Built in input lag?

Street Fighter V
81

Seriously, can’t be arsed to read a a few articles by experts? What does the dates have to do with it? If I write an article for basic snap action microswitches and 10 years later microswitches are still the same (actually, the ones we use in arcade sticks have been around since the 80s at the very least) is it not relevant because of the date?

You don’t have time to read, but you have time to sit on a forum and argue about things you don’t understand?

Oh, but hey, it’s not like that article wasn’t written by an expert in his field, nah no way he could be a veteran programmer who knows his shit. I mean, he didn’t working on twitch based arcade racing games or anything like that… -_- How much programming experience do you have? How many games have you programmed to run at bare minimum input lag?

Hilariously you’re wrong. HDTV standard (you know, what people use to play PS3 games on…) has a native refresh rate of 60 Hz and the camera records at 60 fps. He counts twice because each refresh rate of the monitor syncs with 1 frame of the game at 30 fps. You can then calculate in lag time on a 1/60th of a second time duration of 16.7 ms, or each refresh of the screen being drawn regardless of the game’s frame rate being lower.

This is pretty much a standard. 12 frames of lag at 60 fps = 200.4 ms, and 6 frames of lag at 30 fps still equals 200.4 ms… 1 frame at 30 fps does not equal 1 frame at 60 fps. Learn to do some math please. Learn about HDTV standards please. Read and listen to experts in their fields, please.

Where did I ever write wonky hardware? Please quote me.

Again, what are the controls for those games? Are they vsynced to the refresh rate? Are they unsynced? If unsynced, then as I’ve already written previously you can render many more frames per second. Many FPS tournaments are set up that way, using high refresh monitors, unsynced and uncapped FPS, and extremely low resolutions to avoid taxing the GPU for the fastest twitch input times possible.

Whether you like it or not, FPS is tied to input latency. Why do you think input latency jumps to 133 ms when 60 fps games dip down to 30 fps? Why has there been things like Triple Buffering created to avoid that, at the cost of a little extra latency? Why did companies like nVidia come up with G-Sync so that they can maintain input latency with no screen tearing even when a game dips below a target framerate as an improvement over vsync+triple buffering? Why is there things in my nVidia driver control panels for shit like adaptive vsync which maintains vsync until the GPU is being stressed then drops it and allows tearing to avoid input latency spikes? Oh no, you’re not telling me all this stuff exists just for shits and giggles and it has nothing to do with input latency??? I mean, why would any company spend $$$ doing R&D for such things, we have experts like you!

Look at nearly every single 30 fps input latency test. They’re usually up around 133 ms unless they have massive amounts of post processing creating further latency. Look at nearly every 60 fps input latency test. Usually down to 66 ms. There is literally years of data and tests to back these up. Since things haven’t changed and we’re still using tried and true methods like vsync or no vsync uncapped FPS, these tests and data are still valid. Old school arcade games updated at 60 Hz and we have some good people in the ST community and TT community who have tested actual, real, not emulated by MAME arcade hardware at 4 frames of input lag, or, by the 60 Hz standard which those games were made to update at and which the CRT monitors they were attached to updated at, of 66 ms of latency.

Nah, but you know better. You’re the expert. You have tests and data and experts in the field backing you up. Oh wait

82

I almost decided not to answer at all because of your hostile & edgy sarcastic reddit attitude.

Yes you got that right and still couldn’t limit what you wanted me to read.

I know that TV refresh rates are around 60Hz. I know that the camera recorded at 60fps. I know that the game runs at 30fps. (frames per second). This is not some sacred knowledge that only ‘experts of the field’ know.
In the video, there wasn’t a latency of 12 frames in the game except in the 60fps camera footage, and this is where he went retarded. The game (which was 30fps) had a latency of 6 frames. On his camera, which was recording at 60fps, it looked like 12 frames.
Now, did the game have an input delay of 6 frames or 12 frames?

Alright you were right, you didn’t say that literally. You said:

which I find pretty fun because when we’re talking about experts of the field of input latency testing, I’m pretty sure he comes near the top of people to look for.

I don’t know why his Xrd PC test would need to be verified? This gif literally speaks for itself, it shows Xrd on PC with 1 frame lag on the right settings. And with these settings, the game is STILL capped at 60fps.
It’s not displaying 200fps, it’s not displaying 100fps. It’s still 60fps. You can check this out by yourself by enabling FPS visibility from options of the PC settings and using the same settings. Here’s even a screenshotof that, pay attention to the top.

…and would you look at that, it’s still only 1f delay, in a 60fps game. How do you explain that?

