Tearing & Micro-stuttering In Every Game

[HeroofTime]

Hello all,

I kindly request help in regards to keeping games as fluid and responsive as possible. Please bear with me as I list off my settings, specs, and what I've tried learning about the topic so far. I would greatly appreciate advice and clarification on what I've gathered.

AMD 1800X w/ Phanteks PH-TC14PE (3850MHz, 100MHz BCLK)
G.SKILL 3200MHz CL16 16GBx2 (3066MHz, 14-14-14-14-28-42-4-6-24-1T)
GIGABYTE GTX 1080 Ti Gaming OC (2038MHz, 5940MHz)

I notice small hitches and micro-stuttering at random times when playing games like PUBG, CS:GO, Insurgency, and RB6 Siege. Screen tearing is also an issue. I run my XL2730Z at 144Hz with Blur Reduction enabled. I use driver 399.24 and will not be using driver 417.71's FreeSync capabilities because my XL2730Z disables Blur Reduction if FreeSync is enabled. Maximum pre-rendered frames is set to 1 in the graphics driver. V-sync will not be an option because I want the least amount of input lag. My mouse and keyboard run at 1000Hz. My mouse runs at 8000DPI. Mouse sensitivity is always lowered greatly in-game.

I've read that 144Hz panels should be lowered in order to reduce latency, and that your FPS should be in sync with that refresh rate (for example, 138Hz & 138FPS Vs. 144Hz & 144FPS). https://www.blurbusters.com/wp-content/ ... 7362a2.png

I've read that you should keep your refresh rate at 144Hz and cap your frame rate at 142FPS. https://www.blurbusters.com/gsync/gsync ... ttings/11/

I've read that you should keep your refresh rate at 144Hz and cap your frame rate at 143.XXFPS (don't have a source for it from Blur Busters, but it's recommended occasionally on other sites).

What's better, using RTSS or an in-game frame limiter instead?

I think that's all I had on my mind and what I needed clarification and help with. Thanks a lot for your time everyone.

PS: I've read that you should have your FPS limited to -3 of whatever your G-SYNC or FreeSync range is. Is this true? My brother has a G-SYNC panel and I'd be willing to tweak his system too. I don't have G-SYNC, but I am definitely curious about that too. Thanks again.

[RealNC]

V-sync will not be an option because I want the least amount of input lag.

Run uncapped, or use a very high cap. Without vsync or any other sync method, you want as much FPS as you can get. For this, the higher the single-core performance of you CPU is, the better. A 4-core CPU with higher clocks is better here than an 8-core with lower clocks. I don't know how Ryzen OCing works, but if it's possible to turn off 4 of the cores and then raise the clocks of the remaining cores to 4.2GHz or such, that would probably give you an FPS boost.

My mouse and keyboard run at 1000Hz. My mouse runs at 8000DPI. Mouse sensitivity is always lowered greatly in-game.

That's way too high. Even anything above 800dpi is too much. For CS:GO for example, 800dpi with an in-game sensitivity in the range of 1.0 to 2.0 is normal.

At 8000, how can you even control the mouse cursor?

I've read that 144Hz panels should be lowered in order to reduce latency

Where on earth did you read something like that? That's nonsense.

and that your FPS should be in sync with that refresh rate (for example, 138Hz & 138FPS Vs. 144Hz

You're not using vsync. You want need as high FPS as you can get.

I've read that you should keep your refresh rate at 144Hz and cap your frame rate at 143.XXFPS (don't have a source for it from Blur Busters, but it's recommended occasionally on other sites).

That for vsync.

What's better, using RTSS or an in-game frame limiter instead?

When not using vsync, neither. If you still want an FPS cap, the in-game limiter is better because it can reduce input lag.

PS: I've read that you should have your FPS limited to -3 of whatever your G-SYNC or FreeSync range is. Is this true?

The 14-page "G-Sync 101" article was not written as an April Fools joke

[Notty_PT]

Well, I personally don´t like to have like 200fps on a 144hz refresh rate. It makes everything weird. I would rather deal with the ocasional tear at 138fps, for example, when not using Vsync or VRR.

Also, having a steady consistent framerate will improve Mouse aiming a lot!

