G Sync Lag Question

Everything about latency. Tips, testing methods, mouse lag, display lag, game engine lag, network lag, whole input lag chain, VSYNC OFF vs VSYNC ON, and more! Input Lag Articles on Blur Busters.
Post Reply
Cryogenetic
Posts: 17
Joined: 30 Oct 2021, 19:49

G Sync Lag Question

Post by Cryogenetic » 06 Aug 2022, 04:50

So I've done a lot of reading with the amazing work done on this forum regarding Gsync but I can't seem to find one answer, wondering if someone can clear this up for me.

From what I understand in game FPS limiters cause less lag than RTSS or Nvidia's FPS limiters. For a game like League of Legends the in game cap is 200 or 240 (so in my case with a 240hz Freesync I have to use 200). My alternative is using Low Latency Mode - Ultra which will cap the FPS to 225 instead.

My question is, am I better off using the in game FPS limiter since they cause less input lag and cap it to 200, or does the Nvidia Low Latency Mode use a different method that has comparable input lag to in game limiters to cap the FPS to 225?

For CSGO should I cap 236/237 in game or set Low Latency Mode to Ultra and let it sit at 225? I don't really mind the 10-12 FPS loss for better input lag.

I do seem to aim better with Freesync off but I've been playing for over 10 years so I think a part of it is my eyes just need to adjust to how the panel is presenting frames and I'd like to try to get used to it, as it's my understanding the actual input lag should be not noticeable by my human eye. I'm fairly sensitive to input lag but even I don't notice it so maybe my above question also falls within that range, but I seem to notice a frame or two which RTSS/Nvidia limiter introduces.

Kyouki
Posts: 190
Joined: 20 Jul 2022, 04:52

Re: G Sync Lag Question

Post by Kyouki » 06 Aug 2022, 07:11

I'll post a great link to some super amazing testing subreddit with a great amount of details and accuracy. I stand behind these findings as I tested on my own.

https://old.reddit.com/r/allbenchmarks/ ... idias_fps/
Tentative Conclusion / FPS Limiter Recommendation(s) (UPDATED 05/03/20)
On DX11 scenario:

Stability-wise, the RTSS and NV v3 (NV CP) limiters were almost on par and they were significantly better than all the rest of methods too.
In terms of approximate and expected input lag, the NV v3 (NV CP) was the best method overall.
On DX12 scenario:

Stability-wise, the NV v2 limiter was the best method overall and the in-engine limiter the worst one. RTSS limiter was close to NV v2 one, and significantly better than both the NV v1 & v3 (NV CP) (worse L-Shapes distributions and Lows avg numbers with both NV v1 & v3, stutters).
In terms of approximate and expected latency, the NV v3 (NV CP) was also the best method overall.
On Vulkan scenario:

Stability-wise, the RTSS limiter was clearly superior than all the Nvidia's limiters (worse L-Shapes distributions and Lows avg numbers with all the NVIDIA's limiters, stutters), being the RTSS limiter the best method overall and the in-engine limiter the worst one.
In terms of approximate and expected input lag, there were almost no differences between all the FPS limiters.
CPU: AMD R7 5800x3D ~ PBO2Tuner -30 ~ no C states
RAM: Gskill Bdie 2x16gb TridentZ Neo ~ CL16-16-16-36 1T ~ fine tuned latency
GPU: ASUS TUF 3080 10G OC Edition(v1/non-LHR) ~ disabled Pstates ~ max oced
OS: Fine tuned Windows 10 Pro, manual tuned.
Monitor: Alienware AW2521H ~ mix of ULMB/Gsync @ 240hz/360hz
More specs: https://kit.co/Kyouki/the-pc-that-stomps-you

Shade7
Posts: 222
Joined: 25 May 2022, 18:44

Re: G Sync Lag Question

Post by Shade7 » 06 Aug 2022, 12:59

Is your monitor Gsync or Gsync compatible?

I see you mentioning Freesync a lot- so wanted to make sure.

From what I’ve experienced- The Gsync experience is very monitor-specific (differs from one monitor to the next).

If your monitor is only Gsync-compatible, I’d probably not bother with it. Gsync compatible monitors (at least the few I’ve tried, led to strange tearing/graphical abnormalities when Gsync was on. I play Valorant mainly and my latency numbers (according to the in game latency stats) are lower with Gsync + Reflex (in game version of Nvidia low latency), but the visuals are scuffed, so I just prefer to keep it off and cap fps ingame.

The true tested “Gsync” monitors seem to have implemented Gsync much better.

You will really have to try it out and see for yourself.

Cryogenetic
Posts: 17
Joined: 30 Oct 2021, 19:49

Re: G Sync Lag Question

Post by Cryogenetic » 07 Aug 2022, 02:07

Shade7 wrote:
06 Aug 2022, 12:59
Is your monitor Gsync or Gsync compatible?

I see you mentioning Freesync a lot- so wanted to make sure.

From what I’ve experienced- The Gsync experience is very monitor-specific (differs from one monitor to the next).

If your monitor is only Gsync-compatible, I’d probably not bother with it. Gsync compatible monitors (at least the few I’ve tried, led to strange tearing/graphical abnormalities when Gsync was on. I play Valorant mainly and my latency numbers (according to the in game latency stats) are lower with Gsync + Reflex (in game version of Nvidia low latency), but the visuals are scuffed, so I just prefer to keep it off and cap fps ingame.

The true tested “Gsync” monitors seem to have implemented Gsync much better.

