Yeah I thought this video was going to include all the things and be a one-stop-shop for "watch this to know everything, send this one video to people instead of having them literally go through my entire channel to be told about things"
@WAUZZZI iirc correctly last time he said capping framerate to something you most frequently is better than a crazy framerate. So if you can hit 240hz stable cap it at 240fps versus having it unlocked and sometimes hitting 300hz and dropping back down to 240hz since it creates inconsistencies and creates higher or lower input lag on your end.
@@unkkit151 he already had a vid on that. Ingame cap= lower inputlag, more variance on frametime. Rtss = medium inputlag, perfect frametime. So for competitive settings its better to use the ingame cap in most games.
@@ValkenDnB exactly. uncapped fps = gpu 100% = bad input lag. Capped fps at stable frames = gpu never exceeding 95% = good and stable input lag. This last point is basically what NVIDIAs new inputlag feature does. By throttling the gpu at 95%, so it never goes over.
I love these videos. Exploring content that is niche, but really shouldn't be. The video was very clear and easy to understand information, whilst still going into great depth on complicated topics. Great channel!
That was one hell of a clear explanation 👌 I really loved the description of V-Sync. I knew what V-Sync was about before this video, but the way you illustrated it was just beautiful!
Would love to see you use this to find optimal ways to lower the variance in latency at various levels. Since not everyone can afford a rig that can do more than 144 fps
@@SOMEONE23145 System latency =/ internet latency. Two totally different beasts that can not be compared what-so-ever in a game like Warzone :p I'll take 10ms system latency with 80ms ping over 80ms system latency with 10ms ping in a game like Warzone any freaking day of the week, especially with all that lag-compensation peeker's advantage stuff that happens in every online shooter ever (yes, I'm also looking at you - Valorant).
@@SOMEONE23145 Also worth mentioning is that between the best pro players there are a significant difference in reaction time that exceeds both 10 ms system latency and 80 ms ping : ) Speed is just one factor for a good player. Tactics is another one. Adding skilled vs non-skilled players to that and even a (probably) very controversial statement when I claim that quite a lot of the really good players benefits from netcode related lag giving them a big advantage in TTK and TTD when compared with same/similar weapons and play. Now, I'm a console player so this may be a bit different for PC gaming (?) but probably not that a big difference since we share the same basics. If I didn't have thousands of gaming hours I would never dare claim this, but experience, understanding lag/netcode better (thanks to battlenonsense) and by spectating these players I clearly see that aiming skills is not what makes them great. Certainly better than an unexperienced player, but when seeing how few shots they actually hit the enemy and how few they need to kill makes me pretty confident about my claim. This is a topic I think should be discussed more but it's probably a much too sensitive subject, as no one usually responds when I try to discuss and understand this better.
thats why there is no point anymore in spending money for competetive gaming. hardware companies need game developers to sell their products, but game devs dont care about anticheat politicy anymore, so there is no point in buying games. so hardware companies cant sell their products. they have to work together and we have to put a sign as consumers. blizzard games are cheat free, while in csgo are 90% cheaters. its all about decisions.
Pretty amazing results! Didn't expect such big variations in 144Hz and below. Seems like sample size matters and clearly gives an overall better view on these results. Thank you for your hard work! I am already on a 165Hz GSync Monitor, maybe switching to 240+ in the near future for FPS games.
I've suscribed right now to your channel because you've this particular, clever pedagogical method of going to the point. Thanks a lot for the quality and intellectual honesty of your presentation. You make us earn one thousand of hours of checking some unintelligible others pretending experts. Best regards with all the wishes for this year.
Thank you very much for the visual image of what is happening when FPS exceeds Hz. I was so confused as to how this actually went down so I just could not understand it. The visual was extremely helpful.
So basically, refresh rate does more to reduce variance, while FPS does more to reduce the average values, but both things have both effects, just balanced differently.
thanks, that is a huge work. but i suppose the fact that more hz and more frames(above hz) is better for input lag is pretty obvious to most of your followers. can you please test more complicated and interesting scenarios as well? for example: 60 fps with 144, 240, 360 hz, cpu bound with different sync and antilag options capped vs uncapped, some tweaks like hpet, timer resolution, dram overclocking (both clock and timings), etc UPD: oh i see, there's much more detailed comment with the same idea. guess i should leave this for TH-cam algorithms anyway
Can you please do a video of "fast-sync"? According to nvidia at least that should allow to run games at unlocked framerate, while not getting tearing as they use extra buffer where rendered screens go and then front buffer which updates at the rate your monitor does (basically unused mid-buffers are just thrown out). Or at least that's my interpretation of what i've read. Would be really interesting to see proper tests on this feature, as it might help people with lower refresh rate monitors to both run games faster and avoid tearing.
I would love to see some videos on popular "in progress" games like Escape from Tarkov. Just to get an idea about what optimization really means for games. I always feel like my input lag is so high in that game. Would also love to see how stable the netcode and fps is from a more official source like yourself.
That scatter plot in the end comparing 60hz, 144hz, 240hz, and 360hz looks like a visual indicator of how it feels to game on those refresh rates. How the jump from 60 to 144hz is the biggest, 144hz to 240hz is significantly smaller and 240hz to 360hz has a very minor improvement.
This is a good visual presentation and explanation for people who are not aware. Another video that confirms what I wrote in my blog 3+ years ago and another good, quality one to add there. Keep up the good work, Criss. You are the only youtuber that make this kind of content and explain it very well. I wish I had the Nvidia tools too. It would be nice to see a comparison b\w CPU and GPU bound with relex On. Im interested to see is there a slight delay with Reflex. Maybe the next video? You, already have the CPU bound results from this video. Btw competitive FPS players play on lower resolutuion so they are cpu bound and they dont need Nv Reflex but fast CPU
Thank you for your super good videos and content I love watching your videos! And I hope you sleep feel good with this channel and aren't stressed out or something! You have great fans and followers 🙈☺️
This makes so much sense now. I have 144hz monitor and I always cap my FPS at 144, even though I can run VALORANT at 250fps, leaving it uncapped while I have 144hz makes the game feels so much worse, stuttery and not clean or smooth at all, even with high frames. I guess next step for me is getting 240hz monitor. Also off-topic question. Should I use nvidia reflex if I cap my FPS at 144 and my GPU usage isn't even 40%. I watched your video about it, but could never figure out whats best in my case. Cause I was using nvidia reflex but it felt like it had no impact, or it felt worse at times.
Uncapped framerate will always be unstable as the CPU and GPU are trying its absolute hardest to pump out as many frames as possible. Capping it at a framerate your system can always handle even in worst case scenario should be your framerate so if you think 250 is consistent for you and it barely drops from that value then go for it, obviously you'll have the screen tearing side effect which I wouldn't be too worried about, especially in VALORANT where reaction time is important. As for the Reflex, the feature doesn't usually work when the process is CPU bound, there was some setting you had to turn on which I don't remember (sorry) to make it more GPU bound. You could also just increase the game's video settings to make it more GPU bound
When I had 144hz, I used to cap my framerate in ow to 200 and disable vsync. There is a slight tearing, but not that bad. Now Im using 240hz, and I cap my in game fps to 236 and enable freesync, its really good, smooth, no tearing and extreamly responsive. I always cap my fps so I can have a more consistent latency.
