- This topic has 10 replies, 3 voices, and was last updated 2 years, 3 months ago by
.
-
Posts
-
Hello Everyone,
I’ve recently upgraded to LG 27UP850-W monitor (27″, UHD, 95% DCI-P3). Mostly satisfied with the unit, but have some troubles with calibration. It supports HW calibration, but as with many LG’s that have this feature it seems to be useless: software is nice, but it doesn’t look like it’s applying correct meter corrections and while all delta’s and reports seem fine, the image just looks off/tinted. Probably no point writing to support about it. So I’ve decided to go my trusted Argyll/DisplaCAL/i1D3 route.
Monitor uses MV270QUM-N51 panel from BOE (400nits, 1200:1 contrast). I couldn’t find a proper datasheet for it, but most likely it’s WLED PFS-ish. I’ve calibrated using community-provided spectral correction (3.3nm made with iPro 2), but still it seems to look a bit ‘redish’ (though better that with LG HW solution). I’ve compared the community correction with existing 1nm PFS corrections (WLED PFS and 94% DCI-p3 WLED PFS/Panasonic) and general shape seems to be like the same on all of them, that leads me to belied that it is indeed a PFS-type backlight. The major difference in the graphs are Y scale values, community provided one tops at ~100/R, ~78/B, ~47/G while 1nm official ones top at 33/R and lower for G,B. Side-by-side comparison is attached. Should I try any of the official 1nm PFS corrections or the 3.3nm is the best that I can hope for?
Thanks!P.S. [rant]Really such a shame that LG (and others) make their HW calibration pretty useless with not supplying/applying proper corrections. I’m sure they wouldn’t go bankrupt purchasing a high-end spectrometer and bundling a proper correction for each monitor model/family… but seems like they just don’t care[/rant]
Calibrite Display Pro HL on Amazon
Disclosure: As an Amazon Associate I earn from qualifying purchases.Hello Everyone,
I’ve recently upgraded to LG 27UP850-W monitor (27″, UHD, 95% DCI-P3). Mostly satisfied with the unit, but have some troubles with calibration. It supports HW calibration, but as with many LG’s that have this feature it seems to be useless: software is nice, but it doesn’t look like it’s applying correct meter corrections and while all delta’s and reports seem fine, the image just looks off/tinted. Probably no point writing to support about it.
You can try to forge EDR, once you find which one it is using (INDK), like in CS2731 thread. These HW cal solutions for low cost monitos may have other errors even if measuring properly, like modifiying after calibration status while correcting brightness.
So I’ve decided to go my trusted Argyll/DisplaCAL/i1D3 route.
Monitor uses MV270QUM-N51 panel from BOE (400nits, 1200:1 contrast). I couldn’t find a proper datasheet for it, but most likely it’s WLED PFS-ish. I’ve calibrated using community-provided spectral correction (3.3nm made with iPro 2), but still it seems to look a bit ‘redish’ (though better that with LG HW solution). I’ve compared the community correction with existing 1nm PFS corrections (WLED PFS and 94% DCI-p3 WLED PFS/Panasonic) and general shape seems to be like the same on all of them, that leads me to belied that it is indeed a PFS-type backlight. The major difference in the graphs are Y scale values, community provided one tops at ~100/R, ~78/B, ~47/G while 1nm official ones top at 33/R and lower for G,B. Side-by-side comparison is attached. Should I try any of the official 1nm PFS corrections or the 3.3nm is the best that I can hope for?
Thanks!Use custom 3nm or Panasonic, the other one, PFS family is garbage and a mess (if you split all displays in files of 4 rows you’ll see most of them do not match your red)
P.S. [rant]Really such a shame that LG (and others) make their HW calibration pretty useless with not supplying/applying proper corrections. I’m sure they wouldn’t go bankrupt purchasing a high-end spectrometer and bundling a proper correction for each monitor model/family… but seems like they just don’t care[/rant]
Right now the list of these failures include:
-NEC PA271Q PA311D (both SV2 and Basiccolor. LightIllusion should work)
-ALL Benq in SW series, all of them
-All LGs using PFS (some older models with GB-LED wheer supported after 5 or more years when they did not sell them!)
-Dell UP2716D, UP2516D
-All PFS phoshor Eizos if you use an i1d3, athough it can be corrected in post calibration white point (same NEC). EIZO APAC has a custom EDR frok those, read Lift Gamm Gain Threads, just donwload corrected EDR and replace old.
-Samsung, all widegamut models
-Asus, almost all modelsThanks, Vincent! Already tried community/3nm, will try Panasonic now.
Would you say there’s a clear preference in my case (1nm Panasonic or consumer 3nm), or better to just try and compare?Pana is closer to the one measured for that display (although with lower resolution).
Anyway, those spiky SPDs can be a source of metameric failure, so you can use visual whitepoint (but use whitepoint relative, not absolute, LUT3D).