Here’s something I recorded from another game also capped at 60fps as the game speed is completely tied to framerate. This is made using pausing and resuming the emulation for only 1 frame:

As you can see, in the game itself there is no lag at all. Everything that comes when playing live is from your monitor, controller, computer or whatever, but the game itself is ultimately lagless. It’s Ketsui running on old Shmupmame 2.2 btw.

I don’t know why it’s so hard for you to accept that this is possible. I’ll be honest, I’m not an expert when it comes to coding and I have no shame in admitting that, but I don’t need to be one to observe this material above and conclude that the claim of ‘3-4 frames of lag being unavoidable game-wise’ is false. That is exactly the problem with SFV’s ‘8f’, on top of whatever else might cause lag, there are several (2-4?) unnecessary extra frames of artificial lag in the game itself that cannot be minimized even if you had the perfect setup.

Yeah, those games can have that high amount of lag especially when you take into account controls and the cabinet, just like SFV has around 8 frames when captured from PS4. And?

83

Seriously?

So you admit you aren’t an expert, and when a veteran game developer writes an article about input lag and minimizing latency at 60 fps you go and make the claim it’s false. Why? You have no experience in this field and the person who wrote the article does. This is literally the same thing as an experienced biologist telling you how cells replicate but you claim it’s false because you read on the internet that cells don’t replicate.

facepalm

Where in the chain of an arcade cabinet does input latency come from?? Please educate us with your in depth knowledge of arcade cabinet technology.

84

For fuck’s sake.

You’re essentially saying that I should rather listen to some ‘expert’ instead of looking at the factual data right in front of my (and your!) eyes… which btw are also provided by an ‘expert’.
Well I’m sorry but I’m taking the latter, and I’d recommend you to do the same.
I didn’t “read on the internet” that a game’s own lag can be under 3-4 frames, I experienced it myself, I saw Teyah’s test results that prove it, etc.
Do some extra frames of lag just magically appear in these cases if an ‘expert’ comes in and says that “under 3-4 frames is not possible”?
Is black magic really relevant to how you’re disqualifying Teyah’s Xrd tests, or are you going to say that he has edited the footage next? Or do you finally admit that the game actually can have 1f lag while still staying 60fps?

TLDR: You don’t need to be a ‘veteran game developer’ or an ‘expert’ to observe this material and make conclusions based on that.

When it comes to what the ‘expert’ wrote in the eurogamer article; he was probably talking about designing the game with some of his own criterias, like running the game on a specific platform and control method, while maintaining certain graphical quality.

Probably from the PCBs and wires of the controllers and the game. They’re usually very fast though, so most of ST’s lag comes from the game itself.

85

Which is it, games have no lag, or games have lag? It can’t be both. Are you aware of what you’re writing at this point?

86

Skip that entirely if you want, and just answer this:
We have some test results where the games ran at 60fps and the results of input delay tests are still under 2 frames (even counting display).

Are these results:

  1. True, it is possible
  2. Edited / they’re all lying
  3. Magic, it shouldn’t be possible because a veteran expert said so

EDIT:

I never said that games have no lag at all. I said that the lag can be as low as 1f, and that 3-4f isn’t a requirement unlike you claimed.
The games can have different amounts of lag, depending on how they’re coded. This is what we were talking about. In my example, the game renders on the next frame from the input, while ST (unaltered) on arcade might do it on 3rd frame.
Ketsui on shmupmame and ST have different amount of lag.

Did you seriously take two quotes out of context just to confuse yourself, or did you not understand them in their original context?

87

Not a single thing you’ve said, Cagar, doesn’t mean that the thing should not be peer reviewed. There is plenty of science out there that has video evidence but they still get peer reviewed before they are published, like say, for instance, the EM Drive.

It’s simply good science.

88

Since this is a pretty hot topic right now I felt like testing the latency of various fighting games using a different method that doesn’t involve any controller or graphic related delays (such as gamepad converters or V-sync). It’s supposed to measure the “raw” intrinsic delay imposed by the game mechanics.
The method consists in simply playing games in super slow motion, with Cheat Engine’s Speedhack function set at 0.02, or Frame Advancing through emulators. I press a button and watch when the game responds. My results:


GAME:	Fantasy Strike
TOOLS:	in-game frame advancing (Shift+1)
DELAY:	0 frames

GAME:	Skullgirls 2nd Encore
TOOLS:	Cheat Engine
DELAY:	0 frames

GAME:	Dead or Alive 5 Last Round
TOOLS:	Cheat Engine, Frame Counter tool
DELAY:	you can Guard incoming attacks with 0 frames of delay, but the rest comes out 1 frame later possibly to prevent kara moves

GAME:	Street Fighter III: 3rd Strike
TOOLS:	FBA-RR, input-display.lua
DELAY:	1 frame