Reflex should only be used if ~99% of your GPU is being used. It can actually add input delay if you use it while not GPU bound. In general, if you cap your framerate to a number that your system can sustain, you should leave Reflex off.
Use however you feel best I guess, but if we go by the chart, then setting your fps higher than refresh rate means better consistency (but more tearing).
This is...EXTREMELY useful. Exactly what I was looking for for deciding whether limiting fps to the refresh rate is worth it the sacrifice; i.e., decreasing system latency (by decreasing fps) to stop tearing. Capping fps IS worth it.
@@rashidkhan-xm4rl No. Always cap FPS. Shooter games have another problem: different mouse sensitivity at different FPS and also the higher the FPS, the more unstable it is. It's very easy to test: either get an Arduino or any other board with mouse capability, or get a program that can emulate mouse inputs. Lock your game at 30 FPS. Then make it move exactly 360° at 30 FPS(I got arduino and did something like 3627 mouse moves by 11 points in the x axis to rotate my char by 360° in CSGO at 30 FPS). Now capture FPS at 300 FPS and make the program do exactly the same thing: send the exact same number mouse moves at the exact same distance. What you will observe is that while at 30 FPS your character moved exactly at 360°, at 300 FPS, your character moved either a little more or a little less than 360°. Conclusion we draw from this behavior in tests: FPS affects mouse sensitivity. Now, unlock your FPS and let it be as big as possible. Let the program move your mouse again. You'll get *different* mouse travel distance each time you do it, because of the inconsistency of FPS. Conclusion: lock your FPS to get stable sensitivity in shooter games, so you get better aim. This behavior is repeatable in any video games that I tried: CS 1.6, Source games(Half life, CSGO, whatever), UE4 games(Valorant, Squad, whatever). Why? Well, it's because of the calculations rates and how floating point calculations are done. You get a different result if you move your in-game view 10 times by 1 mouse move point multiplied by your sensitivity, compared to if you move your in-game view 1 time by 10 mouse move points multiplied by your sensitivity, because of how computers calculate and store floating point numbers. And when you do it 3000 times over, the errors accumulate. And so when you're doing it with unstable frequency, you're getting very unstable results. Once you lock FPS, sensitivity becomes more stable but still has errors because of how calculations are done. The proper solution to this problem is to always move the in-game view X number of times by 1 mouse move dot. So, let's say you locked FPS at 60 and you moved your mouse by 10 dots at the period of time. Right now CSGO will move your view only once at the next frame by 10 dots. Instead, it's better to move the in-game view 10 times by 1 dot. That way the sensitivity will become the same at any FPS: 30, 60, 300, unlocked unstable FPS, whatever.
So which is better? Pump as much FPS as possible for better performance at the cost of tearing or cap FPS to refresh rate to remove tearing but at the cost of performance? And what about whether to use vsync/gsync? One important takeaway from this is the diminishing returns in gains. It seems the jump in improvement from 240 to 360 is negligible compared to 144 to 240.
Tearing over 200fps ain’t even an issue, vsync for no tearing is good for 144fps and below where it’s noticeable otherwise push that mf hardware for the newest frame asap aka uncapped
@@Bl00dMalicefree sync/ gysync I believe adds barely even a mili second extra and paired with vsync on the driver level (as long fps doesn’t go over max refresh) it’s negligible to say for the benefit of no tearing. uncapped with vsync off has been noticeable better input lag to me with no visible tearing atleast in apex coupled with anti lag/ reflex on per your gpu
Great Video Mr. Chris. I do have a question for you. I have seen alot of videos of increasing gaming performence by modifying windows 10, unparking cores, deleting windows apps and so on. A few ytubers like Fr33thy, lock outz have alot of these vids up. Was wondering if you have done this to your systems or even tried these modifcations at all, do they even improve your gaming performence? Again great video man.
Gotta admit, I'm still confused about what the best combo is for the best input lag, frame pacing, and odd refresh rates. I have a 180Hz capable monitor for instance, and while I know enabling v-sync kills latency, most games don't have a fps limiter with 1Hz increments, so you can't just go down to 176Hz or even below just standard presets. Still not sure why you need to go 4Hz below your maximum refresh even with v-sync off either, or if in-game caps will fix frame pacing issues. And it gets even more confusing when you see tests that show uncapped framerates above refresh rates making latency worse, or putting an fps cap on games without internal limiters through something like Rtss can be worse than just turning v-sync on. Or hell, blur busters says to leave v-sync on with G Sync, and that rtss adds no lag, and other tests say otherwise. I swear, this stuff all needs to be standardized for consistency already.
I completely agree. I feel the same way. I’ve spent more time tweaking my settings on games than actually playing. I just wish there was a golden standard to go by. I also wish there was more info/testing with Reflex and Reflex + Boost.
@@J_Karnage I don't know why i wrote that, i meant to say: create a custom resolution in the Nvdia Control Panel and use custom 176 Hz (can use whatever value you want from 1-180 Hz). Can't create refresh rate that already exists in the preset. Values above 180 Hz could damage the monitor but probably 181 Hz is safe to test ( i don't recommend to do it tho). About the fps limiter, the game must have a cfg to edit. Example: in Wolfenstein: Enemy Territory you open the console and type /com_maxfps 125 for example or in the configuration (cfg) file.
Hi! I would just like to appreciate how your graphs are so professionally made and easily understandable too! I do would like to know which tools aside from NVidia's LDAT did you use to help you with the data visualizations? Thanks!
Hey Chris, great work! You tested framerates over the refresh rate here, but another great comparison would be the difference from refresh rate alone - e.g. 60fps at 60hz, 144hz, 240hz, 360hz. With that data we would be able to isolate the impact of refresh rate, the impact of frame rate and the impact of both together.
You shouldn't lock a game to 60hz with monitor at 144, the stutter alone will make you play worse. If talking about multiples, like 60 fps on 60, 120, 180, 240 etc, there shouldn't be any difference in input lag, the display will simply repeat the same frame 1 2 or 3 more times, at least in the best case scenario when using some kind of sync. Running completely unsynced would make a higher refresh rate one better though, and cause tearing on some of the frames and not on others (since the frame changes less often than the display). Anyway, running locked 60 fps on 240hz monitor is a bad idea and you should invest your money elsewhere.
