Page 1 of 1

XL2546 Calibration Results (so far) and Questions

PostPosted: 03 Apr 2019, 11:26
by TaDaa
Hi Guys,

New to this forum but wanted to share my XL2546 calibration results to reduce crosstalk so far and ask a few questions. Attached my image so far -- was able to achieve this by essentially using a custom user color mode and reducing the RGB values until cross talk was minimal (then adjust other settings to help compensate). Tried to find a good balance of color to crosstalk - the result actually looks similar but better than ASUS ELMB, which is pretty washed out at least IMO (tried the XG248Q and returned it for various reasons). I am planning to calibrate this some more, but thought it looked pretty good so far (but I'm a noob) and figured it was worth sharing -- photo at 240hz, 960 pixels/sec and taken with an IPhone.


Few questions regarding this monitor:
1) I thought the panel was supposed to be an AU Optronics panel? My service menu shows QISDA BMS Q250HTN00.A000 - is this normal for the XL2546?
2) When turning Dyac off, the screen is extremely blurry even moreso than my XF250Q. It is somewhat fixable with calibration, but wanted to see if this is normal behavior or if this panel is a dud.
3) Is there anyway to save the overdrive settings in the factory settings?
4) Any other tricks/tweaks that people have figured out with this monitor? I've tried the vertical totals/~180hz, but it doesn't seem to improve crosstalk at all for me.

Thanks guys!

Re: XL2546 Calibration Results (so far) and Questions

PostPosted: 03 Apr 2019, 23:58
by TaDaa
Regarding the blurriness, I am seeing something that strikes me as odd in the OSD.

1:1 and Aspect are both disabled/grayed out and there seems to be no way to enable them. The blurriness does look like some type of scaling or something so I am wondering if this is somehow related. The manual suggests that changing picture modes can enable these settings, but none of them do. Will probably end up calling benq and returning if not fixable. Firmware is V002 20180911.

Re: XL2546 Calibration Results (so far) and Questions

PostPosted: 04 Apr 2019, 02:22
by Chief Blur Buster
Crosstalk-wise, that does indeed look really good for 240Hz strobing.
What you did was the "Reduced Contrast Ratio" trick of reducing strobe crosstalk, as I understand.

Mind if you share your RGB settings?

Now, about the blurriness, this may be fixable by switching to a different resolution, as Custom Resolutions are not always successfully at 1:1. You may want to delete your 240Hz Custom Resolutions (after screenshotting them) and making sure you start from a clean slate with the default 240Hz EDID to see if any scaling artifact disappears.

Re: XL2546 Calibration Results (so far) and Questions

PostPosted: 04 Apr 2019, 11:51
by TaDaa
Thanks for the reply Chief!

Yeah I ended up last night on :
Color Temperature - RGB - 43/43/45
Color Vibrancy - 11
Strobe Intensity - 4
Strobe Area - 14
Sharpness - 1
Black Equalizer - 2
Gamma - 4
Brightness - 90
Contrast - 100
AMA - premium
DyAc - premium

Thanks for the tip on EDID. I tried clean slate, but 1:1 and aspect are still disabled. Called tech support to ask, their recommendation was to use "gpu scaling" which they said would put the monitor into passive mode, which did not help either. They did not have any specific recommendations on getting 1:1 mode enabled. Regarding the blurriness with DyAc=off, they suggested that the problems were most likely caused by some type of pixel inversion and that I should return or exchange this monitor.

In case anyone was wondering about the QISDA vs AU panel, according to BenQ support its supposedly the same panel just manufactured at a different location. This is kind of a concern for me as well I guess -- error rates from different factories/locations.

I also noticed some strain last night while attempting to reset my settings to default. Some of the default modes with this monitor are overly bright I could feel my eyes picking up on some type of lines/bars slowly moving in the background. These moving bars felt even more intense on dark backgrounds. After staring at it for a good bit and looking away, it felt like the moving bars image was imprinted on my eyes. Lowering the brightness and other settings does fix this problem though (at least for me)

Regarding freesync - not sure if its just this QISDA panel, but I was unable to set freesync on with this monitor even if disabling DyAc. Any attempt at freesync on would work on windows desktop only but resulted in a permanent black screen in all games.

Hopefully someone finds this info useful, but unfortunately I have to say good bye to this monitor now. Here's to waiting for (hopefully) decent 0.5ms.

Re: XL2546 Calibration Results (so far) and Questions

PostPosted: 04 Apr 2019, 14:40
by Chief Blur Buster
Thank you for the settings! Which leads me to believe, you did a really great job calibrating for low strobe crosstalk. It certainly looks very low-crosstalk (only faintly at top/bottom) for 240Hz strobing.

Photography-taking-technique commentary: Pursuit camera is not necessary for crosstalk analysis of strobe backlights, and static camera photographs -- as long as exposure capture only one strobe flash -- are acceptable for crosstalk analysis. Also, even with an imperfectly taken iPhone photograph, I almost can't see any crosstalk -- just barely at top/bottom edges only if I zoom! And you took a portrait photo of a landscape screen, creating perpendicular scan directions that would sometimes amplify visibility of artifacts. So you actually took a worse-case photograph that may have amplified crosstalk visibility but that didn't happen!

About blurriness with DyAc=OFF is hard to diagnose without some very good careful photographs.

About inversion artifacts:
Does the blurriness only occur during moving images? Fine checkerboard textures in solid colors? thread with screenshots. Fine-lines textures in solid colors? See .... That's not scaling blurriness but an artifact from interaction with the monitor's alternating positive/negative pixel voltage in alternate refresh cycles. This is supposed to be invisible but sometimes becomes visible as artifacts especially during motion tests, and specially of odd-pixel-steps.