Chuck, the best way here is to rent i1Pro (i1Pro2 or even i1Pro 3 are better, but install unsigned driver for the last one from Argyll 2.3) and to make your own spectral correction for i1d3 (set HiRes mode for i1Pro!), then make matrix correction for your instance of i1d3, based on spectral one. This will almost perfectly match your i1d3 measurements with i1Pro. But this may not correct WP perfectly. Even i1Pro’s are not reference devices and white color temperature standard is developed for full-spectrum lights at the first, so you can meet metamerism effect when you look at the screen. Try to find out most stable white in xy white selector in Argyll (you may use display menu RGB bars too). Check standard images (like portraits) afterwards, sometimes you should move white in a bit wrong way and remake all work. Or you may slightly move RGB in display menu after profiling.
Info on LG calibration program: screenshot in manual shows color temperature selection only, there’s no white selection by xy coordinates. If so, it looks like unusable, but I haven’t seen the latest upgrade.
Add on: find out best display brightness for your tasks and set RGB bars to neutral position (50/50/50 for LG, yeah?), then make colorimeter corrections.
Thanks for the suggestions. The community 3nm spectral correction (the one i’m using atm) was done using i1Pro 2. I can rent the same spectro (though a bit pricey… 90EUR/day) to get matrix correction, but i wonder if it will really solve the problem… This used to be so much easier with plain sRGB WLEDs (at least in my experience). Do you know if quantum dot backlights are easier to handle wb and correction-wise (spds seem very smooth and even) ?
just an update: for now ended up using WLED PFS DCI-P3 94% (Panasonic VX) correction, i think it gives me a bit more neutral wb, when doing verification the difference is marginal, calibrating with Panasonic spectro + verification with custom 3nm (and vice versa) result in “PASS” (with slightly higher delta in WB measurement, but still in green zone). Also for some reason 3nm correction was switching the instrument mode to ‘Refresh’ instead of “LCD (generic)” and when I set the mode back to LCD it sets the selected correction to ‘none’. Would the Refresh mode affect the calibration in any noticeable manner?
One more questions: Is the following (see attachments) color temp and RGB grey balance behaviour inherent to the WLED backlight in general or does it happen for another reason (instruments, etc)? My previous LG 27UD68 did similar things in the low luminance portion of the graphs.Chuck, I’m not ready to give you thorough answer on quantum dot LED LCD, in theory it looks like you say. But all in all, color temperature standard is made for full spectrum light sources, but all display lights consist of several spectral peaks. Sometimes I meet PFS-like displays, that show pure white tint and correct portrait colors close to D65 dE=0 measured with my i1Paint (i1Pro 3 Plus device), other PFS show rose tint and perform much better with dE=0 measured by my i1Pro 2, that has little spectral shift to i1Paint and my ColorMunki Design and dE~4.5 color difference. However, my i1Paint shows blackboby temperature close to daylight temperature by my i1Pro 2, so I think, that i1Pro and i1Pro 2 devices may had different calibration standard (note, i1Paint uses 7-peaks full-spectrum LED that emulates D50). My BenQ PV270 spectrum is similar to Samsung quantum dot LED’s (it is quite compressed towards blue bound) and I use i1Paint dE=0 for it. More respectable and well-made Eizo CS2731 uses PFS. I think, panel quality and capacity are more valuable for manufacturers rather than physical model perfection.
Your attachements show correct plots. Yes, LCD displays never get ideal gamma value in deep dark and highlights. I suppose, cause of gamma emulation. Gamma 2.2+/-… was natural function of CRTs (I have never calibrated or tested CRT, though) and their electronics, but is synthetic for digital panels. Not an issue if test is passed well. I don’t know, what ICC stage corrects gamma difference. …Vincent, could you help here?
More respectable and well-made Eizo CS2731 uses PFS. I think, panel quality and capacity are more valuable for manufacturers rather than physical model perfection.
while i agree in general, EIZO’s CS line and other pro models are sometimes also hit-and-miss. I’ve had two CS2740 last month (EIZO refused to tell me what type of backlight it’s using and LCD panel model/SKU). uniformity was better of course than consumer models, but there’s something inherent to either whole batch or this type of panel. bottom was ‘colder’ and top ‘warmer’ both visibly and measurably, and this for 3x the price of my current display. also contrast was 680:1 on 1st unit and 720:1 on the 2nd due to uniformity compensation. Then there was a good deal on Dell UP2720Q, but the unit had 4-5 distinct backlight bleed spots at the bottom and the shop refused to exchange it (only return/refund) due to price jump of +400EUR after purchase). So I decided to sit this round out and settle on 95% DCI-P3 and 89% AdobeRGB with relatively good 1100:1 contrast and not loosing Freesync as a bonus (which would have been a bummer on pro models) for fraction of a cost.
