BenQ Palette Master Ultimate(PMU)

[nick234]

Hi.

Have you seen the new BenQ hardware calibration software?

I have a question, does anyone know if the spectral calibration files have been improved?

and what is your overall opinion about this program?

Regards

[Vincent]

Hi.

Have you seen the new BenQ hardware calibration software?

I have a question, does anyone know if the spectral calibration files have been improved?

Now supports WLED PFS, although with the very generic “PFS_Phosphor_Family_31Jan17”, not one specific for WLED PFS with P3 red and AdobeRGB green.
Split display samples from PFS_Phosphor_Family_31Jan17.ccss in DisplayCAL and you’ll see( n samples x WRGB spectral series).

Older QLED models are not suported, just new ones.

and what is your overall opinion about this program?

Regards

Clone of CN7 which is and upgarde, but due to very low QC on Benq SW series hardware still has the same issues regarding grey range as Dell, LG and other low Q HW calibration software.
Anyway, it is a huge leap forward from PME.

[nick234]

But generally “PFS_Phosphor_Family_31Jan17”, fares much better in this new software?.

You wrote that gray is still bad, but is it even a little better?

It has a gamma linearization function, then there are more gray measurements – it seems that this soft better interferes with the hardware of the monitor. And we can no longer talk about cheating hardware calibration as was the case with PME.

Thanks for the answer Vincent

[Vincent]

But generally “PFS_Phosphor_Family_31Jan17”, fares much better in this new software?.

Closer to actual SW models. Plot them by yourselft. For example take a few samples from DisplayCAL colorimeter database:

https://colorimetercorrections.displaycal.net/

and compare them to PFS_Phosphor_Family_31Jan17 and to RGLED (older EDR)

Now it’s a better match, but still lacking of correction for older QLED models. Xrite software has not them, maybe ASUS has them.

You wrote that gray is still bad, but is it even a little better?

From ambassador videos on youtube, ar default settings same bad range in half of them, > 1.5 in a* using PMU measurements… but their software does not evaluate range, hence goes unnoticed for unexperienced users.
Mostly a problem for B&W and UI design.
Same happens with Dell’s DUCCS and other software. If QC cannot grant you perfect neutral white from factory, it may not be able to correct it

It has a gamma linearization function, then there are more gray measurements – it seems that this soft better interferes with the hardware of the monitor. And we can no longer talk about cheating hardware calibration as was the case with PME.

Thanks for the answer Vincent

Yes, “enhanced gamma calibration” option, but no ambasador has tested it.

As explained before the worse QC is (bad grey color uncalibrated) the more patches are needed in grayscale to correct it.
From what we have seen here “medium” speed in Displaycal can correct most displays (48 patches) while fast (24) may leave erros in some bad behaved displays.
PMU without that new option takes… 18 (grey, combined?) and range is bad. DUCCS took 9×3 (on each channel uincalibrated gamma ramp) and ranges in 1.4-1.6 were common.
That new option should go to ~40 (on W ramp, or for each ramp) to be able to correct most of their displays. We’ll see…

Also after calibration contrast was very bad, even on ambassador models 300-700:1, but I doubt thi is PME/PMU fault. It’s bad QC, bad panels, were agresive UC is applied to pass a minimum quality. That destroys contrast.
OTOH Coloresges have perfect dC color uniformity out of the bot at native contrast and most of SW line prices are close to CS models… so I cannot see these Benq as alternative. Why woudl somebody buy these half baked models at the same price as a CS?

[nick234]

You’re right, higher models seem to make little sense – it’s better to buy NEC or Eizo.

While the SW240 seems to be an interesting alternative – it’s cheap, has good uniformity, hardware calibration for the price, and a wide gamut

I have the SW240 and I’m most interested in how it handles PMU, but unfortunately there are few tests on the Internet of new software with this monitor model.

[Vincent]

I have the SW240 and I’m most interested in how it handles PMU, but unfortunately there are few tests on the Internet of new software with this monitor model.

