How To Fix V-Sync Input Lag

Published 01.05.2019 в 18:00 | Guide rating: 76


Just like tickrates, the understanding of this concept is fairly simple yet overly praised as some kind of necessity which in the end completely is worthless by the law of diminishing returns.
By sheer math and actual research among opinions and facts, there has been quite a lot of ignorance being spreaded around, in vain of perhaps showing PC gamers that the highest framerate is nothing to benefit from if you don't have the hardware for, this guide came to my mind, even though it'll probably get viewed in a negative way.
I'm pretty sure a few will be sane enough to try out what they are missing as essential features.
A true sweet spot.
So unless you are willing to completely open your mind and think with other than just overrated opinions, take your time and read this guide.
Having that out of the way.
I will be covering; V-Sync (Double Buffer / Triple Buffer) Enhanced Sync (AMD) / Fast Sync (Nvidia) FreeSync (AMD) / G-Sync (Nvidia) Adaptive Sync (Nvidia only) As a warning however like mentionned above.
All solutions have their trade offs that you as the user should try to witness first before making a decision, ultimately for a short TD;LR.
V-Sync is the most blunt and aggressive solution, while anything else is either passive and/or investment to go for.

The Problem (Tearing and V-Sync Alone)

How can you achieve gameplay without any slow downs and inconsistencies, tearing even? By capping your framerate to the lowest stable framerate or your maximum refresh rate.
Why is it? If you are one of these players using the lowest resolution with the lowest settings, it'll probably be around 300 though that number is stupidly high as is most likely to drop further when anything occurs in game.
There is no reason to keep it that high, nor stress your components, it won't give you an advantage or better registry, whatever excuses you may hear at this high.
In fact you are more prone to frametime stutters and inconsistencies for every action scenes that may come, specifically 60Hz monitors will look like it's having noticeable slow down with said framedrops because the tearing appears bigger the lower the framerate while 144Hz improves, but does "not" remove any flaws if you get the framedrop in anyway.
For example if you were to have a frame drop from 60 FPS to 40 FPS, it would behave better input wise on 144Hz than 60Hz, but there will still be similar tearing.
The reasons why 144Hz looks "smoother" compared 60Hz in so many aspects is because of the tearing being almost absent by your eyes at higher frames, without forgetting you are able to witness 144 FPS instead of only 60 which contrary to having 200 FPS framedrops to 100 on 60Hz, 144Hz will display more of those and that in mind avoid some tearing because of that, that is still 44 FPS missing from the refresh, making a slow down noticeable, next to being that you are ACTUALLY witnessing more updates a second physically always compliments the real input instead of 60Hz which cannot feel any better than what 60Hz already is.
What is frametime? The time is takes to render a frame and the higher the framerate is, the harder it is to keep it stable.
So how do you fix this lag? Simple, everyone will probably scream at me for saying this, it's possible to use Vertical Syncronization without it's input lag, giving you a smooth display without having to worry about the drops in high framerates as well.
"But V-SYNC causes mouse lag, it's stupid and worthless you don't want to have mouse lag and who really cares, high framerates improves your performance quite by a large margin!" The mouse lag would be true if you were to leave V-Sync uncapped, bear with me, this is not a personal opinion, I have been getting a lot of witnesses to try this, they all had the same answer, which was consistency and "no mouse lag", I repeat "no mouse lag".
Then how the hell is that possible you must ask? Try to imagine how it is to fit 63 frames in a gap of 60Hz? If you thought of mouse lag then you are correct! But how, why is it an issue? Simple, once again, Vertical Syncronization never in it's days, capped the framerate, never it did, I repeat, V-SYNC doesn't cap your frames, it SYNCS them and force the framerate to stay at the current refresh rate by slowing down the current frames to put newer ones ahead, this is where the lag comes from and why people are right to an extend.
Cinque Terre

