Huh! I initially thought this was an issue with a missing deltaTime calculation somewhere that would be easy to fix, but judging by the results of this testing plateauing at 120 FPS, it actually seems like it might be something a bit more complicated. I hope it gets ironed out in the near future, whatever it is; Wolverine being suddenly unable to do his job at certain FPS thresholds is a pretty raw deal.
What else could it be ? Some server side problems ? They may have a manual cap at 120 fps as it seems that 120 is the intended framerate for how theyve designed damage and abilities
@@cctz_1 Given that I don't have the source code and I'm not overly well-versed in UE5, I can't say anything for certain. This is likely an issue that would need to be investigated in a bit more depth. That being said, I can think of a few potential culprits: - Depending on how they're spawning hitboxes/projectiles and whether or not they're tied to the animations somehow, the firing rate bugs might be a quirk of how animations are rendered and looped at lower FPS. This would explain why the issue specifically affects fully-automatic weapons with distinct animations for rounds firing (Punisher, Star-Lord, Wolverine). Tying a game-affecting action like that to the animation seems like a terrible idea to me, but I'm also not on a professional game development team with strict deadlines to clear. - There could be a netcode issue where FPS affects how certain packets are sent and received by the client, with hitboxes and tracers being spawned at the wrong rate or movement being applied at the wrong speed due to game speed discrepancies. This strikes me as highly improbable given that servers are usually supposed to do most of the heavy lifting when it comes to things like player positions and damage and a client's FPS shouldn't affect this, but netcode is a complicated subject and I have no idea what this game handles on the frontend or backend, or how. - We also can't entirely rule out some deltaTime flarkery, though it seems less like "they forgot deltaTime" and more "the calculation involving deltaTime is acting weird." Maybe they accidentally put in deltaTime compensation twice, or they're using some under-the-hood calculation that does weird things at low framerates. Again, can't know for sure without having the equations in front of me.
@@ThunderMan2212 I see. Now that Ive seen some other examples of it I think its indeed a netcode problem. Some ppl are claiming it looks like the prediction system is getting confused by lower fps players and compensating in a way that causes this bug. But theres no real way for us to know rn
He doesn't even play on 30fps my guy, also the guy in the video didn't notice but he ended the animation from the grab sooner on the 120fps that's why wolverine was hitting more. I just tested in the training range now, it's pretty much the same.
Also, they're really underselling the difference in kill time with stuff like Starlord. Those are the times where you just got away with a sliver of health, but got healed or killed them at the last second.. but instead, you're gunna die because of that slight change of 'milliseconds that dont matter'.
I don't blame him but this vid coulda been better researched but i don't expect him to know every intricate timing of each character either i would take this vid with a grain of salt
@gawrgura3767 i'm on Pc and this game takes a lot of resources. I play overwatch2 on ultra setting and run 200 FPS smoothly.. This game i have to lower all the settings just to keep 80 FPS. Consoles are built with the same components making it a network issue most times. PC have many variable parts so trouble shooting is much more difficult
Since the genre is competetive multiplayer, I can use fighting games as example. And as mentioned in 6:51 , there's a whole second of increased time for the low fps player. In ANY fighting game, adding around 62-66 frames _(1.1sec)_ to an attack for one player only and calling it unnoticeable, is clearly a misinformed take on how important every second is in these competetive environments. In short, that 1 second extra to kill, he's getting a triple Loki healing shot, good job you missed the kill because of your shitty CPU and nothing else.
@@MarvelRivalsMerchant 0.1 seconds isnt a lot but it adds up, if theyre getting healed that could go from 0.1 to 0.3, and over the course of an entire match the amount of damage you do would overall be lower so you might end up with like 5% less ult charge or something.
@@BagelBoi4000at 0.1 there are more things that would screw you over than that. In the grand scheme of things if you are good enough it doesn't matter. The horizontal movement and wolverine DPS however from 30-120 are absolutely terrible.
Exactly what I was thinking. Have a conversation with someone who's trained to hit trigger switches on single frames to link moves or cancel animations and tell them its insignificant.
One second is much more important in fighting games than in team games. Fighting games are so much more micro focused. It’s a bigg-ish deal but in a game with healing and damage from external sources there’s already “randomness” for when a combo will truly kill You can see this when watching vods of fighting games vs marvel rivals, going back 10 seconds in a fighting game feels like forever and 10 seconds of a team fight feels minuscule
Edit: I created a video showing in the game engine how and why the bug exists for those who are curious. Game dev here chiming in to answer why it does this with unreal engine. The reason is because the game latency is taking longer to acknowledge and respond to your player input. Due to having a high latency for game client functionality (not to be confused with ping latency), and low frame rate which creates the high latency, things like holding a button down to initiate individual input requests to fire off the output such as wolverine's slash attack is bottlenecked and has a delay. This is actually common with hold X key or button to perform Y action. When using doctor strange's flight ability, the game is initiating similar behavior by scaling the character position to the parameters set in the game code. However, his distance is impacted because of the "real time" script lifetime isn't matching up with the client. In other words, the client is saying it's done because the time has passed but your game response is essentially saying no. It may sound a little confusing but it's important to remember that the script will execute and the time will finish before the game can truly finish the executed operation to the pawn controller. I can upload a video demonstrating this issue in a fresh UE5 project as an example.
another "tip" I haven't really seen mentioned is making sure vanguard is turned off if you have league or valorant installed on your pc. Vanguard can absolutely nuke the performance of your game
just so you guys know, these issues were solved by some kid on reddit and he posted what code to change in one of the config files so animations arent so resource demanding. I can confirm it fixed all these issues but the game developers decided to block access to config files and threatening bans if people mess with it. meaning they know how to fix it atleast but the fact that they havnt probably just means they gonna fix all the other issues like dr strange portal causing frame spikes all in one patch. but yah the main issue is how resourse demanding all these animations are and it causes low end pcs to miss hitboxes or suddenly lag effecting animations. sometimes wandas "shift" gets cancled cause the game thinks you are double pressing even when you are not. so many bugs
@@laWiskaPY the best way to solve this is to force ranked performance to an agreed achieble standard across all platforms but obviously people who spend thousands on equipment would complain and since games cater to them, low end equipment will never work on ranked play. and with games getting more intense faster then affordable hardware can keep up it might be the end for ranked play in most games.
that kinda sucks because i main strange and the flight boost with higher fps can actually be a gamechanger because he can use his ult in midair and cancel it to catch ppl off guard
The game feels more smoother. But you should also keep in mind that in fights high FPS drops can lead to higher gpu latency. For instance if you have 240fps and then it drops to 180fps fast, then you can feel it ingame. So I recommend you to cap it to a "good average" FPS for lower gpu latency and more smoothness.
