- This topic has 7 replies, 2 voices, and was last updated 4 years, 12 months ago by
.
-
Posts
-
Hi, I need some help.
I am on my 4th monitor, so, 3rd RMA, and again the screen uniformity is terrible, and this time I actually told them to open the damn box and check it before they sent it off to me and it still drops by a max of just a few 0.0x points shy of 26% brightness in the bottom right corner compared to the centre. Less sever drops are found in the bottom left and along the right side. This part I can at least easily prove with DC and my calibrator. (Thank-you.)
The second part I was hoping I could get some information on other ways that this company can check contrast ratio, because lower than spec contrast ratio, (880 at best, instead of the spec sheet of 1000, which I know is “theoretical”) has also been a big problem. This isn’t even a cheap monitor. It was $1400 plus tax when I got it when it first came out. It’s a 27″ IPS W-LED.
Now here is the hilarious part. This company told me that their tech/repair/refurb department got back to them and they have no way of checking what the monitors contrast ratio is. I actually lol’d at the CS rep in an email. How can a multi-national company, 1, say their monitors do 1000:1 contrast ratio when they can’t even check it, and 2, how can their repair facility say that a monitor is working to spec if they can’t test for contrast ratio? lol It totally boggles my mind.
So, what I was hoping, was that someone, is probably you Florian, (thank-you, again) could give me a possible alternative way to calculate actual contrast ratio. They have a back port on the monitor especially for maintenance, and they sell per-calibrated monitors so I would think they should have the hardware…. I was just thinking, if they can measure the cd/m2 at 6500k white at a brightness of 31 (that’s what I’ve found to be comfortable), at the centre, and at the corners, could they use those 2 numbers from each spot to come up with the contrast ratio and screen uniformity manually?
This had to stop. I added up the total emails in my in and out boxes and it was about 200 emails starting in Jan 2018. No, that isn’t a typo. 20…. 18…….
Help me. (Or, alternatively, you may also shoot me. Just say’n…)
Thanks!
DM
Theoretical static CR 1000:1 is the maximum for most IPS-like panels. Some of them are a higher like those Panasonic widegamut with W-LED PFS (1500:1) or some sRGB WLEDs (~1100-1200:1) , some of them are slightly lower like most AHVAs from AU Optronics.
That maximum value is achieved at panel NATIVE white point. Usually that white point is cooler than common D65 or even slightly cian (cool white+green) in some widegamuts.
If you wish D65 white, monitor electronics should limit one or more channels maximum output (blue and maybe green).
This lowers contrast (since with RGB input 0,0,0 light leak remains the same)… there is no other way to do it.
So… let’s say that you own an AHVA and reseller applied a D65 calibration. Then 880:1 seems reasonable/expected.
IDNK which model you own.The only current way to solve unifomity issues is paying for premium models, usually photo editing models (so they may be “slow” for some users).
Some models have uniformity compensation functionality… which is done at the expense of contrast, droping it, but some Panasonic IPS WLED PFS panel 1500:1 like those used in latest 27″ PA or CG from NEC/Eizo should be able to keep 1000:1 even with uniformity compensation enabled at max.
They are close to your $1400 + tax refered budget, but they are 60Hz QHD and not strictly oriented towards gaming.Thanks for the reply. And…
Oh, bugger it. It’s an Asus PG279Q, which is a 27″ IPS W-LED, 144OC Refresh Rate monitor with G-Sync. And it was a little over $1400 when I got it.
And yes, as you mentioned, all the panels/monitors I’ve gotten have run heavy in the blue and the green. The green seems more sensitive to contrast changes although I eventually saw a correlation between me upping the contrast, only to have to lower the G and B values. The only way I saw contrast ratio gains was by allowing the white point to skew to blue/green. Through a bit of quick testing I realized that I was basically doing the same thing as keeping the contrast ratio the same, but by other means.
Another concern I have on top of those 2, is the gamut range of this panel. It is lower then the other ones. I can only get 98.3% sRGB instead of 99% or even 99.5% on my first monitor that went south and started this long journey. Also, by looking at DisplayCal’s measurements I’m only getting, Red – 96.9% (248/256), Green – 97.3% (249/256), Blue – 96.9% (248/256) and while not multiplied with the others to get the total number of colours I’m posting the Grey results as well. They are also 96.9% (248/256).
So, you might be asking yourself, what’s the big deal over 0.7%? I’ll tell you. At 100% sRGB coverage a regular 8 bit panel has 3 channels; red, green, and blue. These channels can have any value between 0 and 255 although it is often expressed as 1 to 256. You get the total number of colours (and I’m sorry if this is all old stale news to you) by multiplying 256 x 256 x 256. That gives you a total of a possible 16,777,216 colour variations. I believe my original monitor got about 99.5% sRGB and mid 920’s contrast ratio. That number would be 16,693,330 colours. My best result on this current monitor gives me RGB values of 248/256, 249/256, and 249/256. If you multiply the actual numbers out, you get 15,376,248. That means, compared to my first monitor, I am losing a total of 1,317,082 possible colours! That is a LOT of colour! What you end up with is an effect called “posterization” or “banding” where you can see a line where the colour changes. This is seen most often in the clear sky of a wide angle sunset, or with a polarizing filter causing the sky to go from a light blue to a dark blue within the same image.
Thanks for the reply. And…
Oh, bugger it. It’s an Asus PG279Q, which is a 27″ IPS W-LED, 144OC Refresh Rate monitor with G-Sync. And it was a little over $1400 when I got it.
AHVA + gaming = we do not care about QC, it does not matter how much did you pay. This segment behaves that way… maybe latest 27″ QHD Foris is good, IDNK.