@@rtyzxc you hear about Prerendered frames? if you have 60fps the frames are 1/60 = 16.66 ms "longer" and if you multiply that for the number of prerendered frames you have the total inputlag, 60hz with 300 fps have much less input lag than 60hz with 60fps though vsync on(only fast sync allows framerates higher of refresh rate of the monitor). PD:SORRY FOR MY ENGLISH. Dont forget fast vsync is triple buffered and this introduces some input lag too. That's is why if you lock your framerate below your refreshrate you have better inputlag because prerendered frames = 0
@@rtyzxc "You shouldn't lock a game to 60hz with monitor at 144, the stutter alone will make you play worse." Zero refresh aliasing with adaptive sync any frames at any rate slower than the maximum of the monitor. "there shouldn't be any difference in input lag" There's a very sharp improvement in input lag due to the faster scanout (the time between the frame reaching the front buffer and being fully refreshed onto the monitor is much lower) but it would be great to quanify this (standard deviations etc) with the new tools that allow for sampling hundreds of times. "Anyway, running locked 60 fps on 240hz monitor is a bad idea" 60fps was just an example rate - one that every monitor can do, therefore it's good for isolating the impact that refresh rate has on latency across the entire range of monitors from 60hz to 360hz. Locking framerate below the maximum refresh rate of the monitor - sometimes quite far below - is actually a very legitimate thing to do on an adaptive sync monitor, depending on the game. Other games cannot achieve higher framerates due to CPU/RAM limits, but still experience a reduction in input lag and an improved response consistency when upgrading to a higher refresh rate monitor.
@@AerynGaming I thought you were talking about running fixed refresh rate above 60, now you are talking about adaptive sync, which will of course run the game perfectly at 60 fps, 60hz.
the data at 2:05 follows a periodic, upward sloping latency pattern. Does anyone else notice this? Three distinct groups of data clusters are visible. Weird, ingame rendering oscillations?
Can you do a video on the effects of Nvidia Reflex (on vs off) on input lag / system latency on high tickrate games (CS, Overwatch, Valorant) vs. low tickrate games (Destiny 2)?
I'm new to input delay analysis. Can someone clarify what is actually happening? For example, I'm confused as to whether input in online, player to player games is being relayed to the server, or other player, before it is displayed. Does the visual even matter or is that somehow synonymous with the baseline P2P transmission starting point? I'm familiar with net received delay across web. There are so many issues. In COD a simple "hit scan" firing point doesn't work with all these projectiles. What happens with aim assist as an enemy's frames slow down, motion is interpolated, and impact position is not where it would normally be due to delay? Add actual input delay with now muddied aim assist... I once saw a Valve dev article talking about interpolation making extremes of a bouncing ball impossible to hit despite visual presence. This is similar to the dreaded snaking.
Me too, I'm curious, is 'input lag' just 'visual lag' or does the computer not process any user input during the frame gap? But I kinda think this will be based on each individual game
@@iXpertMan I originally thought it was actual device lag, in the way that pro COD events were using Cronos max to lower that side, but then I saw the delay being measured on same side displays, with their own turnaround time, AND heard that some games update local state only after server side confirm which just sounds insane.
I believe he’s measuring delay between input from mouse or keyboard + system latency + display latency. So the numbers shown is only client side from mouse click to action on display. Nothing to do with server latency or lag compensation.
I'n most games your client will show appropiate responses aswell as actions on its own before it verifies with the server, thats why you can see yourself shooting at enemys in games but the shot not connecting when you are lagging.
@@pandasdreamlygon actually he didn't if you look. I would like to know also what would change if run a game at 144fps on a 240 hz monitor vs a 144hz monitor.
Great video, thank you for your efforts! If I have a 144 Hz monitor, should I enable my games to go past 144 FPS to get it smoother, or limit it at 144 FPS to prevent screen tearing?
Thanks for your videos and effort , greatly appreciated . But i'm completely lost between the different advice you give (i'm sure it's my fault) , i mainly play valorant and i have problems with packet loss . I've tested the method of capping the fps , and it seems to work fine for that particular problem , but now i don't know if i have to cap it to the max fps my pc can sustain , or the 4 fps below the refresh rate of the monitor (144 hz in my case ) . I have an 8700k with a geforce 1070 , i can spikes up to 330 fps and dips to 170fps sometimes . So i don't know if i should find another solution to the packet loss , or just cap my fps , and if so to what number
So if my pc cant handle consistent 240 fps but i have a 240 hz monitor ,would it be better for me to cap my fps to 240 and have it drop a lot or do i cap it at 144 fps so there is more delay but its more consistent?
I would like to add that in games like BF4 not sure about BFV running a multiplier of the server tick rate is also helpful. one more mention is run 3 FPS lower than your G-sync range for better frame times.
What about a reverse situation where you have a 240hz monitor but only get 144-180 FPS in a game. Why does it feel less smooth than lowering that monitors refresh rate to 144hz?
This is just what I am looking for and very helpful, thanks. Now I wonder what FreeSync/G-sync do to the latency. I believe it should reduce the longest delay by the inverse of the refresh-rate(e.g. the latency should be 10.3 ~ 16.3 ms instead of 10.3 ~ 23.2 ms at 144Hz/144FPS with FreeSync on).
_" believe it should reduce the longest delay by the inverse of the refresh-rate(e.g. the latency should be 10.3 ~ 16.3 ms instead of 10.3 ~ 23.2 ms at 144Hz/144FPS with FreeSync on)."_ No. Normally, the monitor just shows the most available frame, and the minimum and maximum delays are at their absolute minimums. There is no syncs to do to decrease them. VSync delays rendering frames until the new monitor refresh is ready to start. GSync delays monitor's refresh until the new rendering frame is ready. GSync can't reduce neither longest, nor shortest delay, because of the way it works: it actually delays the refresh of the monitor until the new frame is ready. So considering you'll have to lock your FPS at 236 when you play at 240 monitor in order for it to work, instead of playing at 300 FPS or 700 FPS, or whatever, you will have bigger input delay with GSync compared to if you played without GSync with FPS locked at a higher number. What GSync does is that it fixes screen tearing and makes input delay more stable in the same parts of the monitor: bottom of the monitor always has bigger input delay than the top of the monitor. And it's not worth it, it's just another marketing trick to make you pay $700 for a monitor instead of $300, and still get worse results for what they advertise it for: gaming. I'll take a $300 240 Hz monitor without the GSync and play at 300 FPS instead of a $700 240 GSync monitor to play at 236 FPS any day of the week. And have smaller input lag all over the place on the $300 monitor. And save $400 while I'm at it, too. I'm playing 300 FPS on my 260 Hz monitor and I never see tearing. Never. And I looked hard. It's there, but it's so small at 300 FPS and so inconspicuous, that all these syncs are absolute nonsense as a fix for it.
What you need to understand is 60fps at 144hz will no longer feed your monitor old frames. because your refresh rate is faster than your gpu output so it will show you the most reason frame.