GAME:	Ultra Street Fighter IV
TOOLS:	Cheat Engine, Frame Trapped IV
DELAY:	1 frame

GAME:	Street Fighter V
TOOLS:	Cheat Engine
DELAY:	2 frames

GAME:	Street Fighter Alpha 3
TOOLS:	FBA-RR, input-display.lua
DELAY:	3 frames

GAME:	Street Fighter II: The World Warrior
TOOLS:	FBA-RR, input-display.lua
DELAY:	3 frames

If you’re suspecting that Final Burn Alpha is adding that 1 frame of latency, well, I’ve managed to make an emulated game react with 0 frames of delay (more specifically, releasing a direction in KOF98 happens instantly)
If you believe that this method is bad and that I should feel bad, feel free to let me know.

89

cagar the new tali

ITOH cagar in spanish is funny because means “taking a shit”

90

What does input lag do exactly? From the time you press a button to the time your character performs the corresponding action, there is a difference of less than .2 seconds? It seems like something that may be valid but I usually just blame myself for execution error, correct my actions and adapt to the situation. Practicing and continuing to play seems to make it non issue because I get more and more consistent (execution) the more I play. also from looking at other posts… I guess others do not like the game enough to want to adapt to this?

91

People can argue the exact values all they like, but the only thing that matters is I can feel the input delay when I play the game. This is especially true on the PS4, which I don’t play on often, but when I do, it is jarring. Maybe it’s 8F, maybe it’s not, but it’s at least 5 or more, which is already terrible. I start to feel delay around 3 frames or more, but it doesn’t get unplayable until – well, around 5 or more. So on PC, SFV is just barely in the playable portion for me, which is really unfortunate.

In regards to the whole “Oh, it benefits this playstyle more” sort of thing. Whatever, it effects everyone, and while brain-dead setplay sort of benefits the most, I’d admit, it hurts everyone. Yeah, it sucks for reactionary players, but it also sucks for players with bad reaction times like myself. I usually block at the last possible second, and I usually tech throws on the last frame – now, instead, I am getting hit or thrown, not because I failed to react, but because the game failed to put the input in fast enough. That is very annoying. I’m sure most people here can sympathize with this feeling in some regard. I hate looking down at my stick, seeing it in downback and looking up to see 5+ hit on screen and half my health gone. “Whelp…”

God, this is a fighting game, the genre built entirely around frames, and around reacting to those frames! Why on earth did they decide to do this? SFV continues to be this bizarre regressive game that blows my mind. Why is it that as fighting games get newer, they get further and further from being an improvement? Hitboxes should be matching what’s on screen more and more, instead it’s less and less. Input delay should be decreasing, instead it’s increasing. Complexity should be increasing, instead it’s decreasing. I said this in my random Ibuki thing, but I’ll say it here too, SFV looks like what fighting games should have started as: simple.

Maybe it’s just me, but, fighting games started out okay, got to this awesome point, then went into this really big downward spiral (exceptions exist). Why are we letting fighting games go in reverse!?

God, GAAAWD! GRRRAAAH! There’s no reason for it to be this high, I’ve played games with better netcode with infinitely less delay! And that’s only when it’s online! This is all just Capcom being bad. They think they are good enough to make their own shit, but they’ve proven time and time again that they are terrible at it. They can’t build a good netcode. This is reminding me of PoE, and how desync was literally built into the very grounds of that game’s code and it took GGG yeeeears before they had the funds to rebuild the code from the ground up.

This is just how the game is. There will always be people who try to justify it, but in the end, we all know deep down it’s just Capcom being stupid. I don’t doubt there is some actual reason that the game has to have the delay it has, but you know what? I bet a better person at making netcode, could have had it running better with far less delay. (And I know they could have.)

(At the least, can we reduce the PS4 delay so I can actually go to local tournaments and not feel like I’m playing people on the west coast when they’re SITTING RIGHT NEXT TO ME! Fuckin’ goooood!)

92

glad I waited to drop loot on a PS4/game/stick. They need to fix or its old school for me.

93

what if i i told your dumb ass that it’s a reason why I win also? nobody discussing is using it as an excuse, and I’m sure most people aren’t under the delusion that the input delay is keeping them from winning evo. so tired of this garbage response to a real issue.

94

I remember the Dreamcast port of ST having 4 frames of input lag and that being enough for it to be banned from tournament. OR am I remembering this wrong?

I don’t know how much input delay arcade games on CRT have, but my guess is Zero frames of input delay (or 1, if you want to get technical all video games have 1 frame but it’s not a delay, it’s the limits of the technology).