[Chief Blur Buster]

I've read that you should keep your refresh rate at 144Hz and cap your frame rate at 142FPS. https://www.blurbusters.com/gsync/gsync ... ttings/11/

Only "cap 3fps below" for Variable Refresh Rate Monitors like GSYNC or FreeSync.

If you hate tearing, then you should not "cap 3fps below" for VSYNC OFF.

GSYNC -- yes, cap 3fps below
FreeSync -- yes, cap 3fps below
Not using GSYNC/FreeSync -- NO, don't cap 3fps below

If you're wanting ultrasmooth tearfree stutterfree on a non-VRR monitor, and you don't want the lag of VSYNC ON .....
..... then you want the "Low Lag VSYNC ON HOWTO" or "Scanline Sync" instead.

[HeroofTime]

@RealNC Thank you for the information. I could definitely read the articles written on the site here and gather the information I need, but I've tried to do so and had a very hard time following the topics written. There are simply too many details for me to comprehend at once, and what's even funnier is that I thought I knew all that there was about monitors. So, if I choose not to use any sync method, running uncapped is the best thing to do in regards to input lag. The trade-off there is that I'll be experiencing screen tearing, for this is the method I've always used (uncapped). In regards to my CPU, I've overclocked my whole system to the brim. If I push the speeds any more, it will become unstable. If I push the voltage more, I'll be entering territory where longevity might be of concern. In regards to my mouse, I use very low in-game sensitivity to order to control the high DPI. For example, I use 0.287 sensitivity in CS:GO.

Please click on the link I provided in my original post in regards to using 138Hz versus 144Hz. The graph shows lower input lag as a result of lowering the panel's refresh rate for some reason. I must be missing something though because 144Hz naturally should have a lower response time than 138Hz. In regards to using 143.XXFPS for my 144Hz panel, I do understand now that it's only for V-sync and the reasoning behind how it works. Simply put, there's no way to reduce screen tear and micro-stuttering outside of any sync method? Just making sure. Lastly, using in-game frame limiters are always best for every instance except when using V-sync correct? Thank you for your time again RealNC.

@Chief Blur Buster Thank you for the information. I looked up the page in regards to the low lag V-sync post you made, and it really does help a lot and I notice almost no input lag. The one game that I notice the slight input lag in is CS:GO, but I can honestly play with it just fine in competitive matches. I like the low lag V-sync method and properly configured it according to your shared method. I tried looking at the scanline sync method, but it seems too complex for me to tackle. As a result, I'll be sticking with the low lag V-sync method you've shared. In regards to having FPS capped to -3 of your G-SYNC range, I'm not too sure why -3 is the special number instead of any other number. Is there any reasoning behind that? Also, should V-sync be enabled, alongside G-SYNC (with -3FPS), in the driver as well as in-game? Lastly, should frame limiting be forced through RTSS or the game itself? Thanks a lot for your time Chief Blur Buster.

I feel like I'm asking the same question over again and I really apologize if this is the case. I'm sorry for sounding repetitive, but I'm trying to make things sound as clear as possible. This is all I should really need though! Thanks again.

https://www.blurbusters.com/howto-low-lag-vsync-on/
viewtopic.php?f=10&t=4916
https://www.blurbusters.com/gsync/gsync ... ettings/2/

[jorimt]

Please click on the link I provided in my original post in regards to using 138Hz versus 144Hz. The graph shows lower input lag as a result of lowering the panel's refresh rate for some reason. I must be missing something though because 144Hz naturally should have a lower response time than 138Hz.

Yes, the way Battle(non)sense labeled that is a bit confusing. That's with G-SYNC/FreeSync. Both of those syncing methods effectively adjust the refresh rate to the framerate, so at 138 FPS, what he's showing is that you're basically at 138Hz.

This is true in one regard, and not quite true in another. With G-SYNC/FreeSync, the panel is refreshing 138 times per second with a 138 FPS limit, but it's still ultimately scanning each frame in at 144Hz:
viewtopic.php?f=5&t=4874&start=30#p37943

So, that test result does not apply to standalone V-SYNC on/off at all. It only applies if you're using G-SYNC or FreeSync.

In regards to having FPS capped to -3 of your G-SYNC range, I'm not too sure why -3 is the special number instead of any other number. Is there any reasoning behind that? Also, should V-sync be enabled, alongside G-SYNC (with -3FPS), in the driver as well as in-game? Lastly, should frame limiting be forced through RTSS or the game itself?

