NVIDIA G-Sync Pulsar monitor - Asus ROG Strix XG27AQNGV

[mango87]

How well does this monitor switch between 1080p HDMI console input and native resolution displayport input? If I enable ULMB2 for 60hz console gaming, can I have a separate Pulsar setting for displayport input?

[hammelgammler]

Do you think there’s any hope that we get an “extreme Pulsar” mode, with a lower tuned pulse width? Given that with 360Hz/240Hz ULMB2, NVIDIA was possible to reduce the MPRT from 100% to 50% or even 25% PW, it looks like in should be theoretically possible?

I need to set the brightness to the minimum value to get my target 120 nits, so there would be plenty of PW headroom left. Maybe a Pulsar brightness target of 200 nits would be a good compromise?

Generally, I am impressed with the motion clarity though, especially 120Hz 10 PW. It would be so awesome to have ~0.9ms MPRT for 120-360Hz Pulsar.

[liquidshadowfox]

Do you think there’s any hope that we get an “extreme Pulsar” mode, with a lower tuned pulse width? Given that with 360Hz/240Hz ULMB2, NVIDIA was possible to reduce the MPRT from 100% to 50% or even 25% PW, it looks like in should be theoretically possible?

I need to set the brightness to the minimum value to get my target 120 nits, so there would be plenty of PW headroom left. Maybe a Pulsar brightness target of 200 nits would be a good compromise?

Generally, I am impressed with the motion clarity though, especially 120Hz 10 PW. It would be so awesome to have ~0.9ms MPRT for 120-360Hz Pulsar.

I only hope they would but I know it's difficult given how complicated it is to do VRR + Backlight strobing. I just found out this week they actually do variable duty cycle based on FPS, 20% duty cycle around 100 hz, 25% above 120 hz I think and so on until they get to like 40% duty cycle towards 300 fps+; Plus take into account they are also doing PWM compensation pulses in between so you don't see any brightness fluctuations and the scaler maybe has a fixed time cost for calculating all of this in real time.

I would love to see a 10% - 20% decrease offset for the duty cycle across the board just to see what kind of results we would get in a "Gsync pulsar extreme" mode even with reduced brightness.

Either way I'm happy with what we have so far, just waiting on the software side to catch up (namely reflex updates, driver updates and nvidia app updates to allow us to automatically switch between ULMB 2 refresh rates/pulse width, pulsar and regular Gsync mode for regular operation when not gaming.

[kyube]

Do you think there’s any hope that we get an “extreme Pulsar” mode, with a lower tuned pulse width? Given that with 360Hz/240Hz ULMB2, NVIDIA was possible to reduce the MPRT from 100% to 50% or even 25% PW, it looks like in should be theoretically possible?

A higher likelihood of that occurring than NVidia 'unlocking' the MT9810 (scaler IC found within the Pulsar panels) to support full HDMI 2.1 FRL6 bandwidth & avoid all the DSC memes 30/40 series users are complaining about
It took them 9 months to make ULMB2 into a usable state (eclipsing the predecessor flagship, PG27AQN) compared to how dreadful it was at launch... maybe next year we'll get a tighter duty cycle range for Pulsar

[mawi]

Do you think there’s any hope that we get an “extreme Pulsar” mode, with a lower tuned pulse width? Given that with 360Hz/240Hz ULMB2, NVIDIA was possible to reduce the MPRT from 100% to 50% or even 25% PW, it looks like in should be theoretically possible?

A higher likelihood of that occurring than NVidia 'unlocking' the MT9810 (scaler IC found within the Pulsar panels) to support full HDMI 2.1 FRL6 bandwidth & avoid all the DSC memes 30/40 series users are complaining about

DSC is not causing this issue, RTX50 still uses DSC above 240 Hz without any problems. According to DrG-S its the dual head mode where they use two logic pipes to drive higher clock speeds. That happens before the signal goes out the DP1.4 with DSC. Is HDMI using different - more capable internal pipes on RTX30/40 cards? If not, it might still be a problem even if they fully support HDMI 2.1.

However, a full 360 Hz signal through HDMI without DSC would be nice, also on RTX50 cards.

[whitespider999]

I purchased this monitor last week, arrived yesterday. Initially I wasnt impressed, but over time (and the latest firmware update) it's been impressing me more and more in a huge variety of ways (many niche related to how fussy i am over specific areas of motion). Could be my favourite screen of the past 6 years. Including my 500hz qd-oled. Many complain about how this ips is as expensive as a higher end 1440p oled.

This screen has a real special nature to it. Im finding.

[hammelgammler]

I purchased this monitor last week, arrived yesterday. Initially I wasnt impressed, but over time (and the latest firmware update) it's been impressing me more and more in a huge variety of ways (many niche related to how fussy i am over specific areas of motion). Could be my favourite screen of the past 6 years. Including my 500hz qd-oled. Many complain about how this ips is as expensive as a higher end 1440p oled.

This screen has a real special nature to it. Im finding.

I agree. Normally, I swore to never buy an IPS monitor ever again and used many OLEDs in the past (LG 48CX -> LG 48C1 -> LG 42C2 -> Samsung S90C 55" -> Samsung S95F 55"). Even though the CX already had great motion clarity at 60Hz and 120Hz, I never used it because of the increase in input lag and of course brightness.

Then in my quest in finding a great low input lag monitor, especially for 60 FPS content, I bought many CRTs (Sony G520, Sony G500, Sony FW900, Sony W900) and they are indeed great. I would even argue the W900 is generally better for 60Hz, as the flickering is less harsh than ULMB2, which is expected given the phosphor decay of CRTs (which is a negative by itself).

