- This topic has 6 replies, 2 voices, and was last updated 6 years, 3 months ago by
.
-
Posts
-
Hello
It’s the first time I write here. I’m a beginner in color management issues and I have some doubts about that, so I hope you can help me.
Well, a few days ago I bought a BenQ PD2700Q monitor. Previously I returned two units because they had dead pixels and bad screen uniformity. Finally I bought a third unit despite having bad uniformity, since my budget is limited. BenQ has disappointed me enough.
I have a Sequel Chroma USB colorimeter (I’m not sure if it’s 4 or 5 model – mine is triangular) that was stored in a drawer for many years.
My computer is an Acer Aspire E15 – Intel Core i5-5200U with an Intel® 5500 HD integrated card and an support card NVIDIA GeForce 820M with 2 GB of dedicated VRAM. The card that drives the monitors is the HD Intel® 5500. The parameters of the Intel card are unmodified.
I have no problems with the Sequel connection and with its controller (i1 Display 1), but when I calibrate and profile with DisplayCal the reports are not very convincing, although I get the profiles without problems. I have tried it in many ways and I can not solve anything.
When I load up the profiles into the system, there isn’t anomaly or dramatic change that could get my attention. I notice that the neutral colors are a little warmer, but the change is minimal. The visible result does not displease me.
I don’t understand why I can’t get the white level value above 185 cd/m2 and the black level below 0.60 cd/m2. The contrast ratio never exceeds 220:1. The monitor specifications indicates that it has 350 cd/m2 maximum brightness and 1000:1 native contrast. I have also observed that data profiles varies slightly using the same parameters in diferents calibratons. I guess that will influence the ambient light…
Once the profile is loaded into the system, the visualization is correct. The blacks are intense and the luminous targets.
The process that I follow in DisplayCal 3.4 is the following (I don’t know if it’s correct or not):– I select in “Settings” Office and web (D65, Gamma 2.2).
– In “Display & instrument” I select the screen BenQ, i1 Display 1 and LCD (Generic). Full range RGB 0-255. Auto correction (none). The other options deselected.
– In “Calibration” White level 120.00 cd / m2. The rest by default.
– In “Profiling” everything by default.
– Click on the “Calibrate and profile” button.
– In the OSD of the monitor I select “User Mode”, “Gamma 2.2” and the rest by default.
– Click “Start measurement”. The current target level is 182.6 cd/m2 with the brightness OSD setting at 100% (much less than 350 cd/m2 monitor specifications, however the light is intense). The white dot indicates x 0.3128 y 0.3292 0.0 ΔE*00 (Destination) and x 0.2960 y 0.3210 8.0 ΔE*00 (Current).
In “White dot/White level” I adjust OSD brightness and color temperature controls on the R-G-B channels, until the DisplayCal interactive monitor changes to green. OSD values: R = 100, G = 91, B = 87, Brightness = 68. Now the interactive monitor tells me that the target level is 120.10 cd/m2 and the white point is x 0.3128 y 0.3292 0.0 ΔE*00 (Destination) yx 0.3126 y 0.3298 0.5 ΔE*00 (Current).
When executing the “adjustment check”, the black level is 0.62 cd/m2 and the black point is x 0.3128 y 0.3292 (Destination) and x 0.2991 y 0.3208 6.4 ΔE*00 (Current). The text has not changed to green.– I continue with the calibration and wait for it to end.
– At the end a window appears that indicates: Profile self check ΔE*76: average 0.13, maximum 3.50, RMS 0.47. Gamma coverage (95.6% sRGB, 74.0% Adobe RGB, 77.2% DCI P3). Gamma volume (109.1% sRGB, 75.2% Adobe RGB, 77.3% DCI P3).
I do not understand what I’m doing wrong.
Is Sequel a bad colorimeter?