I happened to have posted a new FAQ on my G-SYNC 101 article further clarifying these points today, take a look:
https://www.blurbusters.com/gsync/gsync ... ttings/15/

Also, optimal G-SYNC settings (and their reasoning) can be found on this page:
https://www.blurbusters.com/gsync/gsync ... ttings/14/

If you haven't already, I suggest you read the entire article, front to back; it answers nearly everything you've been asking about in this thread.

[RealNC]

So, if I choose not to use any sync method, running uncapped is the best thing to do in regards to input lag. The trade-off there is that I'll be experiencing screen tearing

You will have tearing regardless of whether you cap your FPS or not. Tearing is not the result of frame rate. It's the result of not syncing. The only thing that changes when the frame rate changes is the tear line moves differently across the screen. But it's always there.

In regards to my mouse, I use very low in-game sensitivity to order to control the high DPI. For example, I use 0.287 sensitivity in CS:GO.

There's no reason to do this though.

Please click on the link I provided in my original post in regards to using 138Hz versus 144Hz. The graph shows lower input lag as a result of lowering the panel's refresh rate for some reason.

That graph looks wrong to me. Also I can't see why you'd conclude 138Hz is faster. There's nothing to compare it to! It says that 138Hz with in-game cap of 138FPS, vsync OFF and g-sync ON, there's an average latency of 18.87ms. But where is the measurement at 144Hz? There is none. If there was, it would be lower than 18.87ms.

I can't make any sense from that graph :-/ Why would he measure latency with these settings at 138Hz but then not measure it for 144Hz? I don't get it

Simply put, there's no way to reduce screen tear and micro-stuttering outside of any sync method?

No. If there was, there wouldn't be any need for sync methods, now would there

Lastly, using in-game frame limiters are always best for every instance except when using V-sync correct?

Correct. With some exceptions. If the in-game limiter is too jittery, it can have microstutter. In this case, using RTSS instead might help.

In regards to having FPS capped to -3 of your G-SYNC range, I'm not too sure why -3 is the special number instead of any other number. Is there any reasoning behind that?

Yes. If FPS gets too close to the upper range of g-sync (144Hz), then you get v-sync like input lag if you have vsync enabled, or tearing if you have vsync disabled.

Also, should V-sync be enabled, alongside G-SYNC (with -3FPS), in the driver as well as in-game?

The rule of thumb is vsync OFF in-game, vsync ON in the nvidia panel. There can be exceptions to this if some games don't work well. Most games work best with those settings though.

Lastly, should frame limiting be forced through RTSS or the game itself? Thanks a lot for your time Chief Blur Buster.

Game's own limiter is usually best, unless it has too much microstutter, in which case RTSS will probably fix that.

[jorimt]

That graph looks wrong to me. Also I can't see why you'd conclude 138Hz is faster. There's nothing to compare it to! It says that 138Hz with in-game cap of 138FPS, vsync OFF and g-sync ON, there's an average latency of 18.87ms. But where is the measurement at 144Hz? There is none. If there was, it would be lower than 18.87ms.

I can't make any sense from that graph :-/ Why would he measure latency with these settings at 138Hz but then not measure it for 144Hz? I don't get it

Yeah, unfortunate wording choice. See my post above your last for what I think he meant by that.

[HeroofTime]

Thank you for your time everyone. I understand pretty much everything now. I've learned how to use each sync method, and what's best for each sync method. I'm simply having three issues now. The first one is something I'd really like to get to the bottom of, for it might resolve my other issues if it gets straightened out. Currently, I'm using the low lag V-sync method. I play all my games in borderless windowed mode due to having two monitors (both BenQ 1440p panels, 60Hz and 144Hz). I'm aware of the slight input lag that borderless windowed mode brings. I hate it when my games minimize when I tab out of them.

NVIDIA CP's global V-sync function does nothing. I know this for a fact because of two reasons. First of all, my FPS doesn't lock to my monitor's refresh rate. Secondly, I still notice screen tearing with V-sync enabled in NVIDIA CP. So, what I resorted to doing was setting the function to letting the 3D application decide. I now enable V-sync in my games, and as a result I've noticed a stark difference in how uniform everything looks through this method (V-sync is simply working now). Am I running into this issue due to running my games in borderless windowed mode? Is there something else that I might be missing?

