- This topic has 77 replies, 6 voices, and was last updated 2 years, 11 months ago by
.
-
Posts
-
Hello,
I just bought a BenQ PD2720U and it is advertised as 100% SRGB Coverage, 99% Adobe RGB coverage and 96% P3 Coverage. I’m going insane trying to figure out why I’m not reaching those numbers.
I first started with my 5 year old Spyder5 and the best I could get was 98.5% SRGB, 93.1% Adobe RGB, 81.3% DCI-P3. So I emailed BenQ and their response was just basically try a different monitor cable or I probably have a wrong setting somewhere, but they can’t help because they only do calibration support for their SW line of monitors. Tried a different cable and a million settings it seams and no improvement. Fell down the google rabbit hole and came across a few posts that say Spyders can be inaccurate sometimes from the factory and their accuracy fades as they age.
So I bought an i1 Display Pro as it seems to be everyone’s favourite from a lot of the posts I’ve come across. Now I am getting 99.7% SRGB, 99.5% AdobeRGB and 91% DCI-P3. So the Spyder was at least partly to blame. Now half a percent for SRGB and Adobe RGB is negligible but 5% for P3 seems significant.
The settings that seem to give best results is just using the SRGB Preset, the correction to “Spectral: LCD RG Phosphor LED IPS (BenQ PD2720U (i1 Pro)), and I’ve set a custom white level of 150 based on measured ambient. The rest is default for the preset. Monitor is set to “user” colour mode (Using any of the monitors built in colour mode presets for DCI-P3, AdobeRGB, etc got much worse results) with gamma set to 2.2 and custom white balance and brightness based on DisplayCal’s recommendation. (Ended up being 38 brightness and 100R 95G 93B). I’ve also tried changing calibration speed to medium and test chart to large. Results were the same. I’ve also tried profiling without any calibration at default monitor settings and got 99.7% SRGB, 98.1% AdobeRGB, and 91.9% DCI-P3.
So the questions are:
– Is there some setting I should be changing in DisplayCal to get the advertised 96% P3?
– Or does BenQ lie about their specs and I should return it. There are lots of monitors that advertise 90% P3 coverage for a lot cheaper, if that’s all I’m going to get.
– Or is it just expected for monitors to be at least 5% below spec?
– Or is the i1 Display Pro just still not accurate enough and I’ll never get advertised results without a super expensive calibrator?
– Or Am I obsessing over nothing and should just accept it as is and move on? haha 😛Windows 10 2004 build 19041.746
GeForce RTX 2070 SUPER 461.09 driver
If those affect it any.Thanks,
-MarcelCalibrite Display Pro HL on Amazon
Disclosure: As an Amazon Associate I earn from qualifying purchases.GBLEDs are not able to reach P3 red, hence up to 93% is expected. If you wanted near full P3 you should move to WLED PFS or some QLED.
Anyway, the main concern in these poor quality lowcost widegamut monitor is color uniformity issues. If by (unlikely) chance it is good, keep it.
If you are planing to create P3 content with it you are going to use color managed software (ICC or LUT3D) and content that reaches full saturation P3 red is not that common, it is vry unlikely to happen (and if you were commited to even higher quality standards you won’t choose benq). All in gamut colors (P3D65 colors inside your display colrospace) will be rendered in an accurate way (unless you have some issue in your image pipeline becaus other HW), so 91% is OK if uniformity is good.PS: we can argue about the importance of AdobeRGB coverage, since AdobeRGB was designed to cover printable cyan, but instead of be obsessed with 3% range for people with 97% coverage or such (mostly near non printable 255 green), best to check if AdobeRGB cyan is within reach (dE) to your display capabilities. It’s about content after all.
GBLEDs are not able to reach P3 red, hence up to 93% is expected.
Correction:
Measured Dell 2016 PremierColor GB-r LED completely covers P3, inc red. See attached annotated output from ICC Profile Info for native-mode UP2516D.
I
GBLEDs are not able to reach P3 red, hence up to 93% is expected.
Correction:
Measured Dell 2016 PremierColor GB-r LED completely covers P3, inc red. See attached annotated output from ICC Profile Info for native-mode UP2516D.