The Fix For V-Sync

Okay, so how do you fix the mouse lag? As mentionned earlier, since V-Sync slows down the framerate to keep in Sync with the refresh rate and put the next frames ahead.
You want to set a limit to prevent that slow down, that can be done with an FPS cap, easy as that.
fps_max or whatever you may think can possibly work, but will require your own witness to test, see and feel.
So fps_max 60? Not so fast, while you may assume this is the solution, using the engine fps cap leaves the game to assume it doesn't need as much ressources which in the end results in unwanted drops.
You will want to use external benchmark softwares such as Rivatuner which comes along MSI Afterburner.
Using said external benchmark software, you will need to cap your frames at 59, why is that you may ask? While you cap your frames, the GPU still renders ahead of it's time, therefor the frames that are supposed to render ahead, such as the 61th of them will be introducing slight mouse lag if you cap at 60 because of pre rendered frame that goes above 60Hz, setting 59 will prevents the pre rendered frame to cause the issue because the pre rendered frame will become the 60th.
It's important to note that frames are "always" pre rendered in a set number by the drivers for Nvidia, even without v-sync.
All of this doing will result in having your GPU to work with either 1 or 2 back buffers (Depending on which V-Sync mode you use, Double or Triple respectively.
) and swap those frames entirely complete into the front buffer, which is the panel of the monitor that you see and be tearing free.
Finally at 59 fps capped and V-SYNC enabled.
Assuming that the game has a proper V-Sync implementation and doesn't suck at low framerates (Some engines can be terrible at low framerates, even 60, which if you're stuck at 60Hz it's a nightmare, regardless how high your framerate is can or can't help, bear with me this is a very complex situation.
) You will feel absolutely, and which I promise, absolutely no mouse lag and will be provided with no distracting tearing, no fps drops.
This solution can be achieved in the same contrast of higher refresh rates, hopefully your frames won't drop below the refresh rate of the monitor itself, otherwise you will notice the framerate drops but those are less likely because the GPU has to work less to provide the consistency.

The Downsides Of This Fix

And here we are with the trade offs, to be clear.
No solution is perfect.
You will be able to see other sync methods pros and cons below, but this guide mostly focuses on the v-sync lag at hand this section is dedicated to it.
Capping a frame below the refresh rate causes the monitor to drop frames which causes each time the screen refresh to stutter the amount of time this missing frame would take (IE; 59 at 60Hz would be 0.
30ms), while not as bad as it sounds.
It can noticed by some while not at all for others, but this is the trade off to remove tearing, if v-sync isn't turned on, it'll turn into a consistent and very visible tear at the same spot and slowly vary.
This "stutter" does not affect the input of the mouse at all, hence why this solution is still viable because due to that fact, you can still move your mouse around and flawlessly without thinking about it unless you solely focus on it or slowly turn around at a very consistent pace, which is why it's hard to notice and feel like the gameplay is perfect.
Even I was wondering what it was occasionally as I noticed, but it didn't stop me from playing and enjoying the tearingless gameplay.
This method in FPS games might not be favorable, but I'd say try it for yourself before assuming that this "stutter" is a dealbreaker for you, after all you didn't have to spend a dime to achieve these results.
I'd say, the most ideal games would be those without too much aiming and more clicking.
Finally, this "stutter" is not impossible to mitigate either.
A high refresh rate monitor that does not possess any sync technologies will make it almost non existent, at any framerates! So, if you ever had invested in one before those options were out, I highly recommend trying it.

Frametime Analysis

All tests were executed in de_dust2 deathmatch offline with harmless bots to avoid deaths that cause stutters every respawns, P2K running around killing bots.
Note1: The analysis of frametime differ by pictures, make sure to enlarge and look properly at them, because one looks terrible and wide doesn't mean it's in the same frametime results.
Note2: I forgot that 144Hz was my refresh rate during the testing, this will affect uncapped V-Sync results slightly compared the 120FPS cap however by the ratio of results, we can easily tell that it's no better and would be worst if I had actually been using 120Hz uncapped with V-Sync.
Uncapped, Vertical Syncronization Off Uncapped, Vertical Syncronization On Capped with DXtory, Vertical Syncronization Off Capped with DXtory, Vertical Syncronization On Capped with fps_max, Vertical Syncronization Off Capped with fps_max, Vertical Syncronization On Capped with RTSS, Vertical Syncronization Off Capped with RTSS, Vertical Syncronization On Results are fairly conclusive in term of consistency and if I may add, responsiveness even though that driver/third party cap does add a little bit of button to pixel lag as you may be able to see further down.
RTSS still remains the best alternative out of all options based upon the graphs, 1%'s and 0.
1%'s as those literally are the show case of how consistency your overall framerate will be.
From what we can see here, Vertical Syncronization has little to no impact, if not do better than without.
Of course this is all within margin of error in such testings.
Only the uncapped syncronized results do show a constant stutter which is the input lag that anyone would feel, it isn't as bad on 144Hz, but it can still be felt.
In this case, you're probably better off playing with V-Sync for the full frames that would accurately provide you gameplay that you can rely on.
I've also done the exact same tests in Overwatch, the patterns were the same along RTSS being better.
Cinque TerreCinque TerreCinque TerreCinque TerreCinque TerreCinque TerreCinque TerreCinque Terre