I've known this for a while being in the competitive gaming community but there's a lot of people that think it's pointless to have more fps past the refresh rate. Your videos are informative as always
i have spent all week trying to fix my Stutters thinking something was just wrong with my windows its must be the QPC timer or something turning off background services updating rolling back updates new drivers installing a old version of windows none of it worked why was my 9900k and 1080 running worse then a 2600k and 1650s turns out 1080s just run like dogshit worse then any of the Turing cards if you don't cap there fps i watched this video turned off NULL turned on Gsync and then ran fps_max 235 in csgo and the Stutters are gone
At 8:04 we can see two kind of separated sample groups at ~12 and ~16 ms -- with the 12-ms-samples from sample ~30 to ~130 and the 16-ms-samples especially at the end ... is this an artifact of some kind? (blue dots für 240 Hz and 240 FPS) The other samples seem to be in normal distributions, this is kind of bimodal.
One thing to note is that the worst case input lag at 60hz decreased by 1/60th of a second when going from 60fps to 360fps, a result that checks out on an intuitive level.
Would you like to check the impact of SMT or Hyperthreading on Input Lag? Could be Interesting. In general Bios Settings :) Of course with ingame limited frames. The lower input lag by the higher frame amount with smt we can just calculate and see if it's worth the input lag it gives :)
Doesn't capping a frame rate also reduce system latency? If your system can output 160 fps, but you cap at 144, pegging your frame rate to that cap, the decreased pipeline also gets drawn frames to the monitor faster?
If you have a refresh rate higher than your fps, it will not improve system latency, it will increase it actually. You want a fps as close as possible to your monitor's refresh rate to reduce frame timing to the nearest lowest ms. Low fps and high refresh results in higher frame time which means it lags. For example. I have a 144 hz monitor and at 144 fps my frame time ms is 7.2. at 80 fps, I get 16-20 ms for my frame timing which is lag that is detrimental to competitive pvp fps games.
The problem I have is that my mouse and keyboard strokes don't always register and my mouse feels slightly 'floaty'. It's been like this for some years now I've been unable to find a solution and I'm running an RX6900XT and get around 200-300 FPS.
What about games that are capped at 60FPS? How different is 60FPS@60Hz vs 60FPS@360Hz. Maybe a dumb question, but I felt like it was the only use case that wasn't covered.
This is the most accurate video i have seen in the world, this proves you don't need 120 hz monitor to become deadly. 120fps graphics card with 60hz is good enough for casual BR games.
Hopefully someone can explain this to me. When using MSI AB with RTSS, when I display frame time, my system was getting terrible frame time spikes 60-100ms. Mainly play Fortnite. By capping my fps using RTSS (144hz monitor) capped to 140, this greatly reduced the spikes but they still occasionally occur. Can someone explain what causes frame time spikes and how to reduce or eliminate other than frame cap? When a spike occurs, the game freezes aka seems to hang. 10850k with rtx4090.
What you need to understand is 60fps at 144hz will no longer feed your monitor old frames. because your refresh rate is faster than your gpu output so it will show you the most reason frame.
@@madmax2069 the latency will be lower since you aren't being fed older frames. This is only applicable if you're running the game on exclusive full-screen mode with no sync option. Otherwise you are just running your game on windows triple buffer vsync which result in choppiness.
Why does my pc decide to run sloppy the next day no matter what I do timer resolution anything anything my pc just plays like console with high fps ? I can’t seem find an answer..I always reset it and download all drivers and everything I know and runs flawlessly but the next day or two I turn it back on it just doesn’t feel right and I have to just wait till it decides to.. same thing would happen on ps4 and Xbox ..and what pisses me off more is that even real world shit like my truck it runs great some days and the next it just feels bloated
Where do you think the sweet spot is in the refresh rate if you only have a 60hz monitor where you don't get too much tearing but get to enjoy the lower input latency?
Your 240hz 240fps data is very interesting. You can clearly see a segregation into 2 bands of response times with very minimal in betweens. May have something to do with the poll/scan rate of your hardware, monitor and game causing the visible banding of the results.
Please make a video where you can see the input lag of Battlefield 2042. I still think that it is horribly optimized and I notice input lag that is not normal in an FPS. Thanks for the work and congratulations for the channel. 😘😘😘🤩
3:30 Great explanation of what tearing is and why it happens
What a fabulous video, thank you for your research and for producing such a useful piece! Will link it in my upcoming video.
2 years later channel still huge underrated. your content is a diamond
Can you please test in different games, like Apex Legends, please?
Man you are stepping up the visuals in these videos. So much quality!
I would love to see this tested with G-Sync enabled.
Yeah I thought this video was going to include all the things and be a one-stop-shop for "watch this to know everything, send this one video to people instead of having them literally go through my entire channel to be told about things"
@WAUZZZI iirc correctly last time he said capping framerate to something you most frequently is better than a crazy framerate. So if you can hit 240hz stable cap it at 240fps versus having it unlocked and sometimes hitting 300hz and dropping back down to 240hz since it creates inconsistencies and creates higher or lower input lag on your end.
@WAUZZZI and RTSS vs built in game limiters.
@@unkkit151 he already had a vid on that. Ingame cap= lower inputlag, more variance on frametime. Rtss = medium inputlag, perfect frametime.
So for competitive settings its better to use the ingame cap in most games.
@@ValkenDnB exactly.
uncapped fps = gpu 100% = bad input lag.
Capped fps at stable frames = gpu never exceeding 95% = good and stable input lag. This last point is basically what NVIDIAs new inputlag feature does. By throttling the gpu at 95%, so it never goes over.
I love these videos. Exploring content that is niche, but really shouldn't be. The video was very clear and easy to understand information, whilst still going into great depth on complicated topics. Great channel!
A guats video wie immer
That was one hell of a clear explanation 👌
I really loved the description of V-Sync. I knew what V-Sync was about before this video, but the way you illustrated it was just beautiful!
4:00 i love how you demonstrated the part how v sync is happening. That is next level
Revisiting this video and...
Outstanding, Chris. That's all I need to say.
Would love to see you use this to find optimal ways to lower the variance in latency at various levels. Since not everyone can afford a rig that can do more than 144 fps
can you make a video about the windows setting for hardware-accelerated GPU scheduling and how that impacts system latency?
3,000 euro in hardware to lower the latency to, at the end, face a cheater in warzone 🤦♂️
Lol
But you got a reaaaallyyyy good look of the cheater :-p
@@SOMEONE23145 System latency =/ internet latency. Two totally different beasts that can not be compared what-so-ever in a game like Warzone :p I'll take 10ms system latency with 80ms ping over 80ms system latency with 10ms ping in a game like Warzone any freaking day of the week, especially with all that lag-compensation peeker's advantage stuff that happens in every online shooter ever (yes, I'm also looking at you - Valorant).
@@SOMEONE23145 Also worth mentioning is that between the best pro players there are a significant difference in reaction time that exceeds both 10 ms system latency and 80 ms ping : )
Speed is just one factor for a good player. Tactics is another one. Adding skilled vs non-skilled players to that and even a (probably) very controversial statement when I claim that quite a lot of the really good players benefits from netcode related lag giving them a big advantage in TTK and TTD when compared with same/similar weapons and play.