The monitor is calibrated at the factory with a 100% sRGB coverage. Why I do not get it?Sorry, a lot of questions…
ERRORS IN THE LOG FILE (text bold)
16:47:45,793 Patch 16 of 16 DE 0.090840, W.DE 90.839710, Repeat (got worse)
16:47:48,325 Patch 16 of 16 DE 0.075488, W.DE 75.488084, Fail ( > 0.071729)
16:47:48,345
16:47:48,348 Brightness error = 0.140684 cd/m^2 (is 119.318815, should be 119.178131)
16:47:48,349 White point error = 0.084844 deltaE
16:47:48,351 Maximum neutral error (@ 0.230832) = 1.145519 deltaE
16:47:48,352 Average neutral error = 0.394019 deltaE
16:47:48,354 Failed to meet target 1.272792 delta E, got worst case 75.488084
16:47:48,355 Number of measurements taken = 56
16:47:48,355 Computing update to calibration curves…
16:47:48,375
16:47:48,378 Doing iteration 2/2 with 32 sample points and repeat threshold of 0.900000 DE
16:47:49,063 Patch 1 of 32 DE 0.107435, W.DE 0.107435, W.peqDE 0.107435, OK ( < 0.900000)16:51:21,772 Patch 32 of 32 DE 0.127731, W.DE 127.730613, W.peqDE 127.730613, Fail ( >
16:51:21,772 0.050000)
16:51:21,778
16:51:21,779 Brightness error = 0.148543 cd/m^2 (is 119.326674, should be 119.178131)
16:51:21,779 White point error = 0.095989 deltaE
16:51:21,780 Maximum neutral error (@ 0.114489) = 0.831381 deltaE
16:51:21,782 Average neutral error = 0.347778 deltaE
16:51:21,782 Failed to meet target 0.900000 delta E, got worst case 127.730613
16:51:21,782 Number of measurements taken = 9616:58:25,148 Profile check complete, errors: max. = 3.500277, avg. = 0.128113, RMS = 0.472710
16:58:25,147 Profile check complete, errors: max. = 3.500277, avg. = 0.128113, RMS = 0.472710
A LOT OF THANKS…
1 ) Your log file shows no errors, nothing is wrong with that DisplayCAL execution.
It is just taking measurements in an iterative process to calibrate garyscale.2) 99.999% the culprit is that old colorimeter. Get an i1d3 (i1DisplayPro/Munki Display)
- i1d1 filter aging
- not suitable for LED
- no future proof (no spectral corrections)
- very likely to have poor inter instrument agreement (even out of the box won’t be accurate)
DisplayCAL just works with data provided by colorimeter…
Calibrite Display Pro HL on Amazon
Disclosure: As an Amazon Associate I earn from qualifying purchases.Thank you very much
I’ve tried again with white point 6500 and native gamma (2.14 approx.) and the results have been somewhat better. I think the recorded data by the colorimeter are not real, since the result seems good, I think. The reference image that I normally use, “olé no moiré.tiff”, tells me. For my personal use it could be worth. Also the Photoshop densitometer and histogram never fail, so I should use them more often.
I don’t know if it’s a good idea to buy a new colorimeter to use on a monitor with a painful screen uniformity. I’ll see…
By the way, what does the log red text mean?
Profile check complete, errors: max. = 3.088587, avg. = 0.302474, RMS = 0.619989
Are values outside?
Thank you
Non synthetic profiles are created from measuremet data.
There are several types of profiles (matrix, LUT, with 1xTRC, with 3xTRC…).What you see in logs is a auto-verification of profile data against the set of measured data it was made from.
Average is average error, RMS is root mean square error and max. is the maximum.
The more complex the type of profile, the lower error statistics.Your results are typical for a matrix profile (hence fine if taken with a good colorimeter), BUT if your colorimeter sees very diferent color coordinates for R,G,B primaries than actual ones, that coordinates will be stored in your profile. That may cause color managed applications to clip colors before it was needed, etc. Calibration may OK good visually, and could be OK, but as long as profile information is not accurate you may experience unwanted results in color managed applications.
Try to create a black to R+G+B+W gradient in a color managed application and look for artifacts in the brighter end. It could be banding due to color management rounding errors but that is not what you should looking for, look for clipping-like artifacts.
If you see such problems, a inexpensive trick could be to generate an synthetic profile with EDID R, G & B coordinates and then add your current calibration, the one you see “good” visually, in its VCGT tag. IDNK is there is a fast way to do this in DisplayCAL or if you’ll need to use ArgyllCMS command line to append your current calibration (.cal) to such a synthetic ICM profile. Of course this is a trick, not a proper way to do this.