Input Analysis

Is there any downsides? As people often make the statement more FPS "CAN" can fill sequences to make your mouse feel smoother because of the early frametime "BUT" you still need this framerate to be stable and be seen for bigger effect, you are still limited to the amount of FPS your monitor can display thus not beneficial as much since the ONLY benefit comes from if your framerate is consistently at said high number for the response time to benefit from it.
Next, having more FPS can help the engine process information faster, but that amount goes as far as 2ms in best case scenario, therefor microscopic benefit.
Mostly because those framerates have to be "ridiculously high and stable" in the first place (400 FPS for example.
), there is more factors to worry about too.
A higher refresh rate will make all the difference better in general.
He also tested with different GPU tiers which can greatly affect the 1% and 0.
1% lows due to how much more stable of a framerate a faster GPU can output.
Higher lows = More frames to keep the game responsive and smooth.
Hence why G-Sync seems better.
Some comparisons; We can clearly tell he did not change the pre-rendered frames that is required to use V-Sync properly on Nvidia and uses 143 instead of 141, this statement will explained further below in this guide that the higher the framerate, the more finnicky it gets, 142 is the required amount to enjoy V-Sync due to how sensitive it is, pretty much like Freesync/G-Sync clearly he had to avoid consistency in his tests as you can see one older graph below with 60FPS caps.
By official statement already, he said that going below 2FPS is required to avoid FreeSync/G-Sync threshold, V-Sync needs it too but in the case of nvidia, you need to substract your refresh rate to the pre-rendered frames which is 3 by default, so 141 fps cap.
(On AMD 143 fps could work since, flip size queue is max pre-rendered frames of Nvidia and is locked at 1.
) It also varies by refresh rate for all 3 features, the higher means you need to cut the framerate further away while the lower is more forgiving and can be closer to it.
As we can see, V-Sync nearly has no lag, if he had done it properly he wouldn't have had any, it also shows how Driver fps capping is terrible, G-Sync actually suffers from it by not only that but also that G-Sync by default has V-Sync on to prevent the framerate from exceeding the monitor which obviously prevent tearing but not the lag, Driver fps cap just adds up to the issue for being terrible and not being done properly by him.
More comparisons featuring 180Hz; Windows 10 Game Mode shown to improve button to pixel depending on the game; Indirectly proving that V-Sync has literally no input lag whatsoever if capped below the pre-rendered frames(Nvidia default is 3 so problematic)/Flip Size Queue(AMD default is 1, so irrelevant) and because AMD is 1 unlike Nvidia's result above.
We see no input lag caused by V-Sync whatsoever, see below for more information and why; He claims that V-Sync is "NOT" enabled while FreeSync is on, those claims are complete bias as FreeSync behaves better with V-Sync enabled, that's because of the complete frames assisting the monitor, the same can be said for G-Sync because that's how Nvidia prevents the framerate from exceeding the refresh rate because otherwise if exceeded G-Sync or Freesync will turn off since you can't Sync more frames than your monitor can output, the same can be said for lower ranges, if reached the monitor will also stop syncing altogether, V-Sync won't stop.
Those facts are confirmed by AMDmatt, an AMD employee, despite claiming that FreeSync removes the bad side of V-Sync, if it can be active at the same time, that means V-Sync does not get any input lag to begin with if kept in the range of the refresh rate, as said above he indirectly proven that V-Sync does not cause lag.
Further facts that ray_m, another AMD employee contradicts AMDmatt by saying it should be disabled clearly indicates that V-Sync is still an instance of it's own and Freesync just like G-Sync is just a physical monitor feature that adjusts to your framerate signal as long you're in range.
Source of AMD employee claims; https://community.
com/thread/213346 Showing that third party caps add "some" lag which is pretty much unoticeable unless you test it with tools as he forgets to mention how stuttery engine framerate caps can be that can be shown in the frametime analysis, this is where a trade off has to be made; Older input lag tests he did and those results are far more different than the newer one, very very odd and suspicious to be that much of worst even though it's all the same thing.
Showing why you shouldn't play any windowed mode, even though there is no real difference from Windows V-Sync being tied to the game, he does not mention the desktop variable framerate which breaks the smoothness, unless you use multiple monitors, never play windowed; Even though I tried to like this guy, he only promotes the technology in a rather biased way instead of researching in place and actually test V-Sync and frametimes properly, his button to pixel lag might be accurate for information, but his statements over the technology are not thoroughly done and misses are lot of factors such as frametime analysis and panel issues or just V-Sync in general, the fact that he states FreeSync to disable V-Sync until it reaches the maximum framerate is completely false, V-Sync full frames helps FreeSync to work better consistency wise which is the same for G-Sync, while engine framerate caps are usually stuttery based upon the upper frametime analysis I've made and he still denies and doesn't test V-Sync all alone and properly to evaluate, based on many witnesses and tests, all reports that it's impossible to tell or that their seem to be some sort of different feel to it, from game to game it also varies, Overwatch for example handles lower framerates poorly compared CSGO.
Source: Battle(non)sense https://www.
com/watch?v=mVNRNOcLUuA https://www.
com/watch?v=rs0PYCpBJjc Facts points toward that V-Sync actually works equally as any other syncs in terms of input when properly used, however this does not cover everything, further below there will be all the information for you to know about the consequences of such tweaking.
Looking at the V-Sync results, you can see that without a cap, the button to pixel lag is increased meanwhile that with FreeSync + V-Sync, it is literally within margin of error which is not FreeSync to thank, but that even without FreeSync, V-Sync would behave perfectly fine, G-Sync is also no exception.
Fun fact, the brain takes 13ms to process and varies by individual.
Cinque TerreCinque TerreCinque TerreCinque TerreCinque TerreCinque TerreCinque TerreCinque Terre