Yeah, but the game has some extremely demanding rendering tech available, it's an impressive looking game. Style doesn't mean that much, I've seen game sequels that require 2-4x the resources, such as with series that span multiple console generations. BUT, at least on PC you can nearly double your framerate on low settings. Using state-of-the-art Lumen GI and DLSS on Ultra settings, even my RTX 3090 barely maintains 100fps, but I can get 240 on lower settings
@@Yuusfw Depends on settings because 100 fps can be difficult on RTX 3090 if maxed out on Ultra, but 240fps is easy on lower graphics. Moral of the story is people should tweak their settings because high fps is important
@@EhurtAfy Its a competitive game. Performance is better than Image quality. People are making it demanding because they want to. No one is forcing them to play full resolution everything ultra. Even streamers with good PC reduce the image quality when it comes to competitive games bc its more fluid
I believe this is an animation issue, idk what the distance issue is but the TTK issue is related to how fast the animations play out, star lord can reload in his ult faster when the frames are higher, wolverine can swing more often when the frames are higher. Again the distance thing is crazy but the main issue seems to be the animations and how they were implemented and are involved in the game process thats why its caps at a certain FPS because you cant gain anything more after the cap
I get between 12-40fps depending on the map and the position of Jupiter. But I'm laughing at this because the difference is so negligible compared to Helldivers 2. In that game my FPS is horrid and your movement speed=FPS meaning I move so slow it hurts. Everyone can outrun me so I take the jetpack and medic suit with 6 stims so I can (try to) keep up. New PC coming this Friday, can't wait.
3:50 arent some of these things just discrepancies due to low fps recordings? if you were playing at 1fps, those kill times would have been 2 seconds. not because you kill slower in game, but because on the first frame that takes a second the enemy is alive, and on the other frame after another second they are dead, marking the measurement at 2 seconds. on 30 fps this discrepancy is still noticeable, but on 120+ fps the extra frames add very little difference to the recording, capping out at whatever framerate you record on (120 fps in this case?), meaning that the kill time stays consistent doesnt necessarily have to be true due to the doctor strange flight ability being clearly affected by fps, but if you measured the time to kill by looking back at your recording, then it definitely is the reason (and not the game itself)
By the way, there’s a display setting that shows your current fps in-game. Please use that I’m conjunction with these types of videos in the future. Makes it easier to keep track of things.
I can explain it from a developer perspective: -The most basic stuff is this, some command get calculated client-side and some performed server side, there will never be a multiplayer game that relies everything on client or rely everything on server side, the closest game that relies mostly on server side is Street Fighter V, and people hate that game. -Looks like from the data, the game treat Button Hold as constant command, meaning if you hold left button, the game will keep sending Left input to the server -Here's the kicker, when your client (your PC) didn't send any command, the game treat it as "Input neutral" even if you're holding the key -The number of input you can send is based on your FPS Basically Someone with 120 FPS can send "Left" input 120 times Someone with 30 FPS can only send "Left" input 30 times, the rest of the time, you will be input "neutral" this is why Strange's can fly further left if you're holding Left on a 120 FPS machine Any character with "reload" animation (This includes Wolverine that can stop attacking after 4 hits) also get affected by this, since 30 FPS guy can only send 30 commands (hypothetical) every second, other than that, the game treat them as not pressing anything, this caused the guy with 120 FPS can start the "attack again" command earlier than the 30 FPS guy. At the end of the day, this will show on 2 extreme side (30 FPS guy and 120 FPS guy) to show "the worst thing that can happen in that situation" Most game generally treat this as "whatever", similar to how people will accept that a 30ms guy will have better play experience than people that play with 120ms.
I notice that Laging players also suffer from damage output and received problems… which could potentially stack with the FPS problem. Eg- you either get free kills on the lagging player as they are stuck in place in you pov, however the dmg needed seems to be a lot more than normal Or You shooting the lagging player to then suddenly be chunked for 250hp or more depending on who and the situation. Point is the servers suck which we know they do…. and or it’s possibly an exploitable problem that people could force.
This is such an amazing video, the amount of effort put behind the research, the details on the relevant cases, and the tips at the end are all so incredibly helpful
@@MyKlaudijus i get 144 fps but at a lower resolution game still looks great and runs well loading times are longer and a few shader issues sometimes to load but when all in 120 to 144 fps constant
yo btw for star lord and punisher its due to the lower frame rates + high rates of fire, on 30 fps you might fire a bullet on the 4th frame and on 120fps you might fire a bullet on the 15th.
It's a feature of Unreal Engine 5. Other games made with the engine also have a higher movement speed & DPS at 120 fps. It's not a massively higher rate in Marvel Rivals but in games where they put hundreds of millions of health on some boss enemy and players do millions of DPS, this small difference becomes much more significant as it acts as a multiplier.
Seems like it has something to do with the difference between animations being tied to framerate, while also being tied to damage output (like bullets firing or claws swiping) yet damage HITS are calculated server-side so if the animations slow down, the attacks are sent out slower, and then the bullets hitting the enemy will be registered further apart (or swipes in wolverine's case), at least that's what I'm seeing.
*For the last part, additional feature to increase FPS:* You can decrease *screen resolution* (while keeping it Fullscreen) in accordance with your normal resolution. _For example:_ If your normal resolution is *1920x1080,* you can set up *1280x720* - it would make your game running as a quarter from your normal screen and increase FPS. Works good in many games, especially competitive ones.
Risk of Rain 2's DLC had this issue where the enemy spawn rate was directly tied to the amount of fps. And it could be a real issue for new heroes moving forward.
Old Quake Engine movement was tied to frames. In Quake 3 there was a sweet spot like 125 250 and 333 framerate. Also on quake 1 and 2 you could only do some jumps with higher and lower framecap.
Seeing this got me to think about Winter Soldier primary fire as well 🤔. It feels weirder to aim & shoot on lower fps to the point your weapon basically don't register hits on longer range when visually it looks like it should. I do realize his weapon is not a hitscan but a projectile based (strangle design decision) so placebo wise it makes me feel that character could be a victim of this game flaw as well. If you can, please confirm in your next testing that as well since i'd love to play bucky more if his shots feel consistent in crunch situations.