It is not a GB-LED. It is some early stage WLED PFS. Graeme measured spectral powrr distribution one of these 25″ or 27″. Maybe there are CCSS correction for two in display cal database.
NOT GB-LED 100% sure
For example the 3 first ones, 3nm, PFS variant with their disctinctive red spikes:
This is a GB-LED (PD2720U, the one he is using)
This is another GB-LED (U2413, displaycal bundled sample, 1nm)
This is another GB-LED (PA242W from NEC/basiccolor/others soft)
And we can go on night and day and your UP2516D will not be a GB-LED
I can’t dispute that there is a similarity in the spectral plots…
However, Dell DUCCS (i1 Profiler) reports GB-LED. See attached screenshots (A) shows display readout, (B) shows “Selected calibration matrix GB-LED”
I’m pretty sure the UP2516D and the UP2517D are the same device except for screen size.
The following web resources report GB-r:
Quote:
//Looking past blue diodes
Although it can be nice from some perspectives to reach sRGB or a bit beyond, as it allows slightly greater vibrancy, you would really want to reach the next ‘standard’ of gamut for colour-critical work and to really unlock vibrancy potential. To achieve this initially, LG Display took the approach of using a modified type of WLED backlight called GB-LED (also known as GB-R LED or GB-r LED). …
There are a number of monitors now available that use GB-LED backlights, including the Dell UP2716D whose colour gamut is shown above (red triangle) and compared with sRGB (green triangle) and Adobe RGB (purple triangle).////The monitor uses a 27” WQHD panel from LG Display. This makes use of AH-IPS technology, with support for 10-bit colour by means of 8-bits per channel + FRC dithering … Backlight: GB-LED (Green Blue Light Emitting Diode)//
https://www.displaydb.com/monitor/dell-ultrasharp-up2516d
//Backlight: GB-r LED//
https://www.displayspecifications.com/en/model/da704af
//Backlight: GB-r LED//
Yes, all these sources — including i1 Profiler commissioned by Dell expressly for this model of display — COULD BE WRONG. Or maybe not.
GBLEDs are not able to reach P3 red, hence up to 93% is expected. If you wanted near full P3 you should move to WLED PFS or some QLED.
Anyway, the main concern in these poor quality lowcost widegamut monitor is color uniformity issues. If by (unlikely) chance it is good, keep it.
If you are planing to create P3 content with it you are going to use color managed software (ICC or LUT3D) and content that reaches full saturation P3 red is not that common, it is vry unlikely to happen (and if you were commited to even higher quality standards you won’t choose benq). All in gamut colors (P3D65 colors inside your display colrospace) will be rendered in an accurate way (unless you have some issue in your image pipeline becaus other HW), so 91% is OK if uniformity is good.PS: we can argue about the importance of AdobeRGB coverage, since AdobeRGB was designed to cover printable cyan, but instead of be obsessed with 3% range for people with 97% coverage or such (mostly near non printable 255 green), best to check if AdobeRGB cyan is within reach (dE) to your display capabilities. It’s about content after all.
If the tech can only do 93% why do they advertise it as 96%?
Low cost is subjective haha. My last monitor was $700 CAD and the one b4 that $300 so jumping to $1300 is high cost to me. 😛
Uniformity is another rabbit hole I haven’t even begun to look at yet. I don’t really know what would be considered good. I’ve attached the report. Seems like it fails ISO but passes Average.
I don’t do colour critical work and am not a colourist by any definition. I do Videography and Video Editing for a living and the end destination is usually just web. I do photography and travel videos as a hobby but never do prints and again destination is just web.
I shoot 4k video in slog2 sgamut3.cine and photos in AdobeRGB. My other monitor only covered srgb, so was looking for a reasonably accurate 4k monitor, with the widest gamut possible for around the $1K CAD range, just to see more of what I was actually capturing, before it gets compressed down to SRGB. So already went over budget with this monitor. Is there a different one in the same price range you would recommend instead?
I can’t dispute that there is a similarity in the spectral plots…
However, Dell DUCCS (i1 Profiler) reports GB-LED. See attached screenshots (A) shows display readout, (B) shows “Selected calibration matrix GB-LED”
I’m pretty sure the UP2516D and the UP2517D are the same device except for screen size.
The following web resources report GB-r:
a) Xrite HAS NOT an EDR for that backlight. Hence, Dell does not have it, they are choosing the lesser evil.