Other Types Of Syncronization

Starting off with.
Enhanced Sync / Fast Sync This method of Syncronization is much more passive, but not perfect either due to variables of framerates.
While not stopping the framerate from reaching above and beyond the refresh rate unlike V-Sync which prevents input lag altogether, it will only work perfectly if you can stay set at the refresh rate.
It is recommended to use this method of syncronization with an exact cap equal to your refresh rate for the best benefit, otherwise problems will occur, those are dependent if you are above or below the refresh, so make sure that you keep it check.
Above the refresh rate it'll result in a forward skip, while below it'll be like V-Sync dropping a frame.
Adaptive Sync It claims to remove the "stutter" by disabling V-Sync when the framerate is below the refresh rate, tearing comes back which defeats the purpose of it, that being said before, the stutter is a trade off from tearing, you get one or the other, but remember the stutter or tearing doesn't not affect input! Unlike the forceful aggressive V-Sync that turns back on once your framerate is even to the refresh rate, bringing back input lag.
This method is not recommended in anyway.
FreeSync / G-Sync These types of syncronization while costing a new monitor in which G-Sync turns very expensive are also nearly perfect in what they do, their only downside comes from the fact that drastic frame drops will slow down the refresh rate, which you will see the light, or brightness start to slow down as well or come back up again, it can truly be annoying for some hence why it's recommended to make sure you keep a consistent framerate with this technology to mitigate this issue.
Playing any games that are prone to heavily drop in framerates like MMO's are recommended to use either v-sync or playing borderless windowed.
Because these drops will keep the technologies go on and off constantly which will be extremely distracting.
However if you do not mind it, there is this other feature that comes with it that whenever your framerate starts to slow down and still be within the range of the sync, it will compensate for these lower frames and predictably clone them ahead of time to give an illusion of smoothness.
Put in perspective, you don't need any types of sync when using those.
Just make sure that you keep your framerate in check and at affordable amount, otherwise the unpleasant monitor light flicker will occur.