Easy, SW240 is a WLED PFS AdobeRGB flavor, there are samples in community database, so:

1) PME -> calibrate to a known target, whatever you want.
Lets say Native gamut , D65, native contrast, 120nit. If PME does not allow SW240 to use native gamut choose P3 or AdobeRGB, grey range is worse on native gamut and you want to test it, so do not choose sRGB/Rec709 as target for these tests.
2) Validate with DisplayCAL (use “wrong” RGBLED correction to use same “rules” as PME). Check contrast & grey range (upper combo in HTML report, RGB + gray  balance)
3) Validate with DisplayCAL using PFS_Phosphor_Family_31Jan17. Check if WP is off vs the other.

If your SW240 has good enough uncalibrated grey and if by chance your i1d3 is well behaved, close to std obs where RGB-LED and WLED PFS backlight drift from each othern, then most of PMU measurement upgrades are not meant to you.
Of couse you get the new features like fast ICC change at the same time you change OSD due to tray app. It may be worth installing.

Lets say that your SW240 has bad grey range uncalibrated and when calibrated with PME still has a high grey range, like >1.5 or so. Or let’s say that measurements of WP using those two CCSS/EDR are very diferent in a* axis.
If that happens PMU is an upgrade menat for you, so
4) Uninstall PME
5 ) Install PMU
6) Calibrate to same target as above using PMU
7) Validate with DisplayCAL using PFS_Phosphor_Family_31Jan17. Check WP (CDT & whiteness), contrast and grey range.

If you get an improvement, congratulations. But if grey range is bad:
8) Calibrate to same target as above using PMU, but use “enhanced gamma calibration” checkbox.
9) Validate with DisplayCAL using PFS_Phosphor_Family_31Jan17. Check WP (CDT & whiteness), contrast and grey range.

All displaycal validations must be run without simulation profile, just validate PME/PMU profiles.

• This reply was modified 9 months, 3 weeks ago by Vincent.
• This reply was modified 9 months, 3 weeks ago by Vincent.
• This reply was modified 9 months, 3 weeks ago by Vincent.

Calibrite Display Pro HL on Amazon  
Disclosure: As an Amazon Associate I earn from qualifying purchases.

[nick234]

Thanks, I will check it the way you wrote.

I have one more question, I deduced from your statements that a lot depends on what kind of monitor you get. Doesn’t hardware calibration make every monitor better?

Hardware calibration should correct every single monitor – regardless of how good a copy we got. I think and write well?

[Vincent]

Thanks, I will check it the way you wrote.

I have one more question, I deduced from your statements that a lot depends on what kind of monitor you get. Doesn’t hardware calibration make every monitor better?

Hardware calibration should correct every single monitor – regardless of how good a copy we got. I think and write well?

If you do a calibration (GPU) with the same amount of grey patches… you’ll get the same/close grey range value. For example DisplayCAL on “very fast ” setting (12 patches)

This is the point:
Current HW cal software from vendors take a very small number of patches, hence only vendros with very good QC get very good calibration, regarding grey range of course.
OTOH vendor GPU calibration software like i1Profiler takes about 20-21 (I do not remember) and 3rd party software like ArgylCMS allows you to use 48-96 => they can correct very bad uncalibrated grey in displays.

IDNK why that obsession on these vendors HW cal software to optimize for speed while they cannot guarantee a good uncalibrated grey (QC). Low QC implies using more pqtches to calibrate grey and customers would like it.
Let’s hope that PMU new option, “enhanced gamma calibration” helps to correct those bad behaved monitos…. but in the end it depends on patch number increase. Going from 18 to… 25 would mean nothing.
Test it and share patch count with that new option enabled.

• This reply was modified 9 months, 3 weeks ago by Vincent.
• This reply was modified 9 months, 3 weeks ago by Vincent.
• This reply was modified 9 months, 3 weeks ago by Vincent.