In borderless windowed mode, I noticed a long time ago that I will experience major frame dips if I have anything moving (even .gif images) on my second monitor. It wasn't like this at all in Windows 7, but this became a problem once I upgraded to Windows 10. I simply make sure that I have nothing playing or moving on my second monitor now. Since enabling V-sync in my games, I'll get locked to 60FPS on my primary 144Hz monitor while gaming if I tab out of my game too many times or receive a Steam message (not always, just sometimes). I always have the following applications open on on my second monitor. Firefox, Discord, a Steam chat box, and a command prompt box pinging my ISP constantly. I know for a fact that command prompt doesn't cause any issues on my primary monitor even though there's lines of text constantly moving up as each ping is being executed. Anyways, I'm unsure as to what could be causing my primary monitor to get locked to 60FPS. RTSS is capping frames at 143.90FPS at all times.

With that said, why am I getting periodic bursts of micro-stuttering in CS:GO with the low lag V-sync method? I've tried both double buffering V-sync and triple buffering V-sync in CS:GO to no avail. I do not get these periodic bursts of micro-stuttering in PUBG, so I know it's not an issue in respect to changing the frame rate cap in RTSS to something lower than 143.90FPS.

Thanks a lot for your guys' help.

[jorimt]

Currently, I'm using the low lag V-sync method. I play all my games in borderless windowed mode due to having two monitors (both BenQ 1440p panels, 60Hz and 144Hz). I'm aware of the slight input lag that borderless windowed mode brings. I hate it when my games minimize when I tab out of them.

NVIDIA CP's global V-sync function does nothing. I know this for a fact because of two reasons. First of all, my FPS doesn't lock to my monitor's refresh rate. Secondly, I still notice screen tearing with V-sync enabled in NVIDIA CP. So, what I resorted to doing was setting the function to letting the 3D application decide. I now enable V-sync in my games, and as a result I've noticed a stark difference in how uniform everything looks through this method (V-sync is simply working now). Am I running into this issue due to running my games in borderless windowed mode? Is there something else that I might be missing?

I'm pretty sure the NVCP V-SYNC option for fixed refresh rate syncing solutions (G-SYNC/FreeSync off), only applies to games in exclusive fullscreen. Whereas in-game V-SYNC options can typically override borderless/windowed V-SYNC behavior.

I juts launched Overwatch, disabled G-SYNC, switched to borderless windowed mode via the in-game settings (at 144Hz), and set the FPS limiter to 300.

Borderless + in-game V-SYNC off = DWM triple buffer activates (I've personally never been able to get tearing in borderless with any configuration), allows framerates above the refresh rate.
Borderless + in-game V-SYNC on = DWM triple buffer overridden with in-game double buffer V-SYNC, framerate capped to max refresh rate
Borderless + in-game V-SYNC on + in-game triple buffer option on = same as above, since the in-game triple buffer solution isn't a DWM or Fast Sync variant, but is effectively double buffer with a third buffer solely to prevent half refresh lock when framerate dips below refresh rate.

For all of the above, I had NVCP V-SYNC forced on globally, and it had zero effect in borderless mode.

So, short of some exceptions, what you are describing is generally expected behavior with borderless.

In borderless windowed mode, I noticed a long time ago that I will experience major frame dips if I have anything moving (even .gif images) on my second monitor. It wasn't like this at all in Windows 7, but this became a problem once I upgraded to Windows 10.

Welcome to mixed-refresh multi-monitor configuration on a modern Microsoft OS; it's a mess. Short of not using a second monitor, I have no advice on this one.

With that said, why am I getting periodic bursts of micro-stuttering in CS:GO with the low lag V-sync method? I've tried both double buffering V-sync and triple buffering V-sync in CS:GO to no avail. I do not get these periodic bursts of micro-stuttering in PUBG, so I know it's not an issue in respect to changing the frame rate cap in RTSS to something lower than 143.90FPS.

Probably CS:GO-specific. After their battle royale/FTP update, I've seen multiple reports of unavoidable microstutter. Likely nothing you can do until the devs address it on their side.