Note this kind of issue also exists in Mobile Suit Gundam: Battle Operation 2. The lower framerate, the slower the mobile suit moves in general. Attacking is somewhaylt noticeable, but effects combos players do.
the more concerning thing here, is that it shows there are NO serverside checks for if the player is moving right. that opens the door for all sorts of cheats
While I def want this addressed, I think people don't realize how common an issue this is in games. Plenty of games I've played recently and watched in depth testing/guides on have mentioned listed stats don't always line up with in game results.
I actually did notice a movement speed boost but it wasn’t because of framerate it was because of latency or lag. The symbol that shows up when theres a network issue. Its annoying
It's likely animation based events. If you are ticking animations at a lower rate you can get these kinds of low frame rate delayed responses because you are frame staggered from the game thread and the animation thread. If it were a lack of Delta time multiplication then the impact would continously grow beyond 120 fps. Then there is the network layer. Does Rivals use entirely client authoritative hits? Do the abilities calculate hits in the server?
For the most part, i feel like some of these are so minute especially when considering all the disadvantages you are already at playing below 120fps. Theae arent the games to take competitively seriously when your hardware cant keep up, unfortunately. The majority of opponents will have an inherent advantage over you that far exceeds a couple of miliseconds. The Dr Strange and Wolverine cases are more agregious though unfortunately
It seems like a deltaTime issue and the developers probably developed console first and forgot not all platforms have uniform fps across the matchmaking pool
It's probably the reload that gets slowed down then if ironmans uni beam and cloaks primary are unaffected. So your tests should be against a target with more health so you're more likely to need to reload. Assuming this is in a real game and you dont have the luxury of 100% accuracy this could be more significant (but still not massive) than your tests have proven Vs Luna Snow in training. It's a bug they absolutely must fix either way. Thanks for doing this experiment!
Quite strange that these issues exist when it runs on Unreal 5, yet the flagship Unreal 5 game Fortnite had already fixed this issue back when the game was on Unreal 4.
It looks like the entire game loop is being rendered at 120 frames per second, along with the video update. Technically speaking, if you were only rendering 30 frames per second, you would have a lot more time to be considering things like "does the player have the fire button down.. yes? shoot" It looks like these checks are being made at the same rate as the video updating, rather than at the speed of resource availability.. which means, as you turn up the frames per second, you are also turning up the frequency it checks to see "Hey, is the button down? Yes? send the message to the server that I've fired.."
Honestly this seems like an animation bug more than anything, the one frame of more damage with starlord is completely negligible and within tolerance as a frame is generated every 0.033s, and the damage with wolverine is likely the animation just finishing prematurely leading to massive differences with dps, but it's really just a very small desync with animation timing. I would also deem the distance issues with certain dashes to fall under the same umbrella of an animation or a Lerp finishing a frame early leading to slightly decreased distances of dashing. Overall, not really that gamebreaking, a couple of small issues but thats about it. If you only can get 30fps there are other limiting factors making you perform worse ingame.
Welcome to the world of numerics, where parabolic trajectories overshoot by a mile, velocity damping makes you go brrrrrr and snappy is actually smooth and jerky if it is frequent enough. I bet in case of Wolverine and Star Lord the animations have to wait until the end of the frame, but since the frame pacing is so slow (and the server repeats client's actions as is) it ends up adding to the total attack time. For Doctor Strange it looks like the effect of air friction changes massively. It may have been lazily calculated and forgotten, or some tiny mistake in per-update velocity loss coefficient, or just that the update pacing is so bad that the numeric "overshoot" effect occurs, but instead it slows down too much. And yes, all of those will plateau towards some update rate, as numerics work that way. Anyway, it can be fixed easily.
I wanted to say the same thing. In Destiny you do more damage at high fps but also takes more damage. Wonder if you take more damage in Marvel Rivals as well. But I don't think there are many people that play at 30fps lol
@@fransbrewis4776Yeah youre underestimating the average persons ability to deal with 30fps gameplay. I consistently have 30. Now id rather have higher, of course, but my pc is too shit for that.
I was watching a streamer yesterday a_seagull and he seems to have found that the games ping is also tied to fps. Tested this myself and it checks out even in the training range. It's not a cpu bind either since the scaling of ping to fps I nearly linear from 5 fps up to 300. This much fps causes this much latency. It makes no sense
Riot Vanguard Anti-Cheat can supposedly reduce FPS in your games, so it is possible that this Anti-Cheat does as well, but who knows. It definitely ain't helping FPS though.
It kind of makes sense because if you're rich you can afford faster monitor and if they keep you happy you will spend more money buying cosmetics. Dev: no no thats not it. The calculations we have to do to keep the character aligned throughout different frame per second is not easy. there always is some mathematical adjustments that needs to be made.
from programming perspective, frame render thread are in the line with main thread, the higher your frame rendered means main thread dose not need to wait more. that why its more responsive in almost all games. only few dev cares about this.
idk if anyone will see this but what he said about frame gen is right but, if you have issues like your game is freezing or fps is just tanking I really would recommend just trying it. I was really struggling with this games optimisations putting everything on low and it would still freeze but, turning Frame gen on pretty much fixed it for me. However if your hitting 60fps and below when playing you should not turn on Frame gen as you most likely will just get latency, you need at least 120+ for frame gen to properly work I feel.
Can you test it with Namor (orb ability) or Scarlet Witch (ult)?? They both have a similar lift/float animation to stranges fly. Ive felt Namor orb is so slow it just leaves me trapped....maybe its an fps thing
This is a big issue I think that due to lower frame rates it is sending information to the servers slower, so if u have higher fps then it would send data faster but again this would mean that every single characters time to kill would be different but its only on some characters. I hope this gets ironed out and fixed and I also know about some other bugs I have encountered on this game since I don't have a strong enough pc to run this game on higher fps.
This is actually a massive problem and something that should be addressed when it comes to official games. What should have killed someone and it didn't because the computer was average specs and not giving 120+ frames and that lowers damage. Ya this will make A LOT of problems for official stat tracking.
Playing as rocket ill paint someone with fire for a solid second before it start registering, especially at range, theyll be taking damage behind cover, vise versa
Don't worry guys I'm just testing something out for UA-cam I'll be out to help in a second 😂 Not gonna lie I laughed a little when you got in game and did it
this makes complete sense as to why i stopped playing the game, it was infuriating and nobody would believe me that it was happening, i had no idea what was causing it but now i know, cool.......I know im due for an upgrade but come on, not only is UE5 not ready for games yet (no game should release that cant hold a steady 144fps at 1440p on lowest settings, which is the AVERAGE across steam users) but FPS based mechanics existing in 2024? Come on, this game had so much potential.
