Input Latency on Consoles Considered Harmful

Slade

[14] Master
As many of you probably know, Soulcalibur switched from an in-house-developed engine to Unreal Engine 4 for the development of Soulcalibur 6. UE4 has a problem with input lag. Input lag is the gap between your controller’s hardware sending the signal for an input and the game updating to act on that input. At launch, Street Fighter V had about 8 frames of input lag, and MvC:I had about 6. Tekken 7 has about 6–8 frames of input lag on consoles (slightly more on PS4 than on Xbox).

On PS4, Soulcalibur 6’s input lag feels to be in the same ballpark as the games above. (If anyone has the tools to get exact data, it would be very helpful to the community to bring awareness to this issue. Setting a video camera to capture at 60FPS and recording the delay between an input and the response on a decent gaming monitor should be sufficient to do this.) Players switching from PC to console or vice versa report noticeable differences in the responsiveness of the game, and this affects gameplay. “Borderline reactable” lows such as Grøh’s SCh 1AK or Azwel’s Ax 6B+K become unreactable when you have 6–8 fewer frames in which to block them. Tight whiff punishes cease to be punishes. Moves that should be GIable on reaction become unreactable.

Players will learn to unconsciously shift their playstyles to adapt to this, but the result will be worse: playstyles that rely less on tactics that demand quick reactions and more on flowchart-style algorithms, designed to minimize the amount of quick responses the player has to make, will start to rule the roost. (Expect Azwel, Ivy, Voldo, and other characters with “vortex mixups” to be high in popularity whereas characters who primarily focus on spacing, punishment, and other Soulcalibur fundamentals will wane in use.) The game is simply worse the more input lag it has.

This poses a dilemma for tournament organizers. They are forced to choose between taking the added burden of providing PC setups to avoid this issue, or to provide an inferior version of the game for players to compete in. It will also increase demand from players to increase non-finals matches from best of three to best of five in order to help reduce the effectiveness of flowchart tactics and brain-dead gameplay. This will create a tension between TOs and competitors and shorten the competitive lifespan of the game.

It affects online play as well. Online has an additional fixed amount of lag (more the worse the connection is) added on top of everything to act as a buffer for variations in network latency over time. But the effect of this lag plus the native input lag is that it feels like the characters are trying to fight underwater, even in four- or five-bar connections. Plus, it messes with your muscle memory, so just frames that you practice offline won’t work online. (Practice Geralt 4A:A in training mode for a while, then go try it online. Doesn’t work, does it?)

I’ve heard it proposed that in at least some of these games the input lag is deliberately added (or not properly optimized away for release) so as to create greater parity between how the game feels offline and online. If this is true, it’s a terrible mistake. Not a single serious fighting-game competitor will tell you that intentionally adding input lag to the game is a good thing. And what competitive players think matters. With the growth of e-sports, Twitch.tv, and in some cases the televising of fighting-game tournaments, having tournament organizers continue to run a game and players continue to play it provides long-lasting marketing value and word-of-mouth for the game in question.

In summary: Soulcalibur 6 has some serious problems with the amount of input latency built into the game, especially compared to previous installments. This makes the game less fun to play and less serious competitively. It creates a greater divide between players on PC and consoles, between players and TOs, and will shorten the life of the game. It needs to be patched, the sooner the better.
 

sytus

[11] Champion
The lag is put in because on average hardware and online infrastructure isn't sufficient yet. As someone who remembers Tekken 6 online and offline I can tell you that not having a uniformed experience is a nightmare. Personally we need another 20 years of advancements before we can get back to how the input was in the old days.

Also Tekken 7 lag has been reduced from 8 frames to 4/5 frames on PS4. This was done nearly a year ago, those issues with the Unreal Engine have been fixed.
 

Slade

[14] Master
The lag is put in because on average hardware and online infrastructure isn't sufficient yet. As someone who remembers Tekken 6 online and offline I can tell you that not having a uniformed experience is a nightmare