Now, I'm a console player so this may be a bit different for PC gaming (?) but probably not that a big difference since we share the same basics.
If I didn't have thousands of gaming hours I would never dare claim this, but experience, understanding lag/netcode better (thanks to battlenonsense) and by spectating these players I clearly see that aiming skills is not what makes them great. Certainly better than an unexperienced player, but when seeing how few shots they actually hit the enemy and how few they need to kill makes me pretty confident about my claim.
This is a topic I think should be discussed more but it's probably a much too sensitive subject, as no one usually responds when I try to discuss and understand this better.
thats why there is no point anymore in spending money for competetive gaming. hardware companies need game developers to sell their products, but game devs dont care about anticheat politicy anymore, so there is no point in buying games. so hardware companies cant sell their products. they have to work together and we have to put a sign as consumers. blizzard games are cheat free, while in csgo are 90% cheaters. its all about decisions.
Awesome video! Really informative. Seeing the spread of results really clarifies the picture for input lag testing.
Schweizer Präzisionsarbeit. Ich liebe es wie ausführlich und detailliert auf die verschiedenen Aspekte eingehst. Mach weiter so 👍🏻
Ich glaube er erwähnte mal, er wäre Österreicher, oder ich vertue mich gerade :D
Pretty amazing results! Didn't expect such big variations in 144Hz and below. Seems like sample size matters and clearly gives an overall better view on these results. Thank you for your hard work! I am already on a 165Hz GSync Monitor, maybe switching to 240+ in the near future for FPS games.
I've suscribed right now to your channel because you've this particular, clever pedagogical method of going to the point. Thanks a lot for the quality and intellectual honesty of your presentation. You make us earn one thousand of hours of checking some unintelligible others pretending experts. Best regards with all the wishes for this year.
Thank you very much for the visual image of what is happening when FPS exceeds Hz. I was so confused as to how this actually went down so I just could not understand it. The visual was extremely helpful.
Man, i am just in love with ur channel! Pls keep up the great work!
What A BRILLIANT and nicely presented technical video! I wish there were more such clear graphics related technical content on TH-cam! Thanks Chris!
Really great job!
This explanation is so much easier to understand, thus makes this video even better!
Thanks for the video!
So basically, refresh rate does more to reduce variance, while FPS does more to reduce the average values, but both things have both effects, just balanced differently.
Thank you for this video. Been looking for a video like this for a while but it is finally here.
The dedication you put is creating this videos is incredible! The cherry on top would be testing Avg. delay and STD with G-Sync Enabled!
This is the best video out there explaining this topic. Great job
fantastic work. thank you for explaining how these two are tied together.
thanks, that is a huge work.
but i suppose the fact that more hz and more frames(above hz) is better for input lag is pretty obvious to most of your followers.
can you please test more complicated and interesting scenarios as well?
for example: 60 fps with 144, 240, 360 hz, cpu bound with different sync and antilag options capped vs uncapped, some tweaks like hpet, timer resolution, dram overclocking (both clock and timings), etc
UPD: oh i see, there's much more detailed comment with the same idea. guess i should leave this for TH-cam algorithms anyway
Thank you for all the content and testing !
Can you please do a video of "fast-sync"? According to nvidia at least that should allow to run games at unlocked framerate, while not getting tearing as they use extra buffer where rendered screens go and then front buffer which updates at the rate your monitor does (basically unused mid-buffers are just thrown out). Or at least that's my interpretation of what i've read. Would be really interesting to see proper tests on this feature, as it might help people with lower refresh rate monitors to both run games faster and avoid tearing.
Great video mate!
I would love to see some videos on popular "in progress" games like Escape from Tarkov. Just to get an idea about what optimization really means for games. I always feel like my input lag is so high in that game. Would also love to see how stable the netcode and fps is from a more official source like yourself.
How about playing at 60fos on a 120Hz display?
My gaming laptop have a 120Hz display but it has Thermal problems to I have to cap my frames at 60
That scatter plot in the end comparing 60hz, 144hz, 240hz, and 360hz looks like a visual indicator of how it feels to game on those refresh rates. How the jump from 60 to 144hz is the biggest, 144hz to 240hz is significantly smaller and 240hz to 360hz has a very minor improvement.
This is a good visual presentation and explanation for people who are not aware. Another video that confirms what I wrote in my blog 3+ years ago and another good, quality one to add there. Keep up the good work, Criss. You are the only youtuber that make this kind of content and explain it very well. I wish I had the Nvidia tools too.
It would be nice to see a comparison b\w CPU and GPU bound with relex On. Im interested to see is there a slight delay with Reflex. Maybe the next video? You, already have the CPU bound results from this video.
Btw competitive FPS players play on lower resolutuion so they are cpu bound and they dont need Nv Reflex but fast CPU
Great work as allways, keep it up :)
Thank you for your super good videos and content I love watching your videos! And I hope you sleep feel good with this channel and aren't stressed out or something! You have great fans and followers 🙈☺️
Настолько понятно, наглядно и просто ещё никто не объяснял. Спасибо автору за проделанную работу.
Insane video. Very nice!
Amazing video as always!
This makes so much sense now. I have 144hz monitor and I always cap my FPS at 144, even though I can run VALORANT at 250fps, leaving it uncapped while I have 144hz makes the game feels so much worse, stuttery and not clean or smooth at all, even with high frames. I guess next step for me is getting 240hz monitor.
Also off-topic question.
Should I use nvidia reflex if I cap my FPS at 144 and my GPU usage isn't even 40%. I watched your video about it, but could never figure out whats best in my case. Cause I was using nvidia reflex but it felt like it had no impact, or it felt worse at times.
Uncapped framerate will always be unstable as the CPU and GPU are trying its absolute hardest to pump out as many frames as possible. Capping it at a framerate your system can always handle even in worst case scenario should be your framerate so if you think 250 is consistent for you and it barely drops from that value then go for it, obviously you'll have the screen tearing side effect which I wouldn't be too worried about, especially in VALORANT where reaction time is important. As for the Reflex, the feature doesn't usually work when the process is CPU bound, there was some setting you had to turn on which I don't remember (sorry) to make it more GPU bound. You could also just increase the game's video settings to make it more GPU bound
When I had 144hz, I used to cap my framerate in ow to 200 and disable vsync. There is a slight tearing, but not that bad. Now Im using 240hz, and I cap my in game fps to 236 and enable freesync, its really good, smooth, no tearing and extreamly responsive. I always cap my fps so I can have a more consistent latency.
@@qiyuxuan9437 your doing it right
Reflex should only be used if ~99% of your GPU is being used. It can actually add input delay if you use it while not GPU bound. In general, if you cap your framerate to a number that your system can sustain, you should leave Reflex off.