It’s the same problem as Benq. Benq has not and never had QLED or WLED PFS EDR correction. SO they correct with RGB LED which is not equal (by far in red) but it is the closest they have.
Nec has the same problem with their WLED PFS AdoberGB: PA271Q and PA311D. They correct colorimeter readings with the WRONG correction GB-LED (PA242W plotted above). Xrite has or has licensed an EDR for those backlights, for HP, so they are bundled with DIsplayCAL as “HP Z24x G2”. Newer Eizo CGs, those necs and even benq SW240 should use that correctio when working with displaycal (unless better taylor made correction is available)b) DUCCS latest version was updated to inlude WLED PFS 95% P3 (no adobeRGB green). Green is not yours, but red looks a little like yours.
It is Dells fault not having the proper correction for their displays. Dell using one correction or another is not a proof of some backlight, just a proof of how they care about tehir customers and their devices accuracy.
Error due to wrong EDR/CCSS correction depends on how different colorimeter spectral sensivities are from STD OBS in wavelegths where “actual SPD” and “CCSS/EDR” drift away. On an ideal i1d3 error is none (but they are not so good).
A common well behaved i1d3 usually is very close to that in reds so choosing RGBLED, GBLED or AdobeRGB WLED PFS is expected to produce a low error. Maybe <4de in white? But i1d3 was designed to avoid those uncertainties… so it’s display vendor fault not caring about customers.Quote:
//Looking past blue diodes
Although it can be nice from some perspectives to reach sRGB or a bit beyond, as it allows slightly greater vibrancy, you would really want to reach the next ‘standard’ of gamut for colour-critical work and to really unlock vibrancy potential. To achieve this initially, LG Display took the approach of using a modified type of WLED backlight called GB-LED (also known as GB-R LED or GB-r LED). …
There are a number of monitors now available that use GB-LED backlights, including the Dell UP2716D whose colour gamut is shown above (red triangle) and compared with sRGB (green triangle) and Adobe RGB (purple triangle).////The monitor uses a 27” WQHD panel from LG Display. This makes use of AH-IPS technology, with support for 10-bit colour by means of 8-bits per channel + FRC dithering … Backlight: GB-LED (Green Blue Light Emitting Diode)//
https://www.displaydb.com/monitor/dell-ultrasharp-up2516d
//Backlight: GB-r LED//
https://www.displayspecifications.com/en/model/da704af
//Backlight: GB-r LED//
Yes, all these sources — including i1 Profiler commissioned by Dell expressly for this model of display — COULD BE WRONG. Or maybe not.
Yet the suspect fingerprints matches the weapon fingerprints. So whatever they say is meaningless.
i1Profiler has not that backlight (yours), nor QLED (SW2700PT), nor AdobeRGB WLED PFS (newer Eizo CG, necs,, etc).
They only have licensed for general public WLED PFS 95% P3, without AdoberGB green (greenn channel is a little different).
Blame Xrite (I’d do if I had not DisplayCAL) for this. Same for Dell. They really do not care.GBLEDs are not able to reach P3 red, hence up to 93% is expected. If you wanted near full P3 you should move to WLED PFS or some QLED.
Anyway, the main concern in these poor quality lowcost widegamut monitor is color uniformity issues. If by (unlikely) chance it is good, keep it.
If you are planing to create P3 content with it you are going to use color managed software (ICC or LUT3D) and content that reaches full saturation P3 red is not that common, it is vry unlikely to happen (and if you were commited to even higher quality standards you won’t choose benq). All in gamut colors (P3D65 colors inside your display colrospace) will be rendered in an accurate way (unless you have some issue in your image pipeline becaus other HW), so 91% is OK if uniformity is good.PS: we can argue about the importance of AdobeRGB coverage, since AdobeRGB was designed to cover printable cyan, but instead of be obsessed with 3% range for people with 97% coverage or such (mostly near non printable 255 green), best to check if AdobeRGB cyan is within reach (dE) to your display capabilities. It’s about content after all.
If the tech can only do 93% why do they advertise it as 96%?
Low cost is subjective haha. My last monitor was $700 CAD and the one b4 that $300 so jumping to $1300 is high cost to me.