[nick234]

I don’t have a monitor with me now. I can test in a few days

A while ago on the previous computer I did a quick test, from what I remember, the graph of gamma dots was much better. The RGB balance left a bit to be desired

Both PME and PMU white balance does not agree with DisplayCal. BenQ’s programs say it’s perfectly 6500K, while DisplayCal says ~6350K. Perhaps the difference is due to different EDR files? the white balance “by eye” seems fine…

The only 100% working programs are Eizo CN7 and NEC SVII. The BenQ program has made progress in relation to the old soft, but it is still far from these two legendary giants. Finally, after so many years, there is an ICC profile synchronization function. However, there is still no possibility of doing more calibrations and then freely loading them into the LUT of the monitor. 3 slots is very little. And each new calibration takes a lot of time, you know, than calibrating, for example, 10 presets and then selecting which one we put into the monitor.

My english is not very very perfect, hope you understand something :))

[Vincent]

I don’t have a monitor with me now. I can test in a few days

A while ago on the previous computer I did a quick test, from what I remember, the graph of gamma dots was much better. The RGB balance left a bit to be desired

Easier to inspect as a* b* columns across all grey ramp than RGB, but YMMV.

Both PME and PMU white balance does not agree with DisplayCal. BenQ’s programs say it’s perfectly 6500K, while DisplayCal says ~6350K. Perhaps the difference is due to different EDR files?

There are two sources of error:
-EDR, device measures “wrong”
-Calibration app failes to aim properly.

On option 2 and if you calibrated with old PME, just run a verification in DisplayCAL with old PME correction, RGB LED.
If you measure with the same EDR, even if it is wrong, you can inspect if that calibration app failed to “aim” properly at target.

the white balance “by eye” seems fine…

6350K “on daylight curve” looks white. That’s the meaning of DisplaycCAL test “measured whitepopint vs assumed whitepint”.
6350K “white” should be “assumed 6400 or 6300k” and a ver low dE to dthose daylight whites.
Also you can get  6500K and not white, but pink or green cast. In that situation youll get assumed 6500K but high dE to daylight curve.

The only 100% working programs are Eizo CN7 and NEC SVII.

CN7 does not officially support WLED PFS, only older GB-LED like CS2730. You need to patch it using unofficial EDRs by Eizo Asia Pacific. I assume some problems with royalties regarding IP of EDRs… but who knows.

SV2 does not support WLED PFS (PA271Q, PA311D). It applies some cheap hack, an offset, like default CN7 correction when you configure it to do not use EDRs. Like with all generic offsets they are not portable between i1d3 so… for some displays will get a wrong correction

As an example of a typical WLED PFS like yours and a typical i1d3
-using GB-LED EDR (no patched CN7) will give you 4dE to daylight (vs using actual EDR)
-using default offset (CN7 configured to no EDR) will give you 2-3dE to daylight (vs using actual EDR)
-using patched CN7 will give you <1dE to actual white

However, there is still no possibility of doing more calibrations and then freely loading them into the LUT of the monitor. 3 slots is very little. And each new calibration takes a lot of time, you know, than calibrating, for example, 10 presets and then selecting which one we put into the monitor.

I agree, that will be a very good update for Benq PMU, Dell DUCCS, LG software…. even if you only have 3 slot, let user upload and download calibrations so you can recall calibrations on disk (not very old, so they can validated with an “all green” mark).

• This reply was modified 9 months, 3 weeks ago by Vincent.

[nick234]

Thanks again for your help.

The NEC PA311 is certainly well calibrated at the factory and must keep its parameters well over time because it has advanced sensors. while the EA series that requires calibration works great based on SVII. moreover, NEC withdrew the best of the best screens from the PA series. So probably there will be no update of this program for these monitors … unfortunately, only Eizo surprises me negatively, it’s their well-known model and has a slip-up

After all, BenQ is in a way not well thought out to the very end – but it should be valued for its lower price, the possibility of hardware calibration and a wide gamut that has its uses. The new generations of these monitors are already better thought out with better uniformity. It seems that the bad days of the SW2700PT are over.

What about the calibration in general? I think I even met on this forum that the above-mentioned calibration is a bit deceptive and not as full-fledged as Eizo or NEC. Is it true? I’m also curious if these monitors actually have 3D LUTs or are they matrix calibrated?