You will really have to try it out and see for yourself.
Sorry for being confusing, my setup is a gtx 3060ti so in the control panel it is Gsync but my monitor is the Benq XL2540k which has freesync premium (gsync compatable)

I actually have incredibly clear motion with this monitor and the visuals are great. I just wonder if I'm handicapping myself by using the Reflex or Ultra Low Latency mode for LoL or Valorant.

Cryogenetic
Posts: 17
Joined: 30 Oct 2021, 19:49

Re: G Sync Lag Question

Post by Cryogenetic » 07 Aug 2022, 02:09

Kyouki wrote:
06 Aug 2022, 07:11
I'll post a great link to some super amazing testing subreddit with a great amount of details and accuracy. I stand behind these findings as I tested on my own.

https://old.reddit.com/r/allbenchmarks/ ... idias_fps/
Tentative Conclusion / FPS Limiter Recommendation(s) (UPDATED 05/03/20)
On DX11 scenario:

Stability-wise, the RTSS and NV v3 (NV CP) limiters were almost on par and they were significantly better than all the rest of methods too.
In terms of approximate and expected input lag, the NV v3 (NV CP) was the best method overall.
On DX12 scenario:

Stability-wise, the NV v2 limiter was the best method overall and the in-engine limiter the worst one. RTSS limiter was close to NV v2 one, and significantly better than both the NV v1 & v3 (NV CP) (worse L-Shapes distributions and Lows avg numbers with both NV v1 & v3, stutters).
In terms of approximate and expected latency, the NV v3 (NV CP) was also the best method overall.
On Vulkan scenario:

Stability-wise, the RTSS limiter was clearly superior than all the Nvidia's limiters (worse L-Shapes distributions and Lows avg numbers with all the NVIDIA's limiters, stutters), being the RTSS limiter the best method overall and the in-engine limiter the worst one.
In terms of approximate and expected input lag, there were almost no differences between all the FPS limiters.
Thank you for this, I think I'm still confused though, which of the above does Nvidia Reflex and/or Ultra Low Latency capping my monitor automatically to 225 fall under?

Kyouki
Posts: 190
Joined: 20 Jul 2022, 04:52

Re: G Sync Lag Question

Post by Kyouki » 07 Aug 2022, 04:52

Cryogenetic wrote:
07 Aug 2022, 02:09
Kyouki wrote:
06 Aug 2022, 07:11
I'll post a great link to some super amazing testing subreddit with a great amount of details and accuracy. I stand behind these findings as I tested on my own.

https://old.reddit.com/r/allbenchmarks/ ... idias_fps/
Tentative Conclusion / FPS Limiter Recommendation(s) (UPDATED 05/03/20)
On DX11 scenario:

Stability-wise, the RTSS and NV v3 (NV CP) limiters were almost on par and they were significantly better than all the rest of methods too.
In terms of approximate and expected input lag, the NV v3 (NV CP) was the best method overall.
On DX12 scenario:

Stability-wise, the NV v2 limiter was the best method overall and the in-engine limiter the worst one. RTSS limiter was close to NV v2 one, and significantly better than both the NV v1 & v3 (NV CP) (worse L-Shapes distributions and Lows avg numbers with both NV v1 & v3, stutters).
In terms of approximate and expected latency, the NV v3 (NV CP) was also the best method overall.
On Vulkan scenario:

Stability-wise, the RTSS limiter was clearly superior than all the Nvidia's limiters (worse L-Shapes distributions and Lows avg numbers with all the NVIDIA's limiters, stutters), being the RTSS limiter the best method overall and the in-engine limiter the worst one.
In terms of approximate and expected input lag, there were almost no differences between all the FPS limiters.
Thank you for this, I think I'm still confused though, which of the above does Nvidia Reflex and/or Ultra Low Latency capping my monitor automatically to 225 fall under?
Reflex + Boost = the same as setting to High Performance / Ultra on Low Latency Mode, which will auto cap your framerate. (there are some sources on this forum)

imo,
for LoL I think I played nicely on just using Gsync with it and cap it using nvidia control panel.
for csgo I wouldn't use Gsync, if you wanna play it competitively try using Ultra Low Motion Blur -- or fixed refresh rate at the highest resfresh rate possible. CSGO produces enough frames in order to appear very smooth and the tearing usually is well beyond framerate to be little noticeable.

Supposedly depending on the ingame frame rate limiters (which do it on the CPU) the input latency is less.
Would only use Ultra on Low Latency Mode when you're GPU bound (95-99% GPU usage) otherwise just use On.

Ingame limiter if the implementation is correct should yield better input latency
RTSS/NVCP results in a smoother image with a tiny bit of input latency plus there is also a hidden benefit as often ingame ones tend to limit also the engine limit, doing this on the GPU doesn't and stops it right before it hits your monitor instead and smooths out frametime.
I've hardly ever used ingame options for the most part as I found it nicer to be able to cap on my own and toggle on/off to test.
CPU: AMD R7 5800x3D ~ PBO2Tuner -30 ~ no C states
RAM: Gskill Bdie 2x16gb TridentZ Neo ~ CL16-16-16-36 1T ~ fine tuned latency
GPU: ASUS TUF 3080 10G OC Edition(v1/non-LHR) ~ disabled Pstates ~ max oced
OS: Fine tuned Windows 10 Pro, manual tuned.
Monitor: Alienware AW2521H ~ mix of ULMB/Gsync @ 240hz/360hz
More specs: https://kit.co/Kyouki/the-pc-that-stomps-you

Post Reply