Uniformity is another rabbit hole I haven’t even begun to look at yet. I don’t really know what would be considered good. I’ve attached the report. Seems like it fails ISO but passes Average.
I don’t do colour critical work and am not a colourist by any definition. I do Videography and Video Editing for a living and the end destination is usually just web. I do photography and travel videos as a hobby but never do prints and again destination is just web.
I shoot 4k video in slog2 sgamut3.cine and photos in AdobeRGB. My other monitor only covered srgb, so was looking for a reasonably accurate 4k monitor, with the widest gamut possible for around the $1K CAD range, just to see more of what I was actually capturing, before it gets compressed down to SRGB. So already went over budget with this monitor. Is there a different one in the same price range you would recommend instead?
GBLED coverage? IDNK. But 91-93 is expected whith SPD I plotted (from CCSS correction). You can’t get more.
Uniformity seems pretty good if we exclude that lower right side. If contrast is OK while using the same setup as you used to measure uniformity should be no issues caused by monitor.
27″ and UHD alternatives … CS2740 but contrat may be lower at D65, but you get HW cal, a reliable one (not like benq SW line). Expected same coverage.
Dell had another 27″ UHD but IDNK backlight. If it is like WLED PFS AdobeRGB you’ll get almost full P3 & AdobeRGB + HW cal + a lillte more contrast 1100-1300:1 range… but Dell or Benq or LG HW cal solutions are not exactly reliable as you may find by fast googling, I mean software.
Those 2 go to 1500 euro or more.Uniformity seems OK if contrast is OK… I would not care, there is not su much real flife content in P3 outside your display colospace.
He’s a google result with a bit more technical exposition:
https://displaydaily.com/article/display-daily/there-s-more-than-one-way-to-make-a-wcg-lcd
My sense is that my disagreement comes down to vernacular:
In Argyll CMS CCxx lexicon, GB vs PFS comes down to viewing categories of displays based on generic backlight terms, which of course happen to follow industry vernacular. But display product makers are not going to spend customer eye-share on gory technical details.
And in the case — 2016-era Dell Premier Color sourced from LG — the LED design style is GB-r,, where the phosphor applied as “r” is PFS.
In this case, we may want a GB view of green and a PFS view of red.
There really is nothing to hang a hat on here, because Argyll / DisplayCal’s taxonomy is just a bunch of loosely attached comments to specific corrections from individual users adapting to their gear.
At the same time, the industry has no desire to expose these details to customers, and the corrections are secret-sauce. Secret in the sense that rather than referring to formal sources for standards, DisplayCal itself has a provision where in can look within the bundles for other vendors calibration products for corrections in the hopes that these might be of aid to users.
In this sense our arguments can be apples/apples or apples/oranges depending on how you look at it: I don’t think the trade lingo for these Dell’s LG panels is wrong, and I don’t feel wrong for using it. OTOH with Vincent’s help, we can see that there are finer points.
I notice that while Dell maybe “doesn’t care”, that the display alignment produced by DUCCS (i1 Profiler) may choose a compromise correction based on criteria other than utmost accuracy, the alignment result looks as good to me as DisplayCal. To repeat, for me DUCCS and DisplayCal produce equivalent quality alignments. But my bias to Argyll CMS / DisplayCal hits home here, because I set white by hand according to the CCSS for i1 Pro 2 UP2516D and run DUCCS on native white. It’s my personal style that I don’t rely out-of-box white for anything because white alignment has such a profound effect on my subjective regard for performance. This is just a personal tic.
I admit that I have -not- exhaustively compared results between DUCCS and and DisplayCal, and DisplayCal may be able to handle some situations better. I much prefer DisplayCal. I appreciate DUCCS ability to load display LUTs with a correction. Yet I don’t care about this DUCCS feature either, because loading the CAL slots locks out the brightness control! Garrr. But as I suffer zero trouble with banding, and I run a fully color managed system, and the backlight level is a critical control for my workspace, I forego DUCCS. Yet I praise Dell for creating it and think it’s valuable and works well. They were on a good track with DUCCS, and it can be very useful for some situations. It can create a superb canned-alignment for uses who want to set their display to a reference space for use by unmanaged apps.