Another concern I have on top of those 2, is the gamut range of this panel. It is lower then the other ones. I can only get 98.3% sRGB instead of 99% or even 99.5% on my first monitor that went south and started this long journey.
I should not care about that. Just an intersection… native gamut boundaries would be outside sRGB like in all these WLED sRGB 99%, so games are goint to be a little bit oversaturated unless you use some kind of LUT3D or factory gamut emulation.
Really , you should not care about sRGB intersection.Also, by looking at DisplayCal’s measurements I’m only getting, Red – 96.9% (248/256), Green – 97.3% (249/256), Blue – 96.9% (248/256) and while not multiplied with the others to get the total number of colours I’m posting the Grey results as well. They are also 96.9% (248/256).
If you used RGB gains to fix white, it is not releted to your contrast issues. That % is what you keep in grey levels after fixing grey & gamma.
If you own a GPU with >8bit LUTs and dithered outputs you loose nothing. It’s virtually akin to HW (non LUT3D) calibration in terms of image quality. No banding or posterization due to calibration.
If you don’t have such GPU and care a lot about banding issues caused by calibration… maybe you should try another kind of GPU or get a monitor with HW calibration.Banding you see in games could be caused by games too (visible even on an ideal monitor)… it is not wise to test this issue there. Use a smooth gradient in a non color managed enviroment. Lots of examples in this forum.
Hi, thanks for your reply. I do have a couple questions that I need some clarification on as it seems that English isn’t your first language, which is totally cool. I only know English, so good on you for learning more than one!
When you say AHVA+gaming=no QC, do you mean that most of these monitors will have screen uniformity problems?
Could you explain what you mean by “sRGB intersection?” It’s the “intersection part I don’t quite understand.
I have an Asus Nvidia 1070 video card. I believe it can do 10bit in Directx, although it may be dithered. Anyways, due to, basically money. Nvidia forces you to get a pro card, which isn’t made for gaming. Sadly, they don’t make a 10 bit gaming & pro card. 🙁
When you say AHVA+gaming=no QC, do you mean that most of these monitors will have screen uniformity problems?
I mean that manufacturers usualy do not care at all about quality control in that segment. People buy them because >100Hz stuff and sometimes advertised values of X miliseconds to some unknown paramater.
Could you explain what you mean by “sRGB intersection?” It’s the “intersection part I don’t quite understand.
Intersection of two sets, like “set theory” in school.
Think of gamut as a solid in a 3D space, like L*a*b* (3 dimensions) plots from DisplayCAL or other software.
Intersect you monitor gamut with sRGB colorspace. That gives you a value in %.
This kind of LED sRGB 99% have a native green outside (more saturated than) sRGB, usually same for red. This is not a problem since they are close to sRGB counterparts but color difference between your display native gamut RGB primaries and sRGB primaries is going to be bigger than the tiny 0.x-1.x% faction of colors from sRGB that your display cannot show.
So… do not care about it.If you are worried about getting the “exact colors”:
-external LUT3D or monitos with that fucntionality
-color management & profiles, bu this is not suitable for games
-Reshade LUT3D, I’ve not tested it, there is a FAQ or tutorial.I have an Asus Nvidia 1070 video card. I believe it can do 10bit in Directx, although it may be dithered. Anyways, due to, basically money. Nvidia forces you to get a pro card, which isn’t made for gaming. Sadly, they don’t make a 10 bit gaming & pro card.
There is a thread somewhere in this forum about enabling dither in geforces. I did not test it, but users look happy with the trick.
ATI/AMD gamer cards dither by default since 2005 without tweaks.OK thanks for the explanation. But I do still have 1 question. Using DC to see what my RGB XXX/256 numbers are, does technically still mean that I am losing 1 million subtle shades of colour? Or no, I’m not?
Despite my horrible luck, or their QC, their CS reps have been great. They even suggested that I possibly look at another monitor. I rejected that before because this one has most of the things I want, but since the panel it uses isn’t RGB balanced, it may be great on paper, but in real life you need to be very lucky to get a great display. Now for most, that doesn’t even enter their minds. They just go with what they get. But for a photographer like me, who also likes to game, this was suppose to be a good balance between the two worlds. I’ll probably have to give up the high refresh rate and G-Sync and go with a monitor more for art work then gaming. *sigh*
Thanks! (And thanks, I’ll look for the “dither trick.”)
OK thanks for the explanation. But I do still have 1 question. Using DC to see what my RGB XXX/256 numbers are, does technically still mean that I am losing 1 million subtle shades of colour? Or no, I’m not?
DisplayCAL computes calibration output (correction) at 16bit from a 8bit input. You can imagine this correction as 8bit integer plus 8bit “decimal” correction applied to a 8bit input. Such calibration is stored in VCGT tag in ICM file. You can inspect them or even see its values in human readeable form (*.cal file)
With an 8bit LUT hardware you are going to map at least 1 output value twice unless uncalibrated display matches your desired target.
With 10-12+ bit LUTs you keep most of those decimals.
With dithered output from a 10-12-14-16bit LUT you keep visually 256 steps even if you limit your output to 8bit. This what ATI did many years ago, that’s why people liked/likes them… it works by default (at least in Windows and with propietary drivers in Linux).
I thibk that your geforce should support it too.Monitors with HW cal use a similar appoach in its internal LUTs. Panel could be 8bit native but correction is done with more bits, then dithered to avoid visual truncation errors.
Strong corrections, loosing too many unique greys levels, may be beyond what dither could offer you. But it is not your situatiuon so do not worry about it.
