For OLED Asus notebook I do profile twice, 1st with LUT and Matrix 2nd is 13days later with Matrix. Both 1, 2 show high delta E on some dark color of red and dark color of green. As I said I’ve done twice, 3days later. But result is same. Other than that -those in dark red, green- it fine I think. I think nothing I can do to improve this, or I miss something? I use Xrite i1display pro with medium calibration speed, high profile quality, auto-optimized testchart.
- This reply was modified 1 year, 4 months ago by Prapan Chulapinyo.
Attachments:
You must be logged in to view attached files.2022-11-22 at 9:48 #37772This report is for OLED Asus laptop.
Attachments:
You must be logged in to view attached files.2022-11-22 at 11:08 #37784You’ll need to see if those displays match their profile and actual TRC, in a report. Once you have it we can see what is actually happening.
I have report files already. And from what I see OLED laptop screen in compare with U2410 and it quite normal in dark tone now. I have more to ask:
- Display Dell U2410 profile made 3 month ago show shift in white point but averageE and maxE still in recommend and nominal range. It does mean screen is deteriorate or age? Do you think this display still good condition?
Contrast is a bit low. Main variation is win white point, amybe CCFL is changing too fast, grey calibration range goes high too..
- For OLED Asus notebook I don’t know why uncalibrate report show white ~6254K-64488K normally notebook I’ve calibrate is around 7000-8000K , there is MyAsus app which I set to native, notmal is this cause it?
Native WP depends on whatever technology is used + factory calibration. Your RGB OLED is not a WLED.
- Pixel shift should be turn off during profiling? I do turn off. And good to turn on when not calibrate to safe screen, or am I miss something? For ABL I’ve not found any setting menu however I do turn on white level drift compensation. For screen safer at the time of profiling kick-in sometime I don’t know how to turn off (now I know there is in MyAsus app) so keep move mouse during profiling which is fine.
IDNK ABL limit at patch size 100% for your notebook, or its variation with temperature, you’ll have to measure it by yourself.
- For OLED Asus notebook I do profile twice, 1st with LUT and Matrix 2nd is 13days later with Matrix. Both 1, 2 show high delta E on some dark color of red and dark color of green. As I said I’ve done twice, 3days later. But result is same. Other than that -those in dark red, green- it fine I think. I think nothing I can do to improve this, or I miss something? I use Xrite i1display pro with medium calibration speed, high profile quality, auto-optimized testchart.
Big error near back red (<26) is due to some issues with i1d3 device. Read RGB OLED threads, report your issues to ArgyllCMS maillist. Or ignore those outliers.
Errors over red 38 seems due to display non linear behavior: you can capture it with a 3D mesh (XYZLUT profile 1st report) but you cannot predict that behavior with an idealized perfect additive display ( matrix profile, 2nd & 3rd).
That’s how display behaves… but snce it is localized, maybe it’s nt a big deal
If you wish to use it accurately (over red 26) without XYZLUT porfiles (due to Photoshop & otehr apps issues with 3 TRC curves) you can try to “idealize display” with DWMLUT to some colorspace akin your laptop native… but since non ideal behavior is located in those outlier … maybe it is not worth the time to do it.Grey calibration seems not very good from the start, huge grey range color shifts, try to decrease calibration speed to slow (but calibration stage-grey calibration stage- , before profiling stage, may go over 30min)
2022-11-23 at 8:34 #37785Contrast is a bit low. Main variation is win white point, amybe CCFL is changing too fast, grey calibration range goes high too..
So if it keep changing too fast it mean it old? Well this Dell U2410 what I know so far is >5 years.
Native WP depends on whatever technology is used + factory calibration. Your RGB OLED is not a WLED.
I see, so it depend on that tech it used, and or factory calibration. Ok.
IDNK ABL limit at patch size 100% for your notebook, or its variation with temperature, you’ll have to measure it by yourself.
I see, so far while profiling/calibration, my eye don’t see reduce on brightness yet.
Big error near back red (<26) is due to some issues with i1d3 device. Read RGB OLED threads, report your issues to maillist. Or ignore those outliers.
It issue from kind of hardware of i1d3? RGB OLED threads is mean search forum in displaycal? , I found here:
Or RGB OLED threads on ArgyllCMS? Then for those near black red will I get the correct/acculate result?
Errors over red 38 seems due to display non linear behavior: you can capture it with a 3D mesh (XYZLUT profile 1st report) but you cannot predict that behavior with an idealized perfect additive display ( matrix profile, 2nd & 3rd).
I see, I understand it.
For DWMLUT I’ve no experience about it yet. But I found you mention in some other post, maybe I’ll study it later. In short what is it benefit over XYZLUT and Matrix? Also what is it drawback? I mean I try to understand why you recommend it. From some search I found this:
Grey calibration seems not very good from the start, huge grey range color shifts, try to decrease calibration speed to slow (but calibration stage-grey calibration stage- , before profiling stage, may go over 30min)
Ok if I have a chance next time I’ll try. And yes it take time around 30-45min. Way I choose medium because I’ve experiment at my screen at home show very few improvement of grey range from medium to low. But that with another display not this one (with Dell 2417H, and some IPS display screen)
Grey calibration seems not very good from the start <— you notice this from evaluate RGB + gray balance and check ‘Evaluate gray balance through calibration only’ then look at both RGB gray balance right? As I attach picture, is that the way to evaluate this, correct?
