Good fps limits for 144hz? 72fps? 108 fps?

[Denkis]

What are some good fps caps for 144hz besides 72 fps?

In many games I can't keep 144/120+ fps and the responsiveness is noticeably suffering. I don't want to use G-Sync because I want to avoid any input lag no matter how small for multiplayer games.

I've seen people say 72 fps is a good cap for 144hz. Is it only the half-point that is smooth? Or wil for example any multiple of 24, say 108, work just as well?

If the fps doesn't fluctuate, is the difference between a "good cap" and a "bad cap" only in terms of screen tearing, or input lag as well? If it's just screen tearing I can try a bunch of caps and see which ones don't cause tearing.

It's hard to tell sometimes if the sluggish feeling at <144/120fps is due to fps inconsistency or just lower fps in general. Would be nice to know how this works on paper.

[Supafly]

Provided fps is stable. If you can a 144hz display to 72 thats means you see the same frame twice as it's half of the max. If you want to lower it more as it's not stable/getting tearing divide 144 by 3 and you'd get 48. Setting to 48 will mean you see every frame 3 times.

Don't setting caps somewhere inbetween. If you set it at say 60 that means you'll see 2.4 frames per hz of the display. That will mean constant tearing. So divide 144hz to get whole numbers only. 2,3,4 and 6 which would mean 72, 48,36 and 24. Not 5 as it's a decimal number 28.8.

May want to read more on G-Sync. Have this bookmarked
https://blurbusters.com/gsync/gsync101-input-lag-tests-and-settings/14/

[Denkis]

Suicidal Monkey původně napsal:

Provided fps is stable. If you can a 144hz display to 72 thats means you see the same frame twice as it's half of the max. If you want to lower it more as it's not stable/getting tearing divide 144 by 3 and you'd get 48. Setting to 48 will mean you see every frame 3 times.

Don't setting caps somewhere inbetween. If you set it at say 60 that means you'll see 2.4 frames per hz of the display. That will mean constant tearing. So divide 144hz to get whole numbers only. 2,3,4 and 6 which would mean 72, 48,36 and 24. Not 5 as it's a decimal number 28.8.

May want to read more on G-Sync. Have this bookmarked
https://blurbusters.com/gsync/gsync101-input-lag-tests-and-settings/14/

I'm more concerned about input "lag" (variance). For example if the fps jumps between 100-130, mouse aiming usually feels inconsistent. Will there be any input lag or other problems on for example 100 fps cap? Or is it just the tearing?

I can set my monitor to 100Hz or create a custom resolution with any refresh rate <144. For example I tried creating a custom resolution with 90Hz, and the monitor measures 90 correctly. But it's cumbersome to change per game.

[Gen X]

tripple buffering...

problem is it can increase input lag
not a big concern for me
but thats the whole point of gsync or freesync no need for tripple buffering

[Autumn_]

Denkis původně napsal:

What are some good fps caps for 144hz besides 72 fps?

In many games I can't keep 144/120+ fps and the responsiveness is noticeably suffering. I don't want to use G-Sync because I want to avoid any input lag no matter how small for multiplayer games.

I've seen people say 72 fps is a good cap for 144hz. Is it only the half-point that is smooth? Or wil for example any multiple of 24, say 108, work just as well?

If the fps doesn't fluctuate, is the difference between a "good cap" and a "bad cap" only in terms of screen tearing, or input lag as well? If it's just screen tearing I can try a bunch of caps and see which ones don't cause tearing.

It's hard to tell sometimes if the sluggish feeling at <144/120fps is due to fps inconsistency or just lower fps in general. Would be nice to know how this works on paper.

G-sync won't add ANY input lag (that you will feel.)

If you don't have V-sync (or steady FPS with G-sync), you will always have tearing.

If you're limiting with software like RTSS, NvidiaCP, or AMDs one I forget the name, you will add input latency, but it'll be less than a frame.
Also, stopping your GPU from hitting 95% usage will reduce latency too.

But, when you cap your FPS, and don't use G-Sync, you should nake sure every frame is displayed 1, 2, 3, etc times.
You can see if that's the case with some easy maths;
144÷144=1, each frame is displayed one time.
144÷72=2, each frame displayed 2 times.
144÷60=2.4, some frames will be displayed 2 times, some will display 3 times. This is bad.

Refresh rate devided by frame rate. Integers are good. Decimals are bad.

Denkis původně napsal:

Suicidal Monkey původně napsal:

Provided fps is stable. If you can a 144hz display to 72 thats means you see the same frame twice as it's half of the max. If you want to lower it more as it's not stable/getting tearing divide 144 by 3 and you'd get 48. Setting to 48 will mean you see every frame 3 times.

Don't setting caps somewhere inbetween. If you set it at say 60 that means you'll see 2.4 frames per hz of the display. That will mean constant tearing. So divide 144hz to get whole numbers only. 2,3,4 and 6 which would mean 72, 48,36 and 24. Not 5 as it's a decimal number 28.8.

