NVIDIA G-Sync Pulsar monitor - Asus ROG Strix XG27AQNGV

[MSIfanboy]

i dont know who to trust when processing delay/input lag tests, you can go on a multitude of websites, and get different results, i dont know, some are testing in games with muzzle flashes or something, others are using LDATS, others are using 1000fps cameras, is this black to white software any good from aperture grille

[Alennartsson]

Hello everyone. I have some questions about PULSAR.
I got my xg27aqngv last week and did try it for some CS2 this weekend.

Today I saw an Reddit post that said that gsync doesn't work over 360fps. I played this weekend with 500+ fps. Does that mean I didn't use the advantages of PULSAR?

I had PULSAR enabled on the monitor. But if gsync is deactivated because 500fps, isn't PULSAR working when?
I came from an 240hz screen so the "improvement" I saw was maybe only the difference between 240-360?

In this Reddit post the said that I should enable "gsync + vsync +reflex" in Nvidia control panel and that it should after that automatically cap my fps to something like 327fps and that this would enable gsync.

So, what settings should I use?
And is pulsar active and working even over 360fps?
Or do gsync need to be active for pulsar to work?

[liquidshadowfox]

@Alennartsson, to my understanding if you go above 360 hz that means you have Vsync disabled and since you are rendering above 360 hz you aren't engaging Gsync while the backlight strobing from pulsar is stuck pulsing the backlight at fixed 360 hz when above the max. Technically it'll still work but you'll get some tearing (if you even notice it at that high of a refresh). The benefit of Gsync pulsar is that if you go below the 360 hz the backlight strobing will match the refresh rate so things look clearer at lower refreshes vs if it wouldn't have been backlight strobbed. If you want the lowest possible latency and don't notice the tearing, you can leave vsync off + Gsync pulsar on and it'll still backlight strobe at a fixed 360 hz but you might get tearing (unless you go into nvidia control panel and enable "fast sync" which will get rid of tearing but then you might get inconsistent frame time delivery which leads to a stutter effect sometimes). Otherwise you can enable reflex in CSGO with vsync enable and it should automatically cap the FPS between 324 - 327 where Gsync is enabled at all times and pulsar will strobe between that refresh. Either way you'll get the best motion clarity, just pick your poison.

[kyube]

i dont know who to trust when processing delay/input lag tests, you can go on a multitude of websites, and get different results, i dont know, some are testing in games with muzzle flashes or something, others are using LDATS, others are using 1000fps cameras, is this black to white software any good from aperture grille

Always consult the highest precision total display latency testing data.
As of now, qsxcv/flood with his 2015 dataset (~1 µs precision = 1MHz) is comfortably ahead of all available solutions on the market.
Ideally, each total display latency data set would include the difference between the top, middle & bottom portion of the display (as latency is a gradient; think of LCDs & OLEDs as a transparent DRAM)
Each numerical value should be represented with at minimum -4 decimal places.
A large enough sample size (+3k samples) is non-negotiable.

A 1000FPS (1kHz) camera & LDAT are both ~1ms precision, which is not enough when dealing with a display's processing latency

OSLTT has a higher precision value (~15µs = ~66 kHz), but is limited by the UI to be useful for the representing data properly to the end-user.
One would have to resort to manual visualization methods for properly visualizing the end-effect. E.g.: Using violin plots

OpenLDAT also has a higher precision.

I played this weekend with 500+ fps. Does that mean I didn't use the advantages of PULSAR?

Enable ULMB2 only instead.
The use-case of PULSAR (GSYNC+ULMB2) isn't for CS2, where you achieve +500FPS easily.
It's for games in which you cannot sustain >360FPS consistently (e.g. The Finals, most UE4 games etc.)

[Chief Blur Buster]

I see the numbers, digits, have a shadow of them self occur before the sharp rendition of them. If that is ghosting, or reverse ghosting, I don't know, but its there at 1920 pixels/sec. If I disable Pulsar, the inverse ghosting is actually less perceivable.

We call this strobe crosstalk -- www.blurbusters.com/crosstalk and www.testufo.com/crosstalk

It's very hard to eliminate. Very few displays can go zero-crosstalk. That's why Viewsonic XG2431 was so popular for many years since it was possible to tune it to go zero-crosstalk at low Hz -- Discorz wrote a large piece at www.blurbusters.com/xg2431-discorz

Being that said, Pulsar is still one of the best strobed VRR implementations seen -- other strobed *VRR* implementations have had way more crosstalk than Pulsar has.

[kyube]

Being that said, Pulsar is still one of the best strobed VRR implementations seen -- other strobed *VRR* implementations have had way more crosstalk than Pulsar has.

Do you have any information on whether there's a possibility that Nvidia can release a firmware upgrade where:
• ULMB2 mode would feature a adjustable pulse width setting (akin to the PG27AQN)
• Pulsar being supported through the HDMI port & having the full FRL6 bandwidth available instead of TMDS? This would open up an entire untapped use-case for these displays, for console gaming.
• Removing GPU vendor lock-in
These 3 are the biggest blockers from making these Pulsar models appealing.

[Chief Blur Buster]

Being that said, Pulsar is still one of the best strobed VRR implementations seen -- other strobed *VRR* implementations have had way more crosstalk than Pulsar has.

Do you have any information on whether there's a possibility that Nvidia can release a firmware upgrade where:
• ULMB2 mode would feature a adjustable pulse width setting (akin to the PG27AQN)
• Pulsar being supported through the HDMI port & having the full FRL6 bandwidth available instead of TMDS? This would open up an entire untapped use-case for these displays, for console gaming.
• Removing GPU vendor lock-in
These 3 are the biggest blockers from making these Pulsar models appealing.

From what I was told:
- No plans for adjustable pulse width away from 25% because of the complexity (worse crosstalk, etc)
- Unknown about HDMI, but I thought it would have worked over HDMI if the video source is an NVIDIA card.
- I don't think they plan to remove GPU vendor lock-in; you will need an NVIDIA GPU to drive Pulsar.

[kyube]

From what I was told:
- No plans for adjustable pulse width away from 25% because of the complexity (worse crosstalk, etc)

Just to clear up any possible misunderstanding, I was asking about the fixed refresh rate mode (ULMB2), not the variable refresh rate mode (Pulsar)
Does this hold true for the fixed refresh rate mode?

- Unknown about HDMI, but I thought it would have worked over HDMI if the video source is an NVIDIA card.

According to all the manuals I've read on the 4 available Pulsar models, they all have HDMI 2.0 (labeled as 2.1, but TMDS instead of FRL6)
This means that HDMI is limited to 1440p@120Hz, where Pulsar requires a minimum of 240 Hz.

[liquidshadowfox]

They should also increase the range where pulsar is disabled, for example I might want pulsar to only activate above 180 hz since below that it starts to get some artifacts

[Chief Blur Buster]

They should also increase the range where pulsar is disabled, for example I might want pulsar to only activate above 180 hz since below that it starts to get some artifacts

Don't forget that tighter refreshtime:gtgtime can also increase artifacts too (e.g. 360fps 360Hz, while the crosstalk is smaller distance, is more intense, than 180fps 360Hz). There's less time for pixel response to fully complete between briefer refreshtimes.

There's a lot of interacting factors (less intense crosstalk but at bigger divergence at larger pixel steps VERSUS more intense crosstalk at smaller divergence at smaller pixel steps), for a given physical motionspeed (centimeters/sec of motion).

In short, higher Hz means smaller pixel step (for the same physical motionspeed) but higher Hz also means less time for LCD GtG to finish in total darkness between strobe flashes.