Q: Will capping your framerate to the monitor refresh rate fix the tearing? A: No it doesn't, framerate is still out of sync.
- Q: Will lower framerates reduce tearing? A: No, the reason why at higher framerates there is more tearing in it's sense is because you have many frames to show, revealing a lot in smaller amounts.
While at a lower framerate there is less frames to display, but being out of sync they will become more noticeable in a large image because you have less frames.
It gets worst when going around geometry in-game, such as corners or objects because of their very distinct sharp shapes.
- Q: Why should I do this though? Even if I don't care about tearing.
A: Sparing GPU/CPU load has it's benefits, but the biggest of all would be that never your game will have FPS drops anymore (That is if your computer can handle that much.
), on the flipside will allow you to use that spare power to do multitasking, recording or streaming.
Quite a lot of things you can do when one process is not eating up all the performance of your computer.
In the end, your GPU load will be minimized and increase in perspective of what it needs to maintain a solid frametime.
- Q: I've heard V-Sync pre-renders frames, therefor holding the action back? A: Not really, V-Sync literally SYNCS full frames.
This idea comes from the fact the game gets delayed when you run V-Sync uncapped and get mouse lag, when what truly happens is the process trying to fit more frames than it should in the refresh rate.
All GPU's have a least 1 pre-rendered to display visuals that you witness no matter what, Nvidia by default render 3 frames at the same time, trying to fit into the limit 63 fps over 60Hz if you were to cap at 60 fps, while AMD pre-rendered frames which they call flip size queue is 1 frame which in the same case of Nvidia is responsible for the input lag because the only way to display those frames is put them ahead in the next refresh causing everything to be delayed behind by the amount of pre-rendered frames.
What truly happens behind V-Sync is that the GPU is asked to render 1 complete frame per back buffer so that the image is vertically perfectly synced and doesn't end up with multiples of different frames at once, causing this tearing.
- Q: Should I use Triple or Double Buffering? A: Triple is claimed to provide better visuals and in all it's factual state, using Triple prevents the framerate from choking because V-Sync in it's design has to provide frames as the monitor refresh, therefor the ability of having 2 back buffer instead of 1 gives more headroom to the GPU while swapping to the front buffer unlike Double which only has 1 back.
Hence why some games/engines have FPS cut in half when using Double Buffering when the framerate is not stable, in that case using Triple Buffer fixes the issue.
However CSGO or the whole source engine in general doesn't seem to cut it's framerate in half.
In my personal testing, only OpenGL would cut the framerate in half, this explains why GPU panels only states OpenGL Triple Buffer, DirectX games simply never would do it.
Not to be mistaken, even if you use Triple buffer with it's two back buffers and 2 pre-rendered frames on Nvidia, it will not prevent any input lag as the reason why it occurs is because the framerate is still going above it's respected 60Hz.
Those are apparently complete separate and you need to always keep your framerate cap minus the amount of pre-rendered frames.
(As seen in Nvidia Inspector when using framerate caps, the owner has reported better results with lower framerate caps.
) - Q: Won't having too low of a framerate make the game stutter with V-SYNC? A: You're gonna get the same laggy slideshow as uncapped, except without tearing.
Also to be aware that the lower the framerate, the higher the frametime, meaning that the game will be a lot less responsive, with or without V-SYNC.
- Q: Will having your frames to be rendered in a lower frametime benefits you in term of reaction because you will receive information earlier? A: Hardly, you will receive information accordingly to what your monitor can do and what the server gives you.
Aka respond time, Hz and Ping.
Here's a table displaying a few results.
(All rounded up.
) Time/FPS-Ticks-Hz 100.
0ms/10 33.
3ms/30 16.
7ms/60 < A standard that has been for years.
6ms/64 15.
2ms/66 13.
3ms/75 10.
0ms/100 8.
3ms/120 7.
8ms/128 6.
9ms/144 5.
0ms/200 < Many of todays default of modern engines.
3ms/300 < Default fps_max of Source Engine and *New RTSS Maximum Limit.
(See * for more.
) 1.
0ms/1000 < Beyond that enters the micro/nano seconds.
(Calculate frametime by dividing 1000ms by X amount of FPS.
) Uncapped framerate does cause horrible frametime stutters that uneven the framerate itself very unconsistently.
If your setup is able to maintain at the high refresh rate, it is more than okay to let your framerate reach up to it's refresh rate a frame below and enable V-Sync/G-Sync/Freesync.
Remember also that the higher the framerate, the more likely it is to be inconsistent, make sure to evaluate your hardware.
- Q: Overlocking my monitor won't break it? A: The worst that can happen is going out of range.
Make sure you don't permanently set it that way, otherwise you will have to reinstall your video drivers through safe mode to fix the issue.
- Q: Does this work with other games? A: It depends if you are able to change the in-game refresh rate, source engine games requires "-refresh [enter your refresh rate]" to be set in set launch options to work, while other games can be done in the video settings.
- Q: Can you do this with lower framerates? A: Yes! If one of your games doesn't run well, but can maintain for example 30 FPS, you can cap your game to the said amount to achieve stable gameplay.
Framerates fluctuating from 60 to 30 constantly is very noticeable and can hurt your own performance.
- Q: Playing in windowed mode removed the tearing and I have uncapped, should I use it? A: Never play in windowed mode unless you have multitasking needs or dual monitors, the monitor will be tied to Windows V-SYNC which can hurt the smoothness of your gameplay even when you have a very high framerate.
It is also worth knowing, V-SYNC in-game doesn't work in windowed mode - Q: The game I want to play doesn't support V-SYNC, what can I do? A: It's possible to force V-SYNC with the GPU panel.
AMD drivers are known to fail at this, therefor D3DOverrider which is another RivaTuner Software is a good alternative, however is often prone to crash games if alt tabbed.