Attachments:
You must be logged in to view attached files.2022-11-23 at 9:06 #37787Contrast is a bit low. Main variation is win white point, amybe CCFL is changing too fast, grey calibration range goes high too..
So if it keep changing too fast it mean it old? Well this Dell U2410 what I know so far is >5 years.
Native WP depends on whatever technology is used + factory calibration. Your RGB OLED is not a WLED.
I see, so it depend on that tech it used, and or factory calibration. Ok.
IDNK ABL limit at patch size 100% for your notebook, or its variation with temperature, you’ll have to measure it by yourself.
I see, so far while profiling/calibration, my eye don’t see reduce on brightness yet.
Big error near back red (<26) is due to some issues with i1d3 device. Read RGB OLED threads, report your issues to maillist. Or ignore those outliers.
It issue from kind of hardware of i1d3? RGB OLED threads is mean search forum in displaycal? , I found here:
Or RGB OLED threads on ArgyllCMS? Then for those near black red will I get the correct/acculate result?
It’s clipping on pure reds < 26. If you wish that Graeme Gill finds some kind of solution you’ll have to post it in ArgyllCMS maillist and provide data.
I Mean, USB raw counter values are 0? or they become 0 after being weighted by RGB raw to XYZ from software?
If USB raw counter values read by colorimeter are 0, is it possible to change integration time for these RGB OLEDs, so an argyllcms version can take readings?This has nothing to do with DisplayCAL, although you can read other people issues with those reds. Solution if any has to come from ArgyllCMS new version. If no data and no reports… there will be no solution.
Errors over red 38 seems due to display non linear behavior: you can capture it with a 3D mesh (XYZLUT profile 1st report) but you cannot predict that behavior with an idealized perfect additive display ( matrix profile, 2nd & 3rd).
I see, I understand it.
For DWMLUT I’ve no experience about it yet. But I found you mention in some other post, maybe I’ll study it later. In short what is it benefit over XYZLUT and Matrix? Also what is it drawback?
XYZLUT is a 3D mesh, like a n x n x n cube recording all display behavior irregularities at “n ^3 ” resolution. 17 node per cube edge is about 5000! patches!
Matrix + single curve is an idealization of display: perfectly additive and true neutral grey.
A 3D mesh with enough points can be more accurate but it may be not suported by some iCC software. Also if you use 3 TRC curves instead a single oen, Photoshop and other tools will make ugly rounding errors because of their lack of temporal dithering (maybe 10bit open GL output fixes it… but Illustrator and Indesign do not have it).
If display is well behaved it is advised to use “matrix single curve” profiles because it is an accuarte description (display well behaved) and you avoid these ugly rounding errors due to color management oversimplification in calculations.
In your OLED if you fix grey scale issues, the lower red irregular behavior may be unnoticed so… maybe i’ll stey with matrix singe curve.I mean I try to understand why you recommend it. From some search I found this:
Grey calibration seems not very good from the start, huge grey range color shifts, try to decrease calibration speed to slow (but calibration stage-grey calibration stage- , before profiling stage, may go over 30min)
Ok if I have a chance next time I’ll try. And yes it take time around 30-45min. Way I choose medium because I’ve experiment at my screen at home show very few improvement of grey range from medium to low. But that with another display not this one (with Dell 2417H, and some IPS display screen)
Grey calibration seems not very good from the start <— you notice this from evaluate RGB + gray balance and check ‘Evaluate gray balance through calibration only’ then look at both RGB gray balance right? As I attach picture, is that the way to evaluate this, correct?
If profile is “single curve” (all equal TRC curves), expected values in a*b* for grey are 0, hence “Evaluate gray balance through calibration only” won’t be needed.
2022-11-23 at 10:00 #37788It’s clipping on pure reds < 26. If you wish that Graeme Gill finds some kind of solution you’ll have to post it in ArgyllCMS maillist and provide data.
I Mean, USB raw counter values are 0? or they become 0 after being weighted by RGB raw to XYZ from software?
If USB raw counter values read by colorimeter are 0, is it possible to change integration time for these RGB OLEDs, so an argyllcms version can take readings?Data, report you say. What do I have to provide them? Is report html files output OLED notebook from displaycal is enough? Since it have option to click view raw reference data. Just I’m not sure it good idea or not to ask ArgyllCMS mailing list by provide html output from displaycal?
A 3D mesh with enough points can be more accurate but it may be not suported by some iCC software. Also if you use 3 TRC curves instead a single oen, Photoshop and other tools will make ugly rounding errors because of their lack of temporal dithering (maybe 10bit open GL output fixes it… but Illustrator and Indesign do not have it).