I fired up latest version of X-Rite i1 Profiler and it now lets me choose the backlight correction. This is a recent addition for Mac version. Interestingly, for these UP2516D, it chooses White LED as default. So this adds credence to Vincent’s view of Dell making a compromise in DUCCS.
I stand by my “GB-r correction” post but admit that in sense I may be wrong because I am regarding backlight LED category in a different way than Vincent, and the conversation was not mine. Maybe I am completely wrong. But I think Vincent is generalizing from a specific. The finer point is that GB-r has an overlap with PFS.
As to beating others over the head for how wrong they are regarding one’s 100%-certainty over a completely contrived category of DisplayCal panel backlight types, I see such beatings are just a matter of style.
GBLED coverage? IDNK. But 91-93 is expected whith SPD I plotted (from CCSS correction). You can’t get more.
Uniformity seems pretty good if we exclude that lower right side. If contrast is OK while using the same setup as you used to measure uniformity should be no issues caused by monitor.
27″ and UHD alternatives … CS2740 but contrat may be lower at D65, but you get HW cal, a reliable one (not like benq SW line). Expected same coverage.
Dell had another 27″ UHD but IDNK backlight. If it is like WLED PFS AdobeRGB you’ll get almost full P3 & AdobeRGB + HW cal + a lillte more contrast 1100-1300:1 range… but Dell or Benq or LG HW cal solutions are not exactly reliable as you may find by fast googling, I mean software.
Those 2 go to 1500 euro or more.Uniformity seems OK if contrast is OK… I would not care, there is not su much real flife content in P3 outside your display colospace.
The contrast looks fine as far as I can tell. Doing a verification test, DisplayCal is saying contrast ratio is a little higher then spec at 1100:1. Report attached. As far as uniformity goes, the most important part of whatever I’m shooting will always be within the middle 9 boxes, the edges would just be peripheral stuff, so guess I’m ok there.
HW Cal would be good to have and those look like nice monitors. An extra $1k though is out of my budget for this upgrade cycle. 😛 Perhaps in a few years when I upgrade again. 🙂
And in the case — 2016-era Dell Premier Color sourced from LG — the LED design style is GB-r,, where the phosphor applied as “r” is PFS.
In this case, we may want a GB view of green and a PFS view of red.
No, no need for that. WLED PFS produce the same green… in its AdobeRGB flavor (CG279x etc). B/G height ratio in SPD is just whitepoint.
I notice that while Dell maybe “doesn’t care”, that the display alignment produced by DUCCS (i1 Profiler) may choose a compromise correction based on criteria other than utmost accuracy, the alignment result looks as good to me as DisplayCal. To repeat, for me DUCCS and DisplayCal produce equivalent quality alignments.
No, that is a false statement.
-i1Profiler has no corrections for new baklight because they do not care at all
-even true GB-LED models (2013 models), they use a mix of several backlights where only one is a GB-LED. NEC chose to have separated EDR just for GB-LED.
-DUCCS is a failure because even using the proper correcion (for GB-LED models) for i1displaypro it fixes white at default brightness (actually not just WP, but full lut including gamut simulation), then lowers brigtness to desired target. Since white drifts with OSD brightness contrast going up/down users end with out of daylight curve whites. This is a common issue which lots of users are suffering, just google it. People ending with 6dE from daylight locus…
The only way to correct it is full GPU whitepoint correction in DisplayCAL on top of CALx, or messing arrounf with alt CIE xy white targets trying to predict offset result.-matching readings within i1Profiler variants (“profile whitepoint” in reports) and DisplayCAL (“measured wp vs profile WP” while validating DUCCS profile) is because YOUR particular i1d3 matches std obs to a high degree where the two different SPD, the mix of GBled and others in “RG_phoshor” and the UP2x16D 3nm CCSS sift from each other.
This is the reason of close matching. Not Xrite compromise to quality in their software products.
Same applies to Benq choosing RGB LED correction to correct thery GB-LED, QLED and their AdobeRGB WLED PFS.
They do not supply proper EDRs for their display backlights because they do not care at all, or that they do not want to expend whatever royalties are needed to use/create a new EDR.I admit that I have -not- exhaustively compared results between DUCCS and and DisplayCal, and DisplayCal may be able to handle some situations better. I much prefer DisplayCal. I appreciate DUCCS ability to load display LUTs with a correction. Yet I don’t care about this DUCCS feature either, because loading the CAL slots locks out the brightness control! Garrr.