Some monitors don't always follow specifications in a logic order, one monitor could be 75Hz and still have over 16ms of response time, be sure to check out your monitor specs to figure out what could cause you issues, because if said case above your monitor is 75Hz but 16ms response time, it'll mean that you will have the respond time of 60Hz with the visual of 75Hz and likely bad ghosting, making it laggy still.
Refresh rates can and will affect the way your input works up to it's maximum frequency based upon the framerate you have.
For example, going at 60 FPS cap on 144Hz for a very demanding game can yield better results than 60Hz 59 FPS cap.
It's always best to hold the highest refresh rate to have the best results At last, not all games V-Sync is build equally, some are badly implemented while others are flawless, Driver forced V-Sync can address this issue however, but it'll always be best to use engine V-Sync for the sake of compatibility.
G-Sync/Freesync are better alternative however they may introduce other type of compatibility issues, also the price is another factor.
Monitor syncs also have the unfortunate downside of flickering at lower or huge variable framerates.


Alright so if anyone has more questions, feel free to drop them below in the comment section.
This guide intend was to spread awareness about something people are wasting their money on and with generousity improve their gameplay experience to something consistent and enjoyable to witness.
Check out my other guide related to improving your gaming experience; http://steamcommunity.
com/sharedfiles/filedetails/?id=749450225 http://steamcommunity.
com/sharedfiles/filedetails/?id=816584566 Necessary softwares.
Refresh rate tweaks, will require restart[www.
com] Otherwise your dGPU panel can be used to change the refresh rate, which does not require restart.
dGPU overclock, pre-recording, recording and FPS cap (MSI Afterburner bundled with RTSS responsible for FPS cap.
com] RTSS without MSI afterburner[www.
com] D3DOverrider Standalone[community.
com] DXtory, an alternative recording and FPS cap software[exkode.


FreeSync vs G-Sync Input Lag Comparison by LinusTechTips, yet still misunderstood.
Blur Buster topic showcasing issues with max pre-rendered frames and results, however does not showcase fps_max/engine inconsistencies[forums.
com] .
Vertical Syncronization understanding by Anandtech.


Thanks to this article posted here.
at] Confirming that this solution has been found out about more than just by me years ago.
And many people who gladly tested these discoveries along with me.