May want to read more on G-Sync. Have this bookmarked
https://blurbusters.com/gsync/gsync101-input-lag-tests-and-settings/14/

I'm more concerned about input "lag" (variance). For example if the fps jumps between 100-130, mouse aiming usually feels inconsistent. Will there be any input lag or other problems on for example 100 fps cap? Or is it just the tearing?

I can set my monitor to 100Hz or create a custom resolution with any refresh rate <144. For example I tried creating a custom resolution with 90Hz, and the monitor measures 90 correctly. But it's cumbersome to change per game.

If you want consitency, use RTSS, it gives you perfect frametimes.
The negligible increase in latency is worth it.

Changing refresh rate is pointless. If you're doing that you may aswell just use G-sync.

[Denkis]

Autumn původně napsal:

G-sync won't add ANY input lag (that you will feel.)

Changing refresh rate is pointless. If you're doing that you may aswell just use G-sync.

From what I know there is a slight input delay added by G-sync. But it may very well be negligible.

Autumn původně napsal:

But, when you cap your FPS, and don't use G-Sync, you should nake sure every frame is displayed 1, 2, 3, etc times.
You can see if that's the case with some easy maths;
144÷144=1, each frame is displayed one time.
144÷72=2, each frame displayed 2 times.
144÷60=2.4, some frames will be displayed 2 times, some will display 3 times. This is bad.

Is this only about screen tearing or input lag as well? That's what I wanted to know.

Autumn původně napsal:

Also, stopping your GPU from hitting 95% usage will reduce latency too.

I didn't know this. I just looked it up and it seems to be common knowledge. I'll be sure to look out for this in the future. I don't think it's possible to hard cap the gpu usage. I would need to do this via a fps limit I think.

[Autumn_]

Denkis původně napsal:

Autumn původně napsal:

G-sync won't add ANY input lag (that you will feel.)

Changing refresh rate is pointless. If you're doing that you may aswell just use G-sync.

From what I know there is a slight input delay added by G-sync. But it may very well be negligible.

You could try a blind test, with g-sync, if you have another person to change the settings for you.
Get them to randomly select G-sync on or off, and you play for 5 mins, without them telling you, 5 times. And then see which ones you thought were G-sync vs no G-sync.
(With the same FPS cap obviously, otherwise you'll have loads of variance.)

You shouldn't be able to tell them apart. And if you do, holy ♥♥♥♥, I'm impressed.

Denkis původně napsal:

Autumn původně napsal:

But, when you cap your FPS, and don't use G-Sync, you should nake sure every frame is displayed 1, 2, 3, etc times.
You can see if that's the case with some easy maths;
144÷144=1, each frame is displayed one time.
144÷72=2, each frame displayed 2 times.
144÷60=2.4, some frames will be displayed 2 times, some will display 3 times. This is bad.

Is this only about screen tearing or input lag as well? That's what I wanted to know.

It's not about screen tearing - You can't FULLY remove tearing without the use of V-sync.
And, you can reduce it by using G-sync, but if you have large FPS swings, you'll still get tearing.
You can use RTSS's Scanline-sync, which allows you to move the tear line, to a more acceptable location. (Also limits the FPS to refresh rate.)

As for input lag, you want higher FPS, but remain below <95% usage with said FPS cap. That'll be the best you can do.

But, the main reason why you cap at an integer, is for smoothness.
Each frame is displayed the same amount of times, for the same duration, it means a much smoother experience.
(Only time you wouldn't mind, is if you're at like several hundred (200-300+ FPS, depending on panel refresh rate.)

Denkis původně napsal:

Autumn původně napsal:

Also, stopping your GPU from hitting 95% usage will reduce latency too.

I didn't know this. I just looked it up and it seems to be common knowledge. I'll be sure to look out for this in the future. I don't think it's possible to hard cap the gpu usage. I would need to do this via a fps limit I think.

Lots of people don't know actually, just a few of us here on the forums actually know - or even believe it.
Most people are still under the impression that 99% is the ''THE BEST!! IT'S GOOD!!''
(Some actual logic would dictate otherwise, like, for example, why wouldn't you want to run any other part at it's limit? That causes issues right? So why would the GPU be any different?)

I'm not aware of any way to limit the GPU usage, I would use such a think if I knew about it.
The only way I know of, is RTSS.

But, thinking about it also, you don't want to limit it to NOT use that extra ~5%, you do want it as a 'buffer' of sorts, so if there is a frame that takes a little more processing to do, it doesn't just ♥♥♥♥ the bed.

Edit ; Messed a quote up.