In arcade I can red parry easily in 3S. On ANY home console version on CRT I cannot red parry, with exception to Xbox 360 version, where I can on certain things, but not as well as arcade.

The window for red parry is 3 frames.

PS4 and PC being reduced to 4f of input delay would be fine. I’ve played with VSync forced off, the 4f PC version, it’s an enormous improvement. Night and day. It actually felt like Street Fighter again. If it can be even better than that, cool.

8f is unacceptable, it’s the same as USFIV when it dropped on PS4. Remember that controversy? Only reason there isn’t a similar controversy now is because people didn’t have something better to compare it to (like with 360 and PS4 versions of SFIV). Hell who remembers all the bitching over PS3’s added 1 frame?

Also KUDOS to the man who made this video. THANK YOU https://www.youtube.com/watch?v=3CBEVIDwge0

95

(This whole post is like a rant based on the crap I’ve seen posted elsewhere, and in every topic like this.)

Really, I am fine with all of the people using input delay as an excuse for their loss. I want them to keep saying it. This topic needs to remain alive until season 2 is finally being crafted. Capcom needs to keep this in mind, and I don’t care if it’s simply whining that keeps it in there.

The most perplexing thing about this topic so far, is that there are people defending it, with the most asinine retorts. Like how Capcom would have to change the WHOLE game, and that the frames are built around the input delay. They’re not, and the frame data (while very formulaic) is not that different from other games. If anything, due to SFV’s extreme lack of active frames, you could argue it’s even harder to whiff punish because there are often times less total frames.

Example using Ryu’s cr.HK, something which should always be getting whiff punished:

SFIII: 7/5/25 (37 total)
SFIV: 5/4/28 (37 total)
SFV: 7/2/22 (31 total)

6 frames faster, and of that, most of the decrease in frames is on the recovery. So why on earth do I see people making posts in places about the frame data being based around delay? It’s absurd. At the speed SFV is running at, it should have comparable input delay to SF3 or even lower (not really possible to go lower than arcade version).

Why defend this? Even if you think it’s fine (it’s not), the simple way of looking at is this: less input delay = more responsive game. Less input delay is nothing but a good thing, it doesn’t hurt any twisted game design philosophy you might’ve thought up while you were trippin’ some serious balls, binging on fuckin’ Arcana hearts 3 and masturbating to Konoha (she’s like 10 dude – the fuck!?).

If it’s not apparent, this topic gets me heated. It’s just one of those topics where I feel like there should be no people defending it, but there are. It’s like someone is rallying against me trying to feed the starving dog at the side of the road. Why? Seriously, Capcom could have like a horde of slaves in their basement and people would stand at the door, and be like “Oh, well, maybe they like being slaves?”

Stop letting people shove crap in your mouth and being like “Oh, wow, I really like the corn you had last night, Capcom!”

Gaaawd!

(Apologies if this is like a flag-worthy post, but, I’m in that kind of mood.)

I’m gonna be a stickler for facts and fix that. Even though I know you know.

Also yeah, I approve of that video’s logic in regards to applying the effects of input delay to the mechanics of the game.

96

If the lag is intentional it shouldn’t be hard to fix. if the lag is unintentional it could be very difficult to fix without potentially re-writing HUGE sections of code. This isn’t like fixing the port of USF4 on PS4 where the underlying game engine is already set up correctly and they just need to fix their own mistakes. Part of the game engine’s code may be written in a way that causes the lag and requires a huge re-write.

If it’s intended they most likely wont fix it. If it’s not then they probably wont because of the work. I think it’s gonna stay.

97

Capcom Presentation From The Future

Combofiend: “Okay guys… no more jokes this time! We are introducing our way to help resolve the input lag problems. Introducing… the new kamigusha type 2100d. The game will run two separate games, one with v-sync off, and another with v-sync on without any drastic models/animations going on. It will collaborate the two and result in a rollback feel to even your offline matches to perfectly match the offline moments of the game. With this, you can now rollback offline, while you are rollbacking online. This is Combofiend signing off, and now to bring in Ono for a Shoryuken photo moment!”

98

@skiegh

I see your point, and i wanna rant and rave and make a big show about it. But i play the game and, it does not matter cause most everuthing i did before i can still do. Then Momo wins a tourny the way he did and its like, damn…how much of an issue is this really?

99

It’s weird because I am tempted to believe people about it being “game breaking” but on the other side there is japanese players who did not bitch at all and I saw plenty of footsies, one hit confirm, whiff punish, Ken run being punish sometimes with crush counters and whatnot at Stunfest.

100

Wasnt piledriver 3 frames in SF4? Its hard to say the issue with trying to use piledriver to punish during footsies is entirely to blame in the lag. Those extra frames make a big world of difference.