This seems to be a “feature” of 2016 models. Maybe you are using DDM to change it instead of manual OSD controls. Locking in DDM is true, but for other generations OSD brightness is unlocked.
A good thing of 2016 gen is that they should be included in dell SDK supported models, so there is a way to upload LUTs to display whithout DUCCS…. although you have to code it in c++.
I fired up latest version of X-Rite i1 Profiler and it now lets me choose the backlight correction. This is a recent addition for Mac version. Interestingly, for these UP2516D, it chooses White LED as default. So this adds credence to Vincent’s view of Dell making a compromise in DUCCS.
It’s default value for all displays unless they were included in some XML configuration file. It is not guessing your particular display.
It’s user responsibility to know which backlight hs to apply in i1Profiler. Also user doc is very outdated, not long ago they still recomended using RG_phosphor for P3 mac displays, with thanks to CCSS provided by users we know that are WLED PFS. Plot them.I stand by my “GB-r correction” post but admit that in sense I may be wrong because I am regarding backlight LED category in a different way than Vincent, and the conversation was not mine. Maybe I am completely wrong. But I think Vincent is generalizing from a specific. The finer point is that GB-r has an overlap with PFS.
There are several PFS flavor, all bundled with displaycal. Plot them. Their main differences are explained in other threads
-PFS 95% P3. Common cheap model
-Apple P3. Same as above but green is slighty different to cover exactly P3, nothing more nothing less
-“photo monitor” PFS, with AdobeRGB green
They key difference is wavelength at peak and width of green spikesPlot SPD with displaycal. It’s very educational about different SPD produce different primaries.
You can even (with some spare time) multiply these SPD data per channel at 1nm with ArgyllCMS -bar 1nm std observer fucntions, then integrate (sum) in visible wavelengths interval. This way you can get CIE XYZ coordinates.
You can plot in a 2D graph fuction multiplication only, to see actual contribution of SPD to what you see, since once multiplied they are “human weigthed”.
Also you can plot 2D accumulative sum (accumulative value of CIE XYZ) to check how actual value is accumulated.GBLED coverage? IDNK. But 91-93 is expected whith SPD I plotted (from CCSS correction). You can’t get more.
Uniformity seems pretty good if we exclude that lower right side. If contrast is OK while using the same setup as you used to measure uniformity should be no issues caused by monitor.
27″ and UHD alternatives … CS2740 but contrat may be lower at D65, but you get HW cal, a reliable one (not like benq SW line). Expected same coverage.
Dell had another 27″ UHD but IDNK backlight. If it is like WLED PFS AdobeRGB you’ll get almost full P3 & AdobeRGB + HW cal + a lillte more contrast 1100-1300:1 range… but Dell or Benq or LG HW cal solutions are not exactly reliable as you may find by fast googling, I mean software.
Those 2 go to 1500 euro or more.Uniformity seems OK if contrast is OK… I would not care, there is not su much real flife content in P3 outside your display colospace.
The contrast looks fine as far as I can tell. Doing a verification test, DisplayCal is saying contrast ratio is a little higher then spec at 1100:1. Report attached. As far as uniformity goes, the most important part of whatever I’m shooting will always be within the middle 9 boxes, the edges would just be peripheral stuff, so guess I’m ok there.
HW Cal would be good to have and those look like nice monitors. An extra $1k though is out of my budget for this upgrade cycle.
Perhaps in a few years when I upgrade again. 
It looks OK but this is not profile verification. This is how good your calibrated & profiled display is while showing sRGB images in Photoshop, that’s what profile simulation do while using custom display profile as default display profile.
Also “sRGB F-S XYZLUT+MTX”, people usually want 2.2 as calibration target (rather than sRGB TRC in a 1000:1 display), and if you experience in this display or another some color issues in grey after calibration, don’t choose “fast” calibration speed. But their are just hints if typical configuration, not saying that those are wrong.
It looks OK but this is not profile verification. This is how good your calibrated & profiled display is while showing sRGB images in Photoshop, that’s what profile simulation do while using custom display profile as default display profile.
I’m confused, so what should I be doing to check contrast then?