Use however you feel best I guess, but if we go by the chart, then setting your fps higher than refresh rate means better consistency (but more tearing).
What about lower fps on a high refresh monitor? Such as 144fps on a 240hz monitor vs 144fps on a 144hz monitor.
This is...EXTREMELY useful. Exactly what I was looking for for deciding whether limiting fps to the refresh rate is worth it the sacrifice; i.e., decreasing system latency (by decreasing fps) to stop tearing. Capping fps IS worth it.
if your gpu is not used at full load then dont cap fps
@@rashidkhan-xm4rl No. Always cap FPS. Shooter games have another problem: different mouse sensitivity at different FPS and also the higher the FPS, the more unstable it is.
It's very easy to test: either get an Arduino or any other board with mouse capability, or get a program that can emulate mouse inputs. Lock your game at 30 FPS. Then make it move exactly 360° at 30 FPS(I got arduino and did something like 3627 mouse moves by 11 points in the x axis to rotate my char by 360° in CSGO at 30 FPS). Now capture FPS at 300 FPS and make the program do exactly the same thing: send the exact same number mouse moves at the exact same distance. What you will observe is that while at 30 FPS your character moved exactly at 360°, at 300 FPS, your character moved either a little more or a little less than 360°. Conclusion we draw from this behavior in tests: FPS affects mouse sensitivity.
Now, unlock your FPS and let it be as big as possible. Let the program move your mouse again. You'll get *different* mouse travel distance each time you do it, because of the inconsistency of FPS.
Conclusion: lock your FPS to get stable sensitivity in shooter games, so you get better aim.
This behavior is repeatable in any video games that I tried: CS 1.6, Source games(Half life, CSGO, whatever), UE4 games(Valorant, Squad, whatever).
Why? Well, it's because of the calculations rates and how floating point calculations are done. You get a different result if you move your in-game view 10 times by 1 mouse move point multiplied by your sensitivity, compared to if you move your in-game view 1 time by 10 mouse move points multiplied by your sensitivity, because of how computers calculate and store floating point numbers. And when you do it 3000 times over, the errors accumulate. And so when you're doing it with unstable frequency, you're getting very unstable results. Once you lock FPS, sensitivity becomes more stable but still has errors because of how calculations are done.
The proper solution to this problem is to always move the in-game view X number of times by 1 mouse move dot. So, let's say you locked FPS at 60 and you moved your mouse by 10 dots at the period of time. Right now CSGO will move your view only once at the next frame by 10 dots. Instead, it's better to move the in-game view 10 times by 1 dot. That way the sensitivity will become the same at any FPS: 30, 60, 300, unlocked unstable FPS, whatever.
So which is better? Pump as much FPS as possible for better performance at the cost of tearing or cap FPS to refresh rate to remove tearing but at the cost of performance? And what about whether to use vsync/gsync?
One important takeaway from this is the diminishing returns in gains. It seems the jump in improvement from 240 to 360 is negligible compared to 144 to 240.
Tearing over 200fps ain’t even an issue, vsync for no tearing is good for 144fps and below where it’s noticeable otherwise push that mf hardware for the newest frame asap aka uncapped
@@puffyips Okay, thanks!
@@Bl00dMaliceif tearing is noticeable to you then ofc you can always enable vsync +gysnc/freesync and cap fps 1-3 under max refresh
@@puffyips Do either or both of those cause input lag or harm FPS performance while enabled?
@@Bl00dMalicefree sync/ gysync I believe adds barely even a mili second extra and paired with vsync on the driver level (as long fps doesn’t go over max refresh) it’s negligible to say for the benefit of no tearing. uncapped with vsync off has been noticeable better input lag to me with no visible tearing atleast in apex coupled with anti lag/ reflex on per your gpu
Interesting and informative. Not sure I'd be willing to sell my liver to get a 360Hz monitor, lol!
144hz seems more than good enough for a casual gamer/ user
Great Video Mr. Chris.
I do have a question for you. I have seen alot of videos of increasing gaming performence by modifying windows 10, unparking cores, deleting windows apps and so on. A few ytubers like Fr33thy, lock outz have alot of these vids up. Was wondering if you have done this to your systems or even tried these modifcations at all, do they even improve your gaming performence?
Again great video man.
Gotta admit, I'm still confused about what the best combo is for the best input lag, frame pacing, and odd refresh rates.
I have a 180Hz capable monitor for instance, and while I know enabling v-sync kills latency, most games don't have a fps limiter with 1Hz increments, so you can't just go down to 176Hz or even below just standard presets. Still not sure why you need to go 4Hz below your maximum refresh even with v-sync off either, or if in-game caps will fix frame pacing issues. And it gets even more confusing when you see tests that show uncapped framerates above refresh rates making latency worse, or putting an fps cap on games without internal limiters through something like Rtss can be worse than just turning v-sync on. Or hell, blur busters says to leave v-sync on with G Sync, and that rtss adds no lag, and other tests say otherwise. I swear, this stuff all needs to be standardized for consistency already.
I completely agree. I feel the same way. I’ve spent more time tweaking my settings on games than actually playing. I just wish there was a golden standard to go by.
I also wish there was more info/testing with Reflex and Reflex + Boost.
Can't you use the fps limiter in Nvidia driver settings? Is it trash?
@@Z3t487 It is trash, lol. At least as far as my experience with it goes. In-game limiters are usually your best bet.
@@J_Karnage I don't know why i wrote that, i meant to say: create a custom resolution in the Nvdia Control Panel and use custom 176 Hz (can use whatever value you want from 1-180 Hz). Can't create refresh rate that already exists in the preset. Values above 180 Hz could damage the monitor but probably 181 Hz is safe to test ( i don't recommend to do it tho). About the fps limiter, the game must have a cfg to edit. Example: in Wolfenstein: Enemy Territory you open the console and type /com_maxfps 125 for example or in the configuration (cfg) file.
Hi! I would just like to appreciate how your graphs are so professionally made and easily understandable too! I do would like to know which tools aside from NVidia's LDAT did you use to help you with the data visualizations?
Thanks!
Hey Chris, great work!
You tested framerates over the refresh rate here, but another great comparison would be the difference from refresh rate alone - e.g. 60fps at 60hz, 144hz, 240hz, 360hz.
With that data we would be able to isolate the impact of refresh rate, the impact of frame rate and the impact of both together.
+1000
You shouldn't lock a game to 60hz with monitor at 144, the stutter alone will make you play worse.
If talking about multiples, like 60 fps on 60, 120, 180, 240 etc, there shouldn't be any difference in input lag, the display will simply repeat the same frame 1 2 or 3 more times, at least in the best case scenario when using some kind of sync. Running completely unsynced would make a higher refresh rate one better though, and cause tearing on some of the frames and not on others (since the frame changes less often than the display).
Anyway, running locked 60 fps on 240hz monitor is a bad idea and you should invest your money elsewhere.