About the wolverine combo part when you showed 120 fps you started attacking groot mid air during your feral dash animation and on the 30 fps you started attacks when you were towards finishing the dash. was the dash cancel mid air caused by the fps change? or did you do it like that without noticing?
They should have used fixed timestep to keep deltatimes consistent, but Unreal Engine doesn't offer one as it is not designed for deterministic games. That's the problem. The easiest fix is to just cap framerate to 60 on the engine side.
I'm fairly familiar with Unreal Engine 5 and there are a few different ways this could be happening depending on how they calculate the passage of time in the game. It also seems to be calculated differently for certain characters or abilities as well. I'm very curious as to what is actually happening behind the scenes in engine. I highly doubt we will ever get a peek at the code though...
With black panther, his dash, his most useful ability, gets a fair bit of extra reach when limiting your fps to 30. Uninstalled, they really need to fix these inconsistencies before it can be competitively viable.
It feels like iron man team up with hulk is affected by FPS after all. Had one in ranked as iron man myself... Well, I was noob but it really felt like he dealt damage faster
In case it wasnt obvious, Starlord's ult should basically be the only one that matters, strictly because it applies its damage across so many hits. That's a small sacrifice I'd be willing to make in order to have the game more optimized when its already pretty modestly optimized. Would be verry different if you were shaving off .1s TTK for every weapon but I think its just his ult. Also wondering if it impacts Iron Man's flight maneuverability since that is super reliant on acceleration.
Everytime I use my ultimate and I die before it fires, it just disappear and i lose my ultimate completely. All because of my fps drop on pc. It never happens on my ps5.
9:00 Frame gen needs to go off because it doesn't fucking work. Test it. Lock your FPS to 30, you know how choppy it looks. Now turn frame gen on, and it looks the same. Its not only input latency but also doesn't generate any frames at all. It is broken! Idk if DLSS frame gen works, but FSR for sure does not. If you have Rivatuner you can see your frame time going wild too when you turn it on. Which shouldn't be the case if it worked properly afaik. Also frame gen in general is fine if you have minimum 120 fps but a 240hz screen. The input latency isn't that noticeable. Even 60 base fps is fine if you really wanna smoothen things out.
My computer recently started having issues and im on 20 fps and im a pretty good play and i realize my oneshot combos aren't working because of the delay between where the enemies are on my screen from where i recall the shurikens
Hope they fix these minor issues. Im not autistic enough to think 0.01 second difference will do anything in my games but might screw over some matches for few players in the world. Id rather have it run on 60fps silky smooth than with 120 where it randomly drops between 80-120 depending on what kinda ults someone is throwin around.
Huh! I initially thought this was an issue with a missing deltaTime calculation somewhere that would be easy to fix, but judging by the results of this testing plateauing at 120 FPS, it actually seems like it might be something a bit more complicated. I hope it gets ironed out in the near future, whatever it is; Wolverine being suddenly unable to do his job at certain FPS thresholds is a pretty raw deal.
What else could it be ? Some server side problems ? They may have a manual cap at 120 fps as it seems that 120 is the intended framerate for how theyve designed damage and abilities
Maybe UE5 related
@@cctz_1 Given that I don't have the source code and I'm not overly well-versed in UE5, I can't say anything for certain. This is likely an issue that would need to be investigated in a bit more depth. That being said, I can think of a few potential culprits:
- Depending on how they're spawning hitboxes/projectiles and whether or not they're tied to the animations somehow, the firing rate bugs might be a quirk of how animations are rendered and looped at lower FPS. This would explain why the issue specifically affects fully-automatic weapons with distinct animations for rounds firing (Punisher, Star-Lord, Wolverine). Tying a game-affecting action like that to the animation seems like a terrible idea to me, but I'm also not on a professional game development team with strict deadlines to clear.
- There could be a netcode issue where FPS affects how certain packets are sent and received by the client, with hitboxes and tracers being spawned at the wrong rate or movement being applied at the wrong speed due to game speed discrepancies. This strikes me as highly improbable given that servers are usually supposed to do most of the heavy lifting when it comes to things like player positions and damage and a client's FPS shouldn't affect this, but netcode is a complicated subject and I have no idea what this game handles on the frontend or backend, or how.
- We also can't entirely rule out some deltaTime flarkery, though it seems less like "they forgot deltaTime" and more "the calculation involving deltaTime is acting weird." Maybe they accidentally put in deltaTime compensation twice, or they're using some under-the-hood calculation that does weird things at low framerates. Again, can't know for sure without having the equations in front of me.
If ur not multiple everything with delta time ur committing a crime , it's basic video game programming
@@ThunderMan2212 I see. Now that Ive seen some other examples of it I think its indeed a netcode problem. Some ppl are claiming it looks like the prediction system is getting confused by lower fps players and compensating in a way that causes this bug. But theres no real way for us to know rn
The biggest thing i'm seeing / realizing is that, if i remember correctly the best wolverine player has a "bad" pc. He's being limited.
He doesn't even play on 30fps my guy, also the guy in the video didn't notice but he ended the animation from the grab sooner on the 120fps that's why wolverine was hitting more. I just tested in the training range now, it's pretty much the same.
Also, they're really underselling the difference in kill time with stuff like Starlord. Those are the times where you just got away with a sliver of health, but got healed or killed them at the last second.. but instead, you're gunna die because of that slight change of 'milliseconds that dont matter'.
I don't blame him but this vid coulda been better researched but i don't expect him to know every intricate timing of each character either i would take this vid with a grain of salt
Just buy good GPU lmfao brokies
@@shrekupine44459/11
When strange makes a portal i immediately lose fps
For real 😂 whenever he opens a portal my game turns into 8 bit
R you both on PC? Cuz i have the same issue and my friend on console (Ps5) never had this problem
@gawrgura3767 i'm on Pc and this game takes a lot of resources. I play overwatch2 on ultra setting and run 200 FPS smoothly.. This game i have to lower all the settings just to keep 80 FPS. Consoles are built with the same components making it a network issue most times. PC have many variable parts so trouble shooting is much more difficult
@@ionfilter0304 my ow2 is running on 200 fps aswell.... well ig they will release performance updates if its that crucial
i use his portal to make them lpook at each other and ive crashed people
Since the genre is competetive multiplayer, I can use fighting games as example.
And as mentioned in 6:51 , there's a whole second of increased time for the low fps player.
