Page 1 of 1
Regarding VRR smoothness with FPS cap vs no cap
Posted: 23 Jul 2024, 01:37
by GFresha
Hello,
Could someone please explain the aspect of smooth frame pacing in VRR/G-Sync? Specifically, I'm interested in the difference between having VRR ON and leaving the game uncapped, allowing the monitor to match its Hz to the GPU frame output, versus having VRR ON and capping the framerate to a level the PC can reach 99% of the time.
In my personal experience, I was playing Cyberpunk with full RT/PT at 4K, and my frame rate fluctuated between 60-80. With VRR on, it felt really smooth. However, when I kept VRR on and capped my frames to 70 using RTTS, it didn't feel as smooth. I usually leave my games uncapped and let VRR match Hz to FPS, and my games always feel smooth as long as I don't hit super low frames like 40. I only cap with VRR if my game is a competitive one and it's exceeding my max Hz.
Now, my question is: Is the "smoothness" aspect of VRR (not talking about the low latency zero tearing aspect) more pronounced when you leave your game uncapped and let VRR match Hz to FPS? Or is it the same level of smoothness even when you cap your frames to a level you can hit 99% of the time? I assume even when you cap your FPS, VRR is still matching hz to cap, and the perceived smoothness is the same uncapped VRR vs capped vrr right?
Perhaps it's a placebo effect for me, but I always prefer the smoothness of uncapped VRR in single-player graphics games versus capping my FPS. Yet in competitive games like CS and Valorant, I keep VRR ON but cap FPS below max Hz, and that also feels smooth. I just don't like capping my FPS in graphics games when I am in the low range like 60-80. I'd rather uncap and let VRR handle fluctuations.
Re: Regarding VRR smoothness with FPS cap vs no cap
Posted: 24 Jul 2024, 14:15
by Luviaz
GFresha wrote: ↑23 Jul 2024, 01:37
Now, my question is: Is the "smoothness" aspect of VRR (not talking about the low latency zero tearing aspect) more pronounced when you leave your game uncapped and let VRR match Hz to FPS? Or is it the same level of smoothness even when you cap your frames to a level you can hit 99% of the time? I assume even when you cap your FPS, VRR is still matching hz to cap, and the perceived smoothness is the same uncapped VRR vs capped vrr right?
If the frame rate is already within your monitor's VRR range, a frame limiter isn't necessary. It's only required when the frame rate exceeds the VRR range, causing the VRR to disengage and revert to either V-sync On or Off behavior. As for fluidity, it varies. External limiters like RTSS can interfere with API present calls, which games may not recognize, potentially affecting their rendering pipeline. However, most games I've capped with RTSS FPS don't show any difference compared to using their in-game or driver limiter.
In multiplayer games, a frame cap is essential. Without it, games can run at hundreds of FPS, leading to tearing and irregular frame pacing. For example, League of Legends runs at 600-700 FPS on my system when uncapped, resulting in characters appearing to teleport rather than move smoothly.
Re: Regarding VRR smoothness with FPS cap vs no cap
Posted: 24 Jul 2024, 15:44
by GFresha
Luviaz wrote: ↑24 Jul 2024, 14:15
If the frame rate is already within your monitor's VRR range, a frame limiter isn't necessary. It's only required when the frame rate exceeds the VRR range, causing the VRR to disengage and revert to either V-sync On or Off behavior. As for fluidity, it varies. External limiters like RTSS can interfere with API present calls, which games may not recognize, potentially affecting their rendering pipeline. However, most games I've capped with RTSS FPS don't show any difference compared to using their in-game or driver limiter.
In multiplayer games, a frame cap is essential. Without it, games can run at hundreds of FPS, leading to tearing and irregular frame pacing. For example, League of Legends runs at 600-700 FPS on my system when uncapped, resulting in characters appearing to teleport rather than move smoothly.
Okay thats great to know, yeah for most if not all triple A graphics games I just leave uncapped and let VRR handle it, but thats interesting what you say about a potential API issue with RTTS when it comes to some games, I think I may have experienced that
But overall, VRR will be engaged whether you are using a frame limiter or not and smoothness should be the same right? It will only not be engaged if we exceed hz.
Re: Regarding VRR smoothness with FPS cap vs no cap
Posted: 24 Jul 2024, 17:37
by Luviaz
GFresha wrote: ↑24 Jul 2024, 15:44
But overall, VRR will be engaged whether you are using a frame limiter or not and smoothness should be the same right? It will only not be engaged if we exceed hz.
Yes, smoothness is generally consistent, at least based on my experience; however, I haven't tested every game, so I can't claim it's the same across the board.
The closest to smoothness would be with V-sync on when FPS equals Hz due to buffering; however, this introduces input latency. The higher the frame rate, the less noticeable the V-sync input lag becomes; anything below 120 fps is quite perceptible to me. Moreover, if the frame rate doesn't match the Hz, which is common in demanding games, V-sync stutter occurs. VRR addresses this issue, making it an ideal complement to V-sync. It's important to remember that if the frame rate reaches the max Hz, it defaults to plain V-sync, with the associated input lag. Hence, the consensus is to use VRR + V-sync and set the frame rate a few frames below the max Hz to keep VRR active.
Input latency varies according to personal sensitivity; some people are more susceptible than others. I don't mind V-sync input lag for story-driven RPGs, but for games that require precise input, like rhythm, shoot-em-ups, or competitive multiplayer games, it definitely matters.
Re: Regarding VRR smoothness with FPS cap vs no cap
Posted: 24 Jul 2024, 19:08
by GFresha
Luviaz wrote: ↑24 Jul 2024, 17:37
It's important to remember that if the frame rate reaches the max Hz, it defaults to plain V-sync, with the associated input lag. Hence, the consensus is to use VRR + V-sync and set the frame rate a few frames below the max Hz to keep VRR active.
Thankfully VRR displays are plenty and no reason not to use VRR, but in the case someone is using vsync, say for a story game, where input delay is not crucial, is the general method to match frames to hz for vsync to be active?
How would you enable it? on in NVCP and off in game just like VRR setup? Or we actually have to enable it in game too?
Re: Regarding VRR smoothness with FPS cap vs no cap
Posted: 25 Jul 2024, 01:14
by Luviaz
GFresha wrote: ↑24 Jul 2024, 19:08
Thankfully VRR displays are plenty and no reason not to use VRR, but in the case someone is using vsync, say for a story game, where input delay is not crucial, is the general method to match frames to hz for vsync to be active?
How would you enable it? on in NVCP and off in game just like VRR setup? Or we actually have to enable it in game too?
If input latency isn't a concern, then yes, or if a frame limit is causing issues with the game and V-sync is the only frame cap option available.
NVCP will override the in-game V-sync setting if it's set to On or Off. This means that the in-game setting is irrelevant if NVCP is set to On or Off. Only the NVCP "Use the application setting" will respect the in-game V-sync setting. If a game's V-sync is functioning properly, then it's best to use it for simplicity's sake. You should only attempt to force V-sync through NVCP if the in-game V-sync is not working correctly.