[Gen X]

i dont see the big deal with input lag anyways even with classic vsync and tripple bufferingvi never go above 0.2-0.3 VAR in csgo anyways
guess on some setups it matters
but in my almost 20 years experience i never had an issue with it let alone notice it unless i was compensating for it in another way
im rusty on hw right now so i cant break down into details

[Autumn_]

Gen X původně napsal:

i dont see the big deal with input lag anyways even with classic vsync and tripple bufferingvi never go above 0.2-0.3 VAR in csgo anyways
guess on some setups it matters
but in my almost 20 years experience i never had an issue with it let alone notice it unless i was compensating for it in another way
im rusty on hw right now so i cant break down into details

It's not so much if YOU can feel it.
Lots of people go for THEIR feeling/interpretation.
I, personally, can feel V-sync latency even at 144FPS.
I don't like it, I never will. Some people can't feel it or notice it at 60FPS.

Different people see and feel different things - every sensory perception a person has, is subjective.

Though if you can't feel V-syncs latency, I am happy for you, I really am.

[Gen X]

maybe i should see what its like at 144hz
been using 60 hz whole life

[Autumn_]

Gen X původně napsal:

maybe i should see what its like at 144hz
been using 60 hz whole life

If you're looking to try it, I really do recommend it. Even if you spend most of your days doing work on it.
Was pretty life changing for me, at 60hz I would get constant headaches. Now with 144hz, I barely get any.

[Gen X]

never got a headache either
im not old
i was a gamer actually
just havent touched pc in a year

might rebuild dunno
just need a mobo and some ram
got a brand new zen quad core i only booted to cmos with never used it
might put it to use

[Autumn_]

Gen X původně napsal:

never got a headache either
im not old
i was a gamer actually
just havent touched pc in a year

might rebuild dunno
just need a mobo and some ram
got a brand new zen quad core i only booted to cmos with never used it
might put it to use

You'll notice, my entire last sentence, I was talking about myself. Not anyone else, since it was my experience. I never said you, or OP suffered from it. I just mentioned it - My reason for owning one.

[Snow]

Gen X původně napsal:

i dont see the big deal with input lag anyways even with classic vsync and tripple bufferingvi never go above 0.2-0.3 VAR in csgo anyways

Since you can't use VSync in CS:GO without triple buffering - why do you mention triple buffering at all? I doubt you know how to turn off triple buffering or at least use different types of triple buffering.
VAR is the difference between what's happening on your PC and on the server. VSync can't possibly affect it as it doesn't affect CPU side of things.

For everything else - Autumn did a good job explaining.

[Denkis]

Autumn původně napsal:

After reading your posts and the article you linked I think I've finally been sold on G-Sync. You know I read that article on Blur Busters over a year ago but it went right over my head.

To summerize if anyone's wondering:

They are saying that there is about 1 frame of delay extra with G-sync (+ V-sync) compared to everything off, at the same FPS. But the catch is, that one frame is "delayed" to avoid screen tearing. I read that as: with G-sync off I might get about 1 frame closer to the last render, but only part of the frame. And it will be folded into the last frame. G-sync seems like a good trade off then. 142 FPS @ 144 Hz + G-sync still has lower input lag than 200+ fps @ 60Hz if I understood correctly. Obviously 200+ fps @ 144 Hz has the lowest lag but that's very situational.

All sync settings off with 200+ fps is still the king for competitive games if you are willing to sacrifice graphics to achieve 200+ fps. I think this is why "pros" don't use G-sync. I think I had misunderstood the reason, and have been superstitious towards G-sync.

I will probably use G-sync + V-sync from now on except for ~200 FPS situations: I understand it's best to use only use V-sync via NVCP.

As for fps limit: I think it's still a good idea to cap the fps for games where mouse aiming is crucial. Putting the limit slightly below your average or perhaps lowest FPS in games seems like a good idea (this is with G-sync on btw). Why? If your FPS jumps between 100-120 the input delay will vary. You want to keep that delay as constant as possible for muscle memory. That, and the thing Atumn mentioned about keeping GPU usage around 95%.



Is RTSS more reliable than ingame FPS limiters in your experience? I've actually heard ingame limiters are preferable when available, as external limiters can introduce a slight input lag. NVCP now has a Low Latency option that aims to further reduce input lag. I've read the results are mixed however and it can cause stuttering depending if your CPU usage is high. The options are off / on / ultra. For "ON" you still need to limit FPS yourself. "Ultra" says it minimizes V-Sync latency but not how. I assume it limits FPS though.

[Supafly]

Denkis původně napsal:

Autumn původně napsal:

After reading your posts and the article you linked I think I've finally been sold on G-Sync. You know I read that article on Blur Busters over a year ago but it went right over my head.

I'm the one who linked the blurbusters Article in post #1 and Autumn did the job off explaining things clearly. I don't explain it clearly and Honestly it sometimes goes over my head too. Hence the article I linked as it's much better at explaining that I ever would. Credit to Autumn for explain what I couldn't.

Glad everything is clear now. Have run with G-Sync

EDIT As for the Low latency limiter I've just left it off. Don't play Comp much nowadays so not a major issue if I get a little latency.

FPS limiters I always used RTSS and see no reason to true something else while it works.

Sure Autumn or others will answer those soon enough,