In ANY fighting game, adding around 62-66 frames _(1.1sec)_ to an attack for one player only and calling it unnoticeable, is clearly a misinformed take on how important every second is in these competetive environments.
In short, that 1 second extra to kill, he's getting a triple Loki healing shot, good job you missed the kill because of your shitty CPU and nothing else.
the 1 whole second is a very big deal, but the 0.1 second difference from 60fps to 120fps isn't so much, that's what I tried to explain there
@@MarvelRivalsMerchant 0.1 seconds isnt a lot but it adds up, if theyre getting healed that could go from 0.1 to 0.3, and over the course of an entire match the amount of damage you do would overall be lower so you might end up with like 5% less ult charge or something.
@@BagelBoi4000at 0.1 there are more things that would screw you over than that. In the grand scheme of things if you are good enough it doesn't matter. The horizontal movement and wolverine DPS however from 30-120 are absolutely terrible.
Exactly what I was thinking. Have a conversation with someone who's trained to hit trigger switches on single frames to link moves or cancel animations and tell them its insignificant.
One second is much more important in fighting games than in team games. Fighting games are so much more micro focused. It’s a bigg-ish deal but in a game with healing and damage from external sources there’s already “randomness” for when a combo will truly kill
You can see this when watching vods of fighting games vs marvel rivals, going back 10 seconds in a fighting game feels like forever and 10 seconds of a team fight feels minuscule
Don't call it a "tech" when it's literally just hardware related
But hardware is technology so it is tech
I mean technically tech is accurate here since it has to do with hardware tech that technically affects the technology in the game. On a technicality.
@@megelegend9968 But we're specifically talking about technique not technology
It has to do with the specific animations themselves, they aren’t scaled up or down, properly for different frame rates when set to custom fps.
Ah yes the hidden Doctor Strange tech: having a better pc
This is such an amateur mistake, not multiplying by delta time in the code
Is it something they can fix via patch?
Small indie devs
@@chrisr4066yeah should be straightforward
Now that is true for once@@thefroggybannana
@@thefroggybannana ahaahahaaaaaaaaaaaaaaaaaa nope
I remember Flats was talking about how shifting forward into ult as strange was broken and I didn’t understand why lol
Edit: I created a video showing in the game engine how and why the bug exists for those who are curious.
Game dev here chiming in to answer why it does this with unreal engine. The reason is because the game latency is taking longer to acknowledge and respond to your player input. Due to having a high latency for game client functionality (not to be confused with ping latency), and low frame rate which creates the high latency, things like holding a button down to initiate individual input requests to fire off the output such as wolverine's slash attack is bottlenecked and has a delay.
This is actually common with hold X key or button to perform Y action.
When using doctor strange's flight ability, the game is initiating similar behavior by scaling the character position to the parameters set in the game code. However, his distance is impacted because of the "real time" script lifetime isn't matching up with the client. In other words, the client is saying it's done because the time has passed but your game response is essentially saying no.
It may sound a little confusing but it's important to remember that the script will execute and the time will finish before the game can truly finish the executed operation to the pawn controller.
I can upload a video demonstrating this issue in a fresh UE5 project as an example.
another "tip" I haven't really seen mentioned is making sure vanguard is turned off if you have league or valorant installed on your pc. Vanguard can absolutely nuke the performance of your game
gotta test this one out ty
+1
@MarvelRivalsMerchant can you test this
this is true although i do not turn it off (because i move between games), vanguard has shown to affect fps on other games too
shit really? I have it off everytime, and only on when I play valorant. Good to know if real.
just so you guys know, these issues were solved by some kid on reddit and he posted what code to change in one of the config files so animations arent so resource demanding. I can confirm it fixed all these issues but the game developers decided to block access to config files and threatening bans if people mess with it. meaning they know how to fix it atleast but the fact that they havnt probably just means they gonna fix all the other issues like dr strange portal causing frame spikes all in one patch. but yah the main issue is how resourse demanding all these animations are and it causes low end pcs to miss hitboxes or suddenly lag effecting animations. sometimes wandas "shift" gets cancled cause the game thinks you are double pressing even when you are not. so many bugs
Don't even think about Xbox S vs X vs PS5 players on 4k vs 1080 TV
So the Ranked in this game can't be taken seriously. It's literally a P2W
@@laWiskaPY the best way to solve this is to force ranked performance to an agreed achieble standard across all platforms but obviously people who spend thousands on equipment would complain and since games cater to them, low end equipment will never work on ranked play. and with games getting more intense faster then affordable hardware can keep up it might be the end for ranked play in most games.
that kinda sucks because i main strange and the flight boost with higher fps can actually be a gamechanger because he can use his ult in midair and cancel it to catch ppl off guard
the extra horizontal movement helps with escaping Jeff ult too if you cant get vertical enough in time
The question is... Do you die faster with higher FPS?
lol
I mean like the wolverine example that depends on your opponent if he or she is running 30 or 120 fps lol
i felt that way honestly i was playing on 60 hz switched to 120 feel like i just die faster
The game feels more smoother. But you should also keep in mind that in fights high FPS drops can lead to higher gpu latency. For instance if you have 240fps and then it drops to 180fps fast, then you can feel it ingame. So I recommend you to cap it to a "good average" FPS for lower gpu latency and more smoothness.
I just think that a game that has its models looking like cartoons shouldn't be so hardware demanding like this.
Yeah, but the game has some extremely demanding rendering tech available, it's an impressive looking game. Style doesn't mean that much, I've seen game sequels that require 2-4x the resources, such as with series that span multiple console generations. BUT, at least on PC you can nearly double your framerate on low settings. Using state-of-the-art Lumen GI and DLSS on Ultra settings, even my RTX 3090 barely maintains 100fps, but I can get 240 on lower settings
Its not that demanding, I can play it with my ryzen 5 5600gt, which is a very mid integrated graphics
@@Yuusfw Depends on settings because 100 fps can be difficult on RTX 3090 if maxed out on Ultra, but 240fps is easy on lower graphics. Moral of the story is people should tweak their settings because high fps is important
You think lighting, polycount, and texture resolution magically become cheaper if you give the characters paintbrush textures?
@@EhurtAfy Its a competitive game. Performance is better than Image quality. People are making it demanding because they want to. No one is forcing them to play full resolution everything ultra. Even streamers with good PC reduce the image quality when it comes to competitive games bc its more fluid
Risk of rain 2 flashbacks...
Exactly what I thought of.
