- This topic has 83 replies, 5 voices, and was last updated 9 months, 2 weeks ago by
.
-
Posts
-
Vincent/Roger:
I aligned my Dell IPS WCG white and black to D50 (approximate) contrast was reduced to 450:1 compared to native 1200:1
Unlike Roger’s Dell LED, mine has R G B gain and offset color controls. Roger’s (much more recent design) unifies these into one set of R G B.
These controls are coarse at black, with a resolution of several dE’00.
Interestingly it’s the aligning of black that leads to most of the loss of CR: setting white to 5000K reduces CR from 1200 to 1050. Setting black to 5000K lowers it to 450.
I am using i1Display Plus with a CCSS made for my model of LED WCG Dell. It appears to work accurately at D65, and result matches other devices. But targeting 5000K, the color appearance of Dell LED WCG is horrible. I would not use it because it looks like slime. By comparison with a daylight CCFL display aligned to D50, or 5000K fluorescent desk lamp (CRI > 90) I would just visually match the LED Dell to another D50 device to match white.
By doing a visual match of Dell LED to D50, much less attenuation of LCD is required which helps CR.
Last point: My Dell LED transfer response at D50 is also horrible with lumps and crossovers. Unlike native, it would be impossible to use at a 5000K white without a VCGT cal.
I assume Roger’s later model improves over my older model somewhat, but not substantially, as it’s basically the same panel tech,. Maybe less peaky spectral response which might make it more well-behaved…?
Roger/Vincent,
I went back and retraced every step, visually matching white with the daylight 5000 lamp, resetting display and re-configuring gains and offsets, double-checking all DisplayCAL settings and re-calibrating
Attached is the verification result
Display is tracking D50 at best conformance and contrast ratio is almost 1000:1 at 136 cdm2
This alignment is still looks too yellow by my reckoning, even though I’ve visually checked white against daylight 5000K lamp. pre-cal measurement checks out D50, and verification passe. But my impression is that everything on my desk under the lamp looks cleaner and whiter than on screen. Screen looks yellow —MYSTERY— If I were to adopt this alignment I would make it bluer even though that disturbs measurements.
My main point was to show that no contrast performance needs to be sacrificed to hit D50 on Dell LED and this experiment shows me this is possible.
So please keep investigating settings on your display. I believe there’s no reason why you can’t get similar results.
/wire
See attached verification report
UP2516D/UP2716D are in the same league as U2413/U2713H/U3014 that I mentioned previously in this thread (“older LED widegamuts from Dell” or something like that I wrote), which can get D65 almost within CR spec and D50 >700:1 (older ones spec are 1000:1). UC behavior is a little different and no RTC overshoot but otherwise same segment, updated to 2016s.
This is a totally different market segment than sRGB “U”s like this one (akin to U2715H but bigger and UHD hence the price), so the OSD limitations make sense sv these others.Also we were talking about CR at factory settings, that it should be closer to spec, on a generous range of 700-1000:1. If it keeps <400:1 @ factory settings with the Myro-1/I1pro2/i1d3… something is off, but finding it needs on site testing. So other than provide some guides like HDMI range check or testing with another computer or checking with PGen pattern gen at differebnt settings, I suspect that no further remote support can be provided. Need on site testing by owner, at factory settings if possible to avoid adding “noise” to the problem by RGB gains / other OSD.
Calibrite Display Pro HL on Amazon
Disclosure: As an Amazon Associate I earn from qualifying purchases.I was sending rgb full and tv on auto signal level. It was 1000 contrast on ycvbr with out 4 4 4 chroma on in the tv. Turned on 444 chroma and contrast when to 450. I went to manual on signal type and it was 450 on limited and 1000 on full. The hardware was auto detecting wrong.
He could try change the signal type or look at the manual signal level if there is on. If it is auto and no way to control it could be wrong.
The screen might not be. Some things are better to do onsite than guessing. It needs a little testing. He did test it to factory and it worked at factory but not at his white point. Something goofy to lose a lot a contrast going to D50. It looks like signal levels. Changeing to 422 ycbcr might help. It might be expecting a full level signal.
Vincent wrote:
You are using an backlit LCD (1), and your are driving white level through RGB gains (2). All these things point that you got stuck on CRT era and that you are unwilling (to learn how)to use an LCD/TFT monitor in a proper way.
I believe it’s generally best for fidelity to do color temp as pre-cal in the display, including LCD displays, because this avoids spending scarce values in the display connection on calibration, which can help avoid gradient banding. Internally the display will tend to have a 14 or greater bpc path to the panel because it has to compute linear light Y out from the gamma encoded R’G’B’ signal in, so make use of those bits in the display for pre-cal via internal controls.
But it’s also a good tradeoff to let black temperature stay at LCD display default and have DisplayCAL choose near-black tracking because it can optimize overall TRC and contrast ratio vs appearance of color cast near black.
For 10 bpc data path from gfx to SDR display, pre-cal is probably moot because there’s 2 bpc of head room; so set display to native and do all cal through GPU.
Is this wrong?
No. Display Cal does wonderful and even makes a clipped white not clip and color correction fixes all colors in the color correction apps. I get better results sometimes not changing defaults but changing targets in display Cal. OSD controls avoid banding since it is done internally. Keeping on RGB at default is fine even if 2 RGB is clipped. Leaving one control in default position is good and subtracting colors to gain the other colors in white is a good thing. Hardware calibrated means un color managed is correct to. Unity games do not use color management.
Vincent wrote:
You are using an backlit LCD (1), and your are driving white level through RGB gains (2). All these things point that you got stuck on CRT era and that you are unwilling (to learn how)to use an LCD/TFT monitor in a proper way.
I believe it’s generally best for fidelity to do color temp as pre-cal in the display, including LCD displays, because this avoids spending scarce values in the display connection on calibration, which can help avoid gradient banding. Internally the display will tend to have a 14 or greater bpc path to the panel because it has to compute linear light Y out from the gamma encoded R’G’B’ signal in, so make use of those bits in the display for pre-cal via internal controls.
But it’s also a good tradeoff to let black temperature stay at LCD display default and have DisplayCAL choose near-black tracking because it can optimize overall TRC and contrast ratio vs appearance of color cast near black.
For 10 bpc data path from gfx to SDR display, pre-cal is probably moot because there’s 2 bpc of head room; so set display to native and do all cal through GPU.
Is this wrong?
white level = nits (supposedly driven by RGB gains, in the past), also we were talking about RGB gains linearity, IF they behave wildly stop at some point where they are still well behaved, then do the last jump on VCGT.
When OP showed some of hist test RGB gsins seem to be linear and at factory settings :
After monitor reset, full Brightness / Contrast / Gains, I get :
White 349 cd/m2
Black 1.61 cd/m2CR = 349 / 1.61 = 216
So the original issue after all these tests done by OP is caused by monitor itself (strange bad behavior far bellow spec, candidate to return for refund) or some video/full level issues in GPU (although configuration seems OK).
Further diagnosis of the problem requires testing by owner