Now I'm in a situation, where I experienced awesome HDR content on OLED (400 nits fullscreen, 4000 nits peak), arguably the best static image experience you can get and the great motion clarity of CRTs. Which is why I will try if I can switch back to an IPS panel full time, simply to not have to think whether or not I'm using HDR with "bad" motion or SDR with good motion.

My new setup would then look as follows:

Sony W900 | Asus XG27AQNGC | Sony 20M2MDE
60 Hz | >60 Hz | 240p60

That would be the ultimate tri-factor of motion clarity, each with their own strengths and use-cases.

The only real alternative would be the 1440p 540Hz OLEDs, but achieving ~500 FPS in modern titles is out of the question, even with a 9800X3D and 5090. It has become to a point, where I have so much options what to play on which screen, all with their pro's and con's, that I spend more time thinking about it than actually playing, which bothers me. I hope that I can get back to SDR/IPS and simply live with it, until maybe the next generation of Pulsar comes around.

[kyube]

DSC is not causing this issue, RTX50 still uses DSC above 240 Hz without any problems. According to DrG-S its the dual head mode where they use two logic pipes to drive higher clock speeds. That happens before the signal goes out the DP1.4 with DSC. Is HDMI using different - more capable internal pipes on RTX30/40 cards? If not, it might still be a problem even if they fully support HDMI 2.1.

However, a full 360 Hz signal through HDMI without DSC would be nice, also on RTX50 cards.

Treat DSC as a hidden "FRL4.5" lock, where anything which fits within FRL4 is able to run effortlessly without hitting above single-head.
This implies that you're stuck with QHD@330-340Hz, assuming you use HDMI 2.1.
Depending on timings used, you could squish it into ~350Hz, but that's absolute maximum for a single head.
With DP1.4, the pixel clock is ~970MHz. You run into a cable limitation before you run into the HW limitation of Ampere & Ada Lovelace.

In fact, Blackwell has the same issue as well, just with a higher bandwidth
Here's how DP2.1 scales:
UHBR10 is ~1616 MHz pixel clock
[single head limit of Blackwell is here, 1620MHz]
UHBR13.5 is ~2181 MHz pixel clock
UHBR20 is ~3232 MHz pixel clock

In comparison, HDMI would be like this:
2.1 FRL4 is ~1185 MHz pixel clock
[single head limit of Ampere / Ada Lovelace would be here]
2.1 FRL5 is ~1481 MHz pixel clock
[single head limit of Blackwell would be here]
2.1 FRL6 is ~1777 MHz pixel clock

Ampere/Ada Lovelace would act as "FRL4.5"
Blackwell would act as "FRL5.5"

QHD@360Hz would require ~1399 MHz bandwidth, which comfortably fits within a single head of Blackwell's engine

[wiseude]

I'm awaiting Alder Lake, which is why I'm still using a 9900K.

Better formated like this?
Sorry I meant to say Nova.
Some other person convinced me to get this monitor and I think I'm going to give it a go.(since it's on sale where im getting it I'm going to get it now)
The new PC will follow when Nova lake releases in a couple of months.

Worst comes to shove il have a decent ips with a g-sync module.It's certainly better then what I currently have.

[RonsonPL]

disclaimer:
CPU only post. Related to "Pulsar or OLED" discussion but not about this monitor or Pulsar tech

The only real alternative would be the 1440p 540Hz OLEDs, but achieving ~500 FPS in modern titles is out of the question, even with a 9800X3D and 5090.

Actually, if you'd like to use ULMB2 120, even ignoring UE5s microstutters, shader compilation stutter and traversal stutter, 9800X3D on Nvidia GPUs struggles to maintain 120. I'd say the CPU completely fails if you task it with maintaining 200fps in the more demanding new games. Even well optimized Battlefield 6 cannot squeeze stable 300+fps from this CPU.
Sadly, nobody cares about HFR. We were lucky AMD made X3D, the gaming savior, by accident (they didn't want to make a gaming CPU, it was for servers, and someone realized "hey, this actually is awesome for games")
Instead of focus on that, we get DRMs, UE5, security patches, all with "don't worry, it slows your gaming by almost nothing, like 1%", and sometimes the reality is that 1% lows and 0,1% lows suffer by -20%.

The new PC will follow when Nova lake releases in a couple of months.

Rumors suggest April-May for the only one which could rival or even just get close to 9800X3D - the model with full cache (bLLC)
Meanwhile more and more rumors show up (or being deliberately spread to cause panic, we'll never know) about everything in electronics going up. PCB prices, materials for mainboards, intel already started price hikes, and remember what were the launch prices for the last intel release - only because nobody was buying them and many, many months later, Intel lowered the procing. If Nova with cache is good and can rival AMD, I expect both AMD and Intel raising the prices and on top of that - intel will set the prices in the most stupid way possible. For people who'll buy it no matter what "because it's Intel" and for non-gaming workloads, Nova Lake with 144MB cache may be pricier because it has 28 cores in total, which will give Intel an excuse. Meanwhile AMD may stop or limit the production of 9800X3D like it did with 7800X3D which caused almost instant +50% price hike back then.

I hope we won't see Zen 6 starting with 6 core Ryzen 10600 at 450$ while 9800x3d moves from 450$ to 800$ or goes EOL.

Let's hope it won't be as bad, but we all hoped the same a year ago about RAM, half a year ago about GPUs and a few months ago about SSDs. And yet, the worst scenario happened. Let's hope at least this time it's just fake panic spread to push people into buying now.