Me when elite spawns are tied to fps
Good thing that wasnt a competitive multi-player game tho
at least it's not really noticeable in this game.. loader flying across the map at mach 7 will forever be funny
Go go gadget 7 million elites by stage 3
Good coverage. Wolverine started attacking earlier on 120fps
I believe this is an animation issue, idk what the distance issue is but the TTK issue is related to how fast the animations play out, star lord can reload in his ult faster when the frames are higher, wolverine can swing more often when the frames are higher. Again the distance thing is crazy but the main issue seems to be the animations and how they were implemented and are involved in the game process thats why its caps at a certain FPS because you cant gain anything more after the cap
Yea i did some testing and its, longer the animation=the most dps will be effected. I bet namor has some effect with his hover.
I get between 12-40fps depending on the map and the position of Jupiter.
But I'm laughing at this because the difference is so negligible compared to Helldivers 2.
In that game my FPS is horrid and your movement speed=FPS
meaning I move so slow it hurts. Everyone can outrun me so I take the jetpack and medic suit with 6 stims so I can (try to) keep up.
New PC coming this Friday, can't wait.
Congrats, man! I'm waiting for the new GPUs to come out so that I can buy an used 7800 XT or something.
Before you you completely take a crap on this guy.. know his coming from apex.. you probably aren’t that good.. give him a shot.
I'm sorry, but I find the Bucky just eating your ult to the forehead at 1:51 just absolutely hilarious
GET DOWN MR PRESIDENT
3:50 arent some of these things just discrepancies due to low fps recordings?
if you were playing at 1fps, those kill times would have been 2 seconds. not because you kill slower in game, but because on the first frame that takes a second the enemy is alive, and on the other frame after another second they are dead, marking the measurement at 2 seconds. on 30 fps this discrepancy is still noticeable, but on 120+ fps the extra frames add very little difference to the recording, capping out at whatever framerate you record on (120 fps in this case?), meaning that the kill time stays consistent
doesnt necessarily have to be true due to the doctor strange flight ability being clearly affected by fps, but if you measured the time to kill by looking back at your recording, then it definitely is the reason (and not the game itself)
By the way, there’s a display setting that shows your current fps in-game. Please use that I’m conjunction with these types of videos in the future. Makes it easier to keep track of things.
Look at the top right. It's enabled.
Isnt it already enabled? It's just small
Already has it up
I can explain it from a developer perspective:
-The most basic stuff is this, some command get calculated client-side and some performed server side, there will never be a multiplayer game that relies everything on client or rely everything on server side, the closest game that relies mostly on server side is Street Fighter V, and people hate that game.
-Looks like from the data, the game treat Button Hold as constant command, meaning if you hold left button, the game will keep sending Left input to the server
-Here's the kicker, when your client (your PC) didn't send any command, the game treat it as "Input neutral" even if you're holding the key
-The number of input you can send is based on your FPS
Basically
Someone with 120 FPS can send "Left" input 120 times
Someone with 30 FPS can only send "Left" input 30 times, the rest of the time, you will be input "neutral" this is why Strange's can fly further left if you're holding Left on a 120 FPS machine
Any character with "reload" animation (This includes Wolverine that can stop attacking after 4 hits) also get affected by this, since 30 FPS guy can only send 30 commands (hypothetical) every second, other than that, the game treat them as not pressing anything, this caused the guy with 120 FPS can start the "attack again" command earlier than the 30 FPS guy.
At the end of the day, this will show on 2 extreme side (30 FPS guy and 120 FPS guy) to show "the worst thing that can happen in that situation"
Most game generally treat this as "whatever", similar to how people will accept that a 30ms guy will have better play experience than people that play with 120ms.
I notice that Laging players also suffer from damage output and received problems… which could potentially stack with the FPS problem.
Eg- you either get free kills on the lagging player as they are stuck in place in you pov, however the dmg needed seems to be a lot more than normal
Or
You shooting the lagging player to then suddenly be chunked for 250hp or more depending on who and the situation.
Point is the servers suck which we know they do…. and or it’s possibly an exploitable problem that people could force.
Hopefully they address this and the absurd ping pikes that's been going on.
This is such an amazing video, the amount of effort put behind the research, the details on the relevant cases, and the tips at the end are all so incredibly helpful
My Nvidia GTX 1050 doesn’t like the sound of this 😪
tried to play on mine cant get past 99 even the steam deck is a little better.
GTX 1060 here
@@MyKlaudijus i get 144 fps but at a lower resolution game still looks great and runs well loading times are longer and a few shader issues sometimes to load but when all in 120 to 144 fps constant
Hoe do you get it to 120 ? Mine is max at 60fps in game. With everything set to low and 720p.. !!@@thepandasdoitbetter
I have 970 and the game worked well before the update (Christmas fest)
yo btw for star lord and punisher its due to the lower frame rates + high rates of fire, on 30 fps you might fire a bullet on the 4th frame and on 120fps you might fire a bullet on the 15th.
this is also more apparent in wolverine as he has an insane rate of fire
It's a feature of Unreal Engine 5. Other games made with the engine also have a higher movement speed & DPS at 120 fps. It's not a massively higher rate in Marvel Rivals but in games where they put hundreds of millions of health on some boss enemy and players do millions of DPS, this small difference becomes much more significant as it acts as a multiplier.
like Uncle Ben said
"With great PC
comes great DPS"
Seems like it has something to do with the difference between animations being tied to framerate, while also being tied to damage output (like bullets firing or claws swiping) yet damage HITS are calculated server-side so if the animations slow down, the attacks are sent out slower, and then the bullets hitting the enemy will be registered further apart (or swipes in wolverine's case), at least that's what I'm seeing.
*For the last part, additional feature to increase FPS:*
You can decrease *screen resolution* (while keeping it Fullscreen) in accordance with your normal resolution.
_For example:_
If your normal resolution is *1920x1080,* you can set up *1280x720* - it would make your game running as a quarter from your normal screen and increase FPS.
Works good in many games, especially competitive ones.
Risk of Rain 2's DLC had this issue where the enemy spawn rate was directly tied to the amount of fps. And it could be a real issue for new heroes moving forward.
Old Quake Engine movement was tied to frames. In Quake 3 there was a sweet spot like 125 250 and 333 framerate. Also on quake 1 and 2 you could only do some jumps with higher and lower framecap.
Seeing this got me to think about Winter Soldier primary fire as well 🤔.
It feels weirder to aim & shoot on lower fps to the point your weapon basically don't register hits on longer range when visually it looks like it should. I do realize his weapon is not a hitscan but a projectile based (strangle design decision) so placebo wise it makes me feel that character could be a victim of this game flaw as well.