For Photoshop with XYZLUT I’ve read that that may have compatibility issue. But what happen if I use it? I ask because I see people still use LUT with Adobe software and still happy. What am I missing? Or what should I concern about?
It that what I expect to see from rounding error? A banding? Or un-accurate near dark from rounding error? If I use XYZLUT with Adobe for ex. Photoshop not just this laptop but some other computer as well.
In your OLED if you fix grey scale issues, the lower red irregular behavior may be unnoticed so… maybe i’ll stey with matrix singe curve.
Ok I get it. Thank. 🙂
If profile is “single curve” (all equal TRC curves), expected values in a*b* for grey are 0, hence “Evaluate gray balance through calibration only” won’t be needed.
Got it.
If profile is “single curve” (all equal TRC curves), expected values in a*b* for grey are 0, hence “Evaluate gray balance through calibration only” won’t be needed.
Well for evaluate gray calibration, I not sure should I evaluate it at summary in picture #1 or at overview in picture #2 from color distance deltaC and deltaE.
Or both meaning the samething? I’ve this question for sometime now. Hope you can clarify me.
Attachments:
You must be logged in to view attached files.2022-11-23 at 15:32 #37794It’s clipping on pure reds < 26. If you wish that Graeme Gill finds some kind of solution you’ll have to post it in ArgyllCMS maillist and provide data.
I Mean, USB raw counter values are 0? or they become 0 after being weighted by RGB raw to XYZ from software?
If USB raw counter values read by colorimeter are 0, is it possible to change integration time for these RGB OLEDs, so an argyllcms version can take readings?Data, report you say. What do I have to provide them? Is report html files output OLED notebook from displaycal is enough? Since it have option to click view raw reference data. Just I’m not sure it good idea or not to ask ArgyllCMS mailing list by provide html output from displaycal?
Mailist, something like:
“I’ve and RGB OLED, my i1d3 gets back clipping under RGB 26,0,0. Is there a way to solve this issue by changing some params in HID command to do readings. I’m willing to test beta versions or provide log files you ask me for”.
I do not own an RGB OLED. I cannot do that.
A 3D mesh with enough points can be more accurate but it may be not suported by some iCC software. Also if you use 3 TRC curves instead a single oen, Photoshop and other tools will make ugly rounding errors because of their lack of temporal dithering (maybe 10bit open GL output fixes it… but Illustrator and Indesign do not have it).
For Photoshop with XYZLUT I’ve read that that may have compatibility issue. But what happen if I use it? I ask because I see people still use LUT with Adobe software and still happy. What am I missing? Or what should I concern about?
It that what I expect to see from rounding error? A banding? Or un-accurate near dark from rounding error? If I use XYZLUT with Adobe for ex. Photoshop not just this laptop but some other computer as well.
3xTRC errors: colored banding on grayscale gradiest, this will be the one easier to spot. And teh way to fix it is to get a very neutral grey calibration and use 1x TRC (3 equal TRCs)
In your OLED if you fix grey scale issues, the lower red irregular behavior may be unnoticed so… maybe i’ll stey with matrix singe curve.
Ok I get it. Thank.
If profile is “single curve” (all equal TRC curves), expected values in a*b* for grey are 0, hence “Evaluate gray balance through calibration only” won’t be needed.
Got it.
If profile is “single curve” (all equal TRC curves), expected values in a*b* for grey are 0, hence “Evaluate gray balance through calibration only” won’t be needed.
Well for evaluate gray calibration, I not sure should I evaluate it at summary in picture #1 or at overview in picture #2 from color distance deltaC and deltaE.
Or both meaning the samething? I’ve this question for sometime now. Hope you can clarify me.
inspect a* b* measured values and “grey range a*b*”. Mostly a*. dC will be a summary. dE includes brightness error (gamama) so it has no use here.
2022-11-24 at 8:42 #37799Mailist, something like:
“I’ve and RGB OLED, my i1d3 gets back clipping under RGB 26,0,0. Is there a way to solve this issue by changing some params in HID command to do readings. I’m willing to test beta versions or provide log files you ask me for”.
I do not own an RGB OLED. I cannot do that.
I’ll do that when have chance to get to that notebook again.
3xTRC errors: colored banding on grayscale gradiest, this will be the one easier to spot. And teh way to fix it is to get a very neutral grey calibration and use 1x TRC (3 equal TRCs)
So it trade off, I want accurate capture most display behavior irregularities but live with color banding on grayscale gradiest : XYZLUT, or very neutral grey calibration no banding and accurate description only if display well behaved : Matrix + single curve then. I see.
inspect a* b* measured values and “grey range a*b*”. Mostly a*. dC will be a summary. dE includes brightness error (gamama) so it has no use here.
I see, so it a* and b* measured and dC for summary from a* and b*.
Thank you very much.
- This reply was modified 1 year, 4 months ago by Prapan Chulapinyo.
Attachments:
You must be logged in to view attached files. -
AuthorPosts