NVIDIA G-Sync Pulsar monitor - Asus ROG Strix XG27AQNGV

[edgintheledge]

Went ahead and tested ULMB at 120, 240, and 360 as well just for fun

120

lossless120 ulmb.png (67.6 KiB) Viewed 78114 times

240

lossless240 ulmb.png (72.38 KiB) Viewed 78114 times

360

lossless360 ulmb.png (66.01 KiB) Viewed 78114 times

Not sure what's up with the graphs getting all spiky instead of one clean pulse as before. Possibly the individual "steps" of brightness as the roll goes by? (Assuming that the lossless adapter isn't the bottleneck... again, it's not exactly meant for this!). This could suggest that the 60hz implementation is indeed different than the higher refresh rate modes. My bet is still on it not rolling at 60hz. The clarity I see with my eyes, plus scan out skew disappearing in the 60hz strobe (yet present in 120+ ULMB & pulsar) seems to agree. When I flick my eyes, every pulse with the ulmb 60hz mode stays perfectly straight. The behavior matches what I have personally observed with OLED BFI, VR headsets, and Dyac screens which are all global strobes to my knowledge.

I also have a hunch that partially lit screen images in the 60hz mode is because of phones forcing variable frame rate videos. I don't know of any way to get an iphone to not record VFR. If I select 240fps, it mostly gives me random frame rates like 207.13 or 202.79 instead of 240. These videos will show partially lit screens that look like a roll, but a close to "true" 240 (239.97) video gives me 1 whole screen frame on, 3 full off frames in a repeating pattern as I posted before (or at least for a while until there's a slight desync as the 60hz mode says 59.94 in the display settings and I can't record at exactly 239.76). I would expect a 60hz roll filmed at 240 to have 4 frames of the scanout moving from top to bottom, but that never happens at 60hz on the pulsar monitor right now. Hope the tftcentral deep dive sheds some light on this once it goes public.

[brownvim]

TFTCentral confirmed in their review that the 60 Hz ULMB mode is still using a rolling strobe (I won’t post the paywalled part here).

If you want to test it yourself for comparison, you could turn Pulsar and ULMB2 completely off and try ShaderBeam instead. ShaderBeam has settings for both global strobe and rolling scan emulation (you can switch by changing the Scan Direction parameter).

That way you could run the same measurement test on both modes and compare the graphs directly to what you’re seeing now.

Just a suggestion if you’re interested in getting your own data.

[edgintheledge]

TFTCentral confirmed in their review that the 60 Hz ULMB mode is still using a rolling strobe (I won’t post the paywalled part here).

If you want to test it yourself for comparison, you could turn Pulsar and ULMB2 completely off and try ShaderBeam instead. ShaderBeam has settings for both global strobe and rolling scan emulation (you can switch by changing the Scan Direction parameter).

That way you could run the same measurement test on both modes and compare the graphs directly to what you’re seeing now.

Just a suggestion if you’re interested in getting your own data.

Dang... Im definitely surprised then. I wonder if it's doing something like rolling the backlight faster (i.e rolling at 360hz speed at the end of each 60hz interval?) cause it doesn't quite make sense to me how else it would roll at 60hz without noticeable artifacts. It's been a bit since I last tried ShaderBeam on my previous 240hz monitor but I recall the "jelly effect" with flicking my eyes being more pronounced than the 60hz ulmb. Might be interesting to take a measurement that way, but I'm also gonna trust the people that actually got the equipment and experience testing this stuff probably know more on the topic than I do

[edgintheledge]

Interesting, it’s been a few years since I played on a FW900 CRT but I do remember the phosphor trails it would leave.

From memory the 10 pulse width at 60hz looks cleaner than CRT, how’s it compare to your CRT?

It’s too dim for me, I use it around 100 to 50 which still blows away an OLED at 60hz.

I still think CRTs got a certain magic going on with the combo of the color and motion performance together. Even more so if you have a higher hz unit. My CRT maxes at 120hz, but ulmb and pulsar can't keep up with the motion there. At 60hz though, even with my bias toward CRT I gotta hand the win to ulmb. In dark scenes, the phosphor trails can get very noticeable. And as I've been made aware, ulmb doesn't show scan out skew despite rolling which is noticeable on a CRT.

I usually like the PW 10-20. I find the monitor usually resets down to 100 brightness when I engage ulmb for some reason. Turning it all the way to 500 and then adjusting the PW up helps alot if you don't already do that because anything above 30PW seems like it's burning my retinas on white screens with how bright it is.

[brownvim]

This 60hz looks magical to me, mainly because console gaming has looked so bad with sample and hold displays. I was amazed at 100 pulse width!

The Acer stays at 500 brightness so I only have to adjust pulse width, which I agree is the best way to adjust brightness.

[kyube]

It seems to report the ULMB pulses around 1800nits. Take that with a grain of salt, this thing is meant to be a gamecube adapter with a lag test feature, not a calibrated display tester
---snip---

Nice find! It seems that the Pulsar QHD@360Hz models have a similar voltage boosting amplitude to the one found in the XL2566K.
A shame that the possible duty cycle range on higher Hz is narrowed compared to 60Hz... which leads to the situation that 60Hz is clearer than +120Hz, which should never happen.

[brownvim]

I managed to recreate the ghosting errors in the new firmware using SmoothFrog.

SmoothFrog v0.5p (simple UI version here):https://www.aperturegrille.com/software ... it0.5p.zip

How to test it:
Open the program and press F to bring up the long FPS slider.
Gently wobble the slider left and right to simulate fluctuating frame rates (like real VRR behaviour).

On the new firmware Above ~180 FPS it stays perfectly clean with no ghosting. However, once I drop it between 120 FPS- 180 FPS (or let it fluctuate in that range), I immediately see a sharp ghosted/double image.

Before the new firmware: Ghosting only appeared at 90 FPS and below. The 90–360 FPS range was clean. After the new firmware: Ghosting now starts appearing from around 180 FPS downwards.

I rolled back the firmware today to confirm this problem goes away, 90-360hz is clean. Current firmware even a 1 FPS adjustment up or down causes ghosting between 90-180 FPS which negatively effects the clarity.

[mawi]

I rolled back the firmware today to confirm this problem goes away, 90-360hz is clean. Current firmware even a 1 FPS adjustment up or down causes ghosting between 90-180 FPS which negatively effects the clarity.

This is amazing! Thanks for that. Hopefully this will be adressed soon... Now having clear proof etc.

[Argus]

Does anyone have confirmation whether any of these monitors use DSC or not? I found I was unable to use DLDSR on the Acer monitor but I'm not sure if that's indicative of DSC being on or not.

I think I can use DSR on mine, if this is what you mean:

Untitled.jpg

Strange. DSR doesn't even show up on mine and I can't make custom resolutions or even make one with CRU. Anyone know why?

[kyube]

Strange. DSR doesn't even show up on mine and I can't make custom resolutions or even make one with CRU. Anyone know why?

DP or HDMI port? What GPU model?
Try the HDMI port on the new FW update.