@@rtyzxc you hear about Prerendered frames? if you have 60fps the frames are 1/60 = 16.66 ms "longer" and if you multiply that for the number of prerendered frames you have the total inputlag, 60hz with 300 fps have much less input lag than 60hz with 60fps though vsync on(only fast sync allows framerates higher of refresh rate of the monitor). PD:SORRY FOR MY ENGLISH.
Dont forget fast vsync is triple buffered and this introduces some input lag too.
That's is why if you lock your framerate below your refreshrate you have better inputlag because prerendered frames = 0
@@rtyzxc "You shouldn't lock a game to 60hz with monitor at 144, the stutter alone will make you play worse."
Zero refresh aliasing with adaptive sync any frames at any rate slower than the maximum of the monitor.
"there shouldn't be any difference in input lag"
There's a very sharp improvement in input lag due to the faster scanout (the time between the frame reaching the front buffer and being fully refreshed onto the monitor is much lower) but it would be great to quanify this (standard deviations etc) with the new tools that allow for sampling hundreds of times.
"Anyway, running locked 60 fps on 240hz monitor is a bad idea"
60fps was just an example rate - one that every monitor can do, therefore it's good for isolating the impact that refresh rate has on latency across the entire range of monitors from 60hz to 360hz.
Locking framerate below the maximum refresh rate of the monitor - sometimes quite far below - is actually a very legitimate thing to do on an adaptive sync monitor, depending on the game. Other games cannot achieve higher framerates due to CPU/RAM limits, but still experience a reduction in input lag and an improved response consistency when upgrading to a higher refresh rate monitor.
@@AerynGaming I thought you were talking about running fixed refresh rate above 60, now you are talking about adaptive sync, which will of course run the game perfectly at 60 fps, 60hz.
the data at 2:05 follows a periodic, upward sloping latency pattern. Does anyone else notice this? Three distinct groups of data clusters are visible. Weird, ingame rendering oscillations?
Can you do a video on the effects of Nvidia Reflex (on vs off) on input lag / system latency on high tickrate games (CS, Overwatch, Valorant) vs. low tickrate games (Destiny 2)?
What happens when i have a 240hz Monitor but i have lower fps? I feel like having a huge disadvantage compared to 144hz with 144fps.
what if i have 200 fps on 240 hz? Is it Bad or better then 240fps on 144hz?
Great video again!
I'm new to input delay analysis. Can someone clarify what is actually happening? For example, I'm confused as to whether input in online, player to player games is being relayed to the server, or other player, before it is displayed. Does the visual even matter or is that somehow synonymous with the baseline P2P transmission starting point? I'm familiar with net received delay across web. There are so many issues. In COD a simple "hit scan" firing point doesn't work with all these projectiles. What happens with aim assist as an enemy's frames slow down, motion is interpolated, and impact position is not where it would normally be due to delay? Add actual input delay with now muddied aim assist... I once saw a Valve dev article talking about interpolation making extremes of a bouncing ball impossible to hit despite visual presence. This is similar to the dreaded snaking.
Me too, I'm curious, is 'input lag' just 'visual lag' or does the computer not process any user input during the frame gap? But I kinda think this will be based on each individual game
@@iXpertMan I originally thought it was actual device lag, in the way that pro COD events were using Cronos max to lower that side, but then I saw the delay being measured on same side displays, with their own turnaround time, AND heard that some games update local state only after server side confirm which just sounds insane.
I believe he’s measuring delay between input from mouse or keyboard + system latency + display latency. So the numbers shown is only client side from mouse click to action on display. Nothing to do with server latency or lag compensation.
I'n most games your client will show appropiate responses aswell as actions on its own before it verifies with the server, thats why you can see yourself shooting at enemys in games but the shot not connecting when you are lagging.
Great work!!!
What about the inverse? Would love to see a game running 144 fps at 144, 240, and 360hz
Exactly!
@@IslamGhonaym he showed it dumb
Did you watch the video?
You guys dont understand what OP meant. He would like to see what happens if you achieve a x FPS < x Hz scenario
@@pandasdreamlygon actually he didn't if you look. I would like to know also what would change if run a game at 144fps on a 240 hz monitor vs a 144hz monitor.
what if you have higher refresh rate and lower fps? missed opportunity in video ;v
danke. das ist das video auf das ich gewartet hab. ob fps > hz einen Unterschied macht.
und danke für das Datum links oben in den Diapositiven.
Your graphs are looking nice and crispy man. And were getting GamersNexus levels data density. Nice work.
Great video, thank you for your efforts! If I have a 144 Hz monitor, should I enable my games to go past 144 FPS to get it smoother, or limit it at 144 FPS to prevent screen tearing?
What happens when you get 144fps on a 240hz monitor?
Thanks for your videos and effort , greatly appreciated . But i'm completely lost between the different advice you give (i'm sure it's my fault) , i mainly play valorant and i have problems with packet loss . I've tested the method of capping the fps , and it seems to work fine for that particular problem , but now i don't know if i have to cap it to the max fps my pc can sustain , or the 4 fps below the refresh rate of the monitor (144 hz in my case ) . I have an 8700k with a geforce 1070 , i can spikes up to 330 fps and dips to 170fps sometimes . So i don't know if i should find another solution to the packet loss , or just cap my fps , and if so to what number
So if my pc cant handle consistent 240 fps but i have a 240 hz monitor ,would it be better for me to cap my fps to 240 and have it drop a lot or do i cap it at 144 fps so there is more delay but its more consistent?
Huge video! Thanks for uploading
What about 144 FPS in 240Hz monitors?
I would like to add that in games like BF4 not sure about BFV running a multiplier of the server tick rate is also helpful. one more mention is run 3 FPS lower than your G-sync range for better frame times.
What about a reverse situation where you have a 240hz monitor but only get 144-180 FPS in a game. Why does it feel less smooth than lowering that monitors refresh rate to 144hz?
can test a old CRT with GTX 750 TI vs 240hz with 2080 ti abount latency? ;D the conversors to vga do a big imput lag =//
insane WORK!
This is just what I am looking for and very helpful, thanks.
Now I wonder what FreeSync/G-sync do to the latency. I believe it should reduce the longest delay by the inverse of the refresh-rate(e.g. the latency should be 10.3 ~ 16.3 ms instead of 10.3 ~ 23.2 ms at 144Hz/144FPS with FreeSync on).
_" believe it should reduce the longest delay by the inverse of the refresh-rate(e.g. the latency should be 10.3 ~ 16.3 ms instead of 10.3 ~ 23.2 ms at 144Hz/144FPS with FreeSync on)."_
No. Normally, the monitor just shows the most available frame, and the minimum and maximum delays are at their absolute minimums. There is no syncs to do to decrease them.
VSync delays rendering frames until the new monitor refresh is ready to start.
GSync delays monitor's refresh until the new rendering frame is ready.
GSync can't reduce neither longest, nor shortest delay, because of the way it works: it actually delays the refresh of the monitor until the new frame is ready.
So considering you'll have to lock your FPS at 236 when you play at 240 monitor in order for it to work, instead of playing at 300 FPS or 700 FPS, or whatever, you will have bigger input delay with GSync compared to if you played without GSync with FPS locked at a higher number.
What GSync does is that it fixes screen tearing and makes input delay more stable in the same parts of the monitor: bottom of the monitor always has bigger input delay than the top of the monitor.
And it's not worth it, it's just another marketing trick to make you pay $700 for a monitor instead of $300, and still get worse results for what they advertise it for: gaming.
I'll take a $300 240 Hz monitor without the GSync and play at 300 FPS instead of a $700 240 GSync monitor to play at 236 FPS any day of the week. And have smaller input lag all over the place on the $300 monitor. And save $400 while I'm at it, too.
I'm playing 300 FPS on my 260 Hz monitor and I never see tearing. Never. And I looked hard. It's there, but it's so small at 300 FPS and so inconspicuous, that all these syncs are absolute nonsense as a fix for it.
What about the latency of 60fps to 144hz? Great video by the way. Keep it up.
What you need to understand is 60fps at 144hz will no longer feed your monitor old frames. because your refresh rate is faster than your gpu output so it will show you the most reason frame.
I've known this for a while being in the competitive gaming community but there's a lot of people that think it's pointless to have more fps past the refresh rate. Your videos are informative as always
So you think that gaining 10ms (avg human reaction is 300ms) is a big advantage? or maybe just marketing? ))
i have spent all week trying to fix my Stutters thinking something was just wrong with my windows its must be the QPC timer or something turning off background services updating rolling back updates new drivers installing a old version of windows none of it worked why was my 9900k and 1080 running worse then a 2600k and 1650s turns out 1080s just run like dogshit worse then any of the Turing cards if you don't cap there fps
i watched this video turned off NULL turned on Gsync and then ran fps_max 235 in csgo and the Stutters are gone
I hope we get to see some of these graphs for other games/engines too
At 8:04 we can see two kind of separated sample groups at ~12 and ~16 ms -- with the 12-ms-samples from sample ~30 to ~130 and the 16-ms-samples especially at the end ... is this an artifact of some kind? (blue dots für 240 Hz and 240 FPS)
The other samples seem to be in normal distributions, this is kind of bimodal.
One thing to note is that the worst case input lag at 60hz decreased by 1/60th of a second when going from 60fps to 360fps, a result that checks out on an intuitive level.
what abt 60fps but with 144hz monitor?
i wonder what would the latency would be like
I wonder what 60 fps will give you on a 144 or 240 hz displays. Should bring lower lag.
Would you like to check the impact of SMT or Hyperthreading on Input Lag? Could be Interesting. In general Bios Settings :) Of course with ingame limited frames. The lower input lag by the higher frame amount with smt we can just calculate and see if it's worth the input lag it gives :)
Doesn't capping a frame rate also reduce system latency? If your system can output 160 fps, but you cap at 144, pegging your frame rate to that cap, the decreased pipeline also gets drawn frames to the monitor faster?
Great Question: Why don't monitors scan from the middle outwards?
Excellent video
Could you show some tests with G-sync on? In theory it should be more stable but would be cool to see it on graph.
nice can't wait for gsync / gsync compatible retest with games like valorant and csgo
Chris @Battle(non)Sense, what is the effect of having a refresh rate greater than the FPS? Would that improve the system latency in some way?
If you have a refresh rate higher than your fps, it will not improve system latency, it will increase it actually. You want a fps as close as possible to your monitor's refresh rate to reduce frame timing to the nearest lowest ms. Low fps and high refresh results in higher frame time which means it lags. For example. I have a 144 hz monitor and at 144 fps my frame time ms is 7.2. at 80 fps, I get 16-20 ms for my frame timing which is lag that is detrimental to competitive pvp fps games.
The problem I have is that my mouse and keyboard strokes don't always register and my mouse feels slightly 'floaty'. It's been like this for some years now I've been unable to find a solution and I'm running an RX6900XT and get around 200-300 FPS.
What about games that are capped at 60FPS? How different is 60FPS@60Hz vs 60FPS@360Hz.
Maybe a dumb question, but I felt like it was the only use case that wasn't covered.
won't do nothing, it can't put out more frames or a part of it so you'll always get the same latency assuming that your fps are constant.
Your videos are a god sent, do you have any plans to test 2.4Ghz Wireless keyboards/Bluetooth Keyboards or wireless keyboards in general?
This is the most accurate video i have seen in the world, this proves you don't need 120 hz monitor to become deadly. 120fps graphics card with 60hz is good enough for casual BR games.
omg your work is outstanding
great video, thanks for sharing
Hopefully someone can explain this to me. When using MSI AB with RTSS, when I display frame time, my system was getting terrible frame time spikes 60-100ms. Mainly play Fortnite.
By capping my fps using RTSS (144hz monitor) capped to 140, this greatly reduced the spikes but they still occasionally occur.
Can someone explain what causes frame time spikes and how to reduce or eliminate other than frame cap? When a spike occurs, the game freezes aka seems to hang. 10850k with rtx4090.
this is what we've been waiting for! comment to help the channel 😉
I would have loved to see the results of running a game at 60fps on a 120hz or 144hz monitor
What you need to understand is 60fps at 144hz will no longer feed your monitor old frames. because your refresh rate is faster than your gpu output so it will show you the most reason frame.
@@acid1787 oh I understand that, I want to see the effects of latency, that's all.
@@madmax2069 the latency will be lower since you aren't being fed older frames. This is only applicable if you're running the game on exclusive full-screen mode with no sync option. Otherwise you are just running your game on windows triple buffer vsync which result in choppiness.
Why games can't update input in a different thread so that it does not get delayed because of vertical sync?
Why does my pc decide to run sloppy the next day no matter what I do timer resolution anything anything my pc just plays like console with high fps ? I can’t seem find an answer..I always reset it and download all drivers and everything I know and runs flawlessly but the next day or two I turn it back on it just doesn’t feel right and I have to just wait till it decides to.. same thing would happen on ps4 and Xbox ..and what pisses me off more is that even real world shit like my truck it runs great some days and the next it just feels bloated
Where do you think the sweet spot is in the refresh rate if you only have a 60hz monitor where you don't get too much tearing but get to enjoy the lower input latency?
Your 240hz 240fps data is very interesting. You can clearly see a segregation into 2 bands of response times with very minimal in betweens.
May have something to do with the poll/scan rate of your hardware, monitor and game causing the visible banding of the results.
Please make a video where you can see the input lag of Battlefield 2042. I still think that it is horribly optimized and I notice input lag that is not normal in an FPS. Thanks for the work and congratulations for the channel. 😘😘😘🤩
Could you please do net code analysis on FIFA on consoles?