[Vincent]

Thanks again for your help.

The NEC PA311 is certainly well calibrated at the factory and must keep its parameters well over time because it has advanced sensors. while the EA series that requires calibration works great based on SVII. moreover, NEC withdrew the best of the best screens from the PA series. So probably there will be no update of this program for these monitors … unfortunately, only Eizo surprises me negatively, it’s their well-known model and has a slip-up

Latest Basiccolor 6 has PFS_Phosphor_Family_31Jan17(for macbooks! not for newer nec PAs!) and NEC integration… but it will be easier to manually correct white on these PAs since grayscale shoudl be ok for very long time.

After all, BenQ is in a way not well thought out to the very end – but it should be valued for its lower price, the possibility of hardware calibration and a wide gamut that has its uses. The new generations of these monitors are already better thought out with better uniformity. It seems that the bad days of the SW2700PT are over.

Hack manually EDRs to include QLED SPD liek discussed in many threads if you want to use proper EDR for SW2700PT

What about the calibration in general? I think I even met on this forum that the above-mentioned calibration is a bit deceptive and not as full-fledged as Eizo or NEC. Is it true? I’m also curious if these monitors actually have 3D LUTs or are they matrix calibrated?

It does not matter, you have no mesh data of actual display behavior (25 profiling patches! 3 node per cube edge!)… same for Eizo CG-X using vanilla CN7 calibration.

If you want to use its LUT3D in these SW  models you need to rely on 3rd party like Light Illusion (more expensive than monitor itself!!! ).

OTOH for CG-X EIzo allows you to upload a LUT3D file manually to act as a chain for a native gamut. AFAIK all other vendors have not this feature.
This means that you calibrate your GC-X to certain white & nits with CN7, profile XYZ in DIsplayCAL, calculate a LUT3D and manually upload that LUT3D to monitor through an option on CN7.

It will be a nice feature for future PMU releases… if you’re his customer… ask them to add it. They have the SDK to do that, but it looks like they only licensed it to LightIllusion & others.
Given the price target of SW models, releasing such SDK will mean bad news for those  calibration software companies… so it is more lileky taht Benq won’t care about his customers and promote paid 3rd party integration.

Anyway, you have DWMLUT in Windows, free, dithered output… visually akin to HW calibration

• This reply was modified 9 months, 3 weeks ago by Vincent.

[nick234]

Okay, now I learned a lot of interesting things from you. I can get back in a few days with the reports if you’re curious. Just I remember there is some function to simulate, you described it, but I still don’t understand. I will be grateful if you send a picture with the correct settings in DisplayCal to do a good validation.

Have you personally used this new PMU?

[Vincent]

If you mean Benq’s “Paper Color Sync”, that useless app calculates a LUT3D to simulate another decive ICC (a printer-ink-paper), but as in DisplayCAL LUT3D  with matrix display profiles as destination, you’ll be using an idealized description of display, not a full 17 node ^3 dexription, not even a 7 node (400 patch) of a simple XYZLUT profile.
Also such apps are useless without a fine tune  option for whitepoint after calibration. Your reference light source for print evaluation may be slightly off from D50 SPD, or your eves from std. obs.

AFAIK PMU has not such feature, but if it has it, great news!

Full LUT3D calibration implies (no matter which vendor) full mesh characterization of display.
What is the point of simulating sRGB on a LUT3D based on a native gamut “matrix-like” description of a display?
matrix (primaries+TRC)  -> to -> matrix (primaries+TRC)
A lut-matrix-lut can do the same, with greater precission (1024 entry instead of 17 node per ramp from Black to WRGB)

As said before a vanilla calibration using CN7 on a CG-X is matrix-like because there is no full mesh data. Optionally you can load a true LUT3D after calibration in CN7, later, a LUT3D transformation based on a 3d mesh charaterization of display (XYZLUT DisplayCAL, Lightillusion… whatever you want)

[nick234]

 I will be grateful if you send a picture with the correct settings in DisplayCal to do a good validation

In reference to this question, can you please send a screenshot with the correct settings?

[Author]

Posts