If you can, please confirm in your next testing that as well since i'd love to play bucky more if his shots feel consistent in crunch situations.
Note this kind of issue also exists in Mobile Suit Gundam: Battle Operation 2.
The lower framerate, the slower the mobile suit moves in general. Attacking is somewhaylt noticeable, but effects combos players do.
the more concerning thing here, is that it shows there are NO serverside checks for if the player is moving right. that opens the door for all sorts of cheats
This game has serious optimization and mechanical problems
Btw thanks great video
Makes me curious if there’s a difference between PC and console🤔
unless they play on series x or ps5 with a monitor as that is the only way to get 120 fps on console
While I def want this addressed, I think people don't realize how common an issue this is in games. Plenty of games I've played recently and watched in depth testing/guides on have mentioned listed stats don't always line up with in game results.
I am stating the obvious here but if you've got 30 FPS you've got bigger problems than this.
Not the point
Strange himself tanks it.
Stop making excuses for this.
No the problem is this games optimisation, stop making excuses for the massive company you boot licker
The game isn't very optimized. The art style is nice, but graphically there's no reason it should run as poorly as it does.
I actually did notice a movement speed boost but it wasn’t because of framerate it was because of latency or lag. The symbol that shows up when theres a network issue. Its annoying
It's likely animation based events. If you are ticking animations at a lower rate you can get these kinds of low frame rate delayed responses because you are frame staggered from the game thread and the animation thread.
If it were a lack of Delta time multiplication then the impact would continously grow beyond 120 fps.
Then there is the network layer. Does Rivals use entirely client authoritative hits? Do the abilities calculate hits in the server?
The devs can work on this and shouldn't be hard.
For the most part, i feel like some of these are so minute especially when considering all the disadvantages you are already at playing below 120fps. Theae arent the games to take competitively seriously when your hardware cant keep up, unfortunately. The majority of opponents will have an inherent advantage over you that far exceeds a couple of miliseconds. The Dr Strange and Wolverine cases are more agregious though unfortunately
It seems like a deltaTime issue and the developers probably developed console first and forgot not all platforms have uniform fps across the matchmaking pool
Hey man, amazing work. Your testing will be invaluable for the developers. They will sort it out much faster than expected because of your input.
Big Respects to you thank you truly for testing and informing this out
It's probably the reload that gets slowed down then if ironmans uni beam and cloaks primary are unaffected. So your tests should be against a target with more health so you're more likely to need to reload. Assuming this is in a real game and you dont have the luxury of 100% accuracy this could be more significant (but still not massive) than your tests have proven Vs Luna Snow in training. It's a bug they absolutely must fix either way. Thanks for doing this experiment!
I feel like this might be related to hitscan characters. Things like seem to only appear with certain hitscans.
Quite strange that these issues exist when it runs on Unreal 5, yet the flagship Unreal 5 game Fortnite had already fixed this issue back when the game was on Unreal 4.
It looks like the entire game loop is being rendered at 120 frames per second, along with the video update. Technically speaking, if you were only rendering 30 frames per second, you would have a lot more time to be considering things like "does the player have the fire button down.. yes? shoot" It looks like these checks are being made at the same rate as the video updating, rather than at the speed of resource availability.. which means, as you turn up the frames per second, you are also turning up the frequency it checks to see "Hey, is the button down? Yes? send the message to the server that I've fired.."
So some things in the game were coded to react to frame rates which is probably some mistakes or miscommunications in the dev cycles.
Dang. A lot learned while I've been away from the game for a bit. 👏🏻
Honestly this seems like an animation bug more than anything, the one frame of more damage with starlord is completely negligible and within tolerance as a frame is generated every 0.033s, and the damage with wolverine is likely the animation just finishing prematurely leading to massive differences with dps, but it's really just a very small desync with animation timing. I would also deem the distance issues with certain dashes to fall under the same umbrella of an animation or a Lerp finishing a frame early leading to slightly decreased distances of dashing.
Overall, not really that gamebreaking, a couple of small issues but thats about it. If you only can get 30fps there are other limiting factors making you perform worse ingame.
Welcome to the world of numerics, where parabolic trajectories overshoot by a mile, velocity damping makes you go brrrrrr and snappy is actually smooth and jerky if it is frequent enough.
I bet in case of Wolverine and Star Lord the animations have to wait until the end of the frame, but since the frame pacing is so slow (and the server repeats client's actions as is) it ends up adding to the total attack time.
For Doctor Strange it looks like the effect of air friction changes massively. It may have been lazily calculated and forgotten, or some tiny mistake in per-update velocity loss coefficient, or just that the update pacing is so bad that the numeric "overshoot" effect occurs, but instead it slows down too much. And yes, all of those will plateau towards some update rate, as numerics work that way. Anyway, it can be fixed easily.
Venom's wallcrawl/run is directly tied to your refresh rate.
Destiny 2 has infamously had lots of issues tied to fps. Weird game engine issues.
I wanted to say the same thing. In Destiny you do more damage at high fps but also takes more damage. Wonder if you take more damage in Marvel Rivals as well. But I don't think there are many people that play at 30fps lol
@@fransbrewis4776Yeah youre underestimating the average persons ability to deal with 30fps gameplay. I consistently have 30. Now id rather have higher, of course, but my pc is too shit for that.
The strange one is actually a big deal since he uses that tech specifically for his ult.
I was watching a streamer yesterday a_seagull and he seems to have found that the games ping is also tied to fps. Tested this myself and it checks out even in the training range. It's not a cpu bind either since the scaling of ping to fps I nearly linear from 5 fps up to 300. This much fps causes this much latency. It makes no sense
Ping also affects fps. If you aren't able to consistently cap the fps, ping can bring you down to over half of your normal frames
Does the kernel anticheating software reduce framerates on pc?
Riot Vanguard Anti-Cheat can supposedly reduce FPS in your games, so it is possible that this Anti-Cheat does as well, but who knows. It definitely ain't helping FPS though.
7:47 Jarvis clip that
But you can see his teammates are red and enemies are blue so there were no enemies left when the shot landed
So not only that Iron Man flew above your head but so did the joke
What was the joke? Clip a moment where nothing happened?
Yeah
He whiffed his ult
It kind of makes sense because if you're rich you can afford faster monitor and if they keep you happy you will spend more money buying cosmetics.
Dev: no no thats not it. The calculations we have to do to keep the character aligned throughout different frame per second is not easy. there always is some mathematical adjustments that needs to be made.
from programming perspective, frame render thread are in the line with main thread, the higher your frame rendered means main thread dose not need to wait more. that why its more responsive in almost all games. only few dev cares about this.
would be interesting to see how things change if you change mouse refresh rate, also how is this on console?
nothing changes with mouse polling rate (other than increased CPU load which may cause lower fps)
console it doesn’t matter at all
idk if anyone will see this but what he said about frame gen is right but, if you have issues like your game is freezing or fps is just tanking I really would recommend just trying it.
I was really struggling with this games optimisations putting everything on low and it would still freeze but, turning Frame gen on pretty much fixed it for me. However if your hitting 60fps and below when playing you should not turn on Frame gen as you most likely will just get latency, you need at least 120+ for frame gen to properly work I feel.
this is what helped me as well lowkey
I couldn't understand, what are you saying?
"idk know"
@@blindn1nj4 damn bro
@@deadsoldiertr how? I feel like I explained it pretty well and its not complicated. what do you not understand
Can you test it with Namor (orb ability) or Scarlet Witch (ult)?? They both have a similar lift/float animation to stranges fly. Ive felt Namor orb is so slow it just leaves me trapped....maybe its an fps thing
Are you jumping then using orb and holding space to float?
This is a big issue I think that due to lower frame rates it is sending information to the servers slower, so if u have higher fps then it would send data faster but again this would mean that every single characters time to kill would be different but its only on some characters. I hope this gets ironed out and fixed and I also know about some other bugs I have encountered on this game since I don't have a strong enough pc to run this game on higher fps.
This is actually a massive problem and something that should be addressed when it comes to official games. What should have killed someone and it didn't because the computer was average specs and not giving 120+ frames and that lowers damage. Ya this will make A LOT of problems for official stat tracking.
Might be better if you actually animation cancelled Starlord's ultimate start-up by just holding your primary fire down first.
Playing as rocket ill paint someone with fire for a solid second before it start registering, especially at range, theyll be taking damage behind cover, vise versa
It makes sense, because you are six universes ahead of your opponent.
Maybe this has something to do with frame rounding on low hertz servers, like how the old cod games suffered
Don't worry guys I'm just testing something out for UA-cam I'll be out to help in a second 😂
Not gonna lie I laughed a little when you got in game and did it
Interesting that if this where in Overwatch people would be picking pitchforks already XD
this makes complete sense as to why i stopped playing the game, it was infuriating and nobody would believe me that it was happening, i had no idea what was causing it but now i know, cool.......I know im due for an upgrade but come on, not only is UE5 not ready for games yet (no game should release that cant hold a steady 144fps at 1440p on lowest settings, which is the AVERAGE across steam users) but FPS based mechanics existing in 2024? Come on, this game had so much potential.
I definitely notice a difference, just upgraded from a 1070 mobile to a 4070 super. It's like night and day.
About the wolverine combo part when you showed 120 fps you started attacking groot mid air during your feral dash animation and on the 30 fps you started attacks when you were towards finishing the dash. was the dash cancel mid air caused by the fps change? or did you do it like that without noticing?
They should have used fixed timestep to keep deltatimes consistent, but Unreal Engine doesn't offer one as it is not designed for deterministic games. That's the problem. The easiest fix is to just cap framerate to 60 on the engine side.
I'm fairly familiar with Unreal Engine 5 and there are a few different ways this could be happening depending on how they calculate the passage of time in the game. It also seems to be calculated differently for certain characters or abilities as well. I'm very curious as to what is actually happening behind the scenes in engine. I highly doubt we will ever get a peek at the code though...
With black panther, his dash, his most useful ability, gets a fair bit of extra reach when limiting your fps to 30.
Uninstalled, they really need to fix these inconsistencies before it can be competitively viable.
good video! however i believe to truly make a conclusion, a macro should have been used to avoid human imperfections.
It feels like iron man team up with hulk is affected by FPS after all. Had one in ranked as iron man myself... Well, I was noob but it really felt like he dealt damage faster
I wish youd just show the clips side by side, would make the comparison easier to see
I didn’t know that strange could go that far horizontally bc my pc sucks. i thought that ability was mainly for vertical movement😂
yeah the doctor strange and wolverine thing is absolutely busted. is definitely a major advantage in mirror matchups whoever has more fps
We only just started hearing about this over the last couple of days lets be honest.. Hardly a game breaking issue!
thats insane how is this not fixed also games starts lagging when using constant abilites like glichy lag
+the lag compensation and the tick rate are absolutely awful, like battlefield, but for a competitive FPS
Yes, NetEase’s NARAKA has the same problem, it’s very terrible
In case it wasnt obvious, Starlord's ult should basically be the only one that matters, strictly because it applies its damage across so many hits.
That's a small sacrifice I'd be willing to make in order to have the game more optimized when its already pretty modestly optimized.
Would be verry different if you were shaving off .1s TTK for every weapon but I think its just his ult.
Also wondering if it impacts Iron Man's flight maneuverability since that is super reliant on acceleration.
That’s it. I’m going out buying an RTX 5090 gpu for this.
6:31
I like how he's gaslighting players that 1 second is a little bit of time in a shooting game
It's the ultimate form of pay to win!
Can you make a video on the best video setting for the 3 main types of graphics people prefer?
For example: Performance, Balanced, and Fancy
DLSS
Turning super resolution sharpening down lowers my FPS
Everytime I use my ultimate and I die before it fires, it just disappear and i lose my ultimate completely. All because of my fps drop on pc. It never happens on my ps5.
9:00 Frame gen needs to go off because it doesn't fucking work. Test it. Lock your FPS to 30, you know how choppy it looks. Now turn frame gen on, and it looks the same. Its not only input latency but also doesn't generate any frames at all. It is broken! Idk if DLSS frame gen works, but FSR for sure does not. If you have Rivatuner you can see your frame time going wild too when you turn it on. Which shouldn't be the case if it worked properly afaik.
Also frame gen in general is fine if you have minimum 120 fps but a 240hz screen. The input latency isn't that noticeable. Even 60 base fps is fine if you really wanna smoothen things out.
Man im poor, and my pc is poor, so this is so unfair
My computer recently started having issues and im on 20 fps and im a pretty good play and i realize my oneshot combos aren't working because of the delay between where the enemies are on my screen from where i recall the shurikens
Hope they fix these minor issues. Im not autistic enough to think 0.01 second difference will do anything in my games but might screw over some matches for few players in the world. Id rather have it run on 60fps silky smooth than with 120 where it randomly drops between 80-120 depending on what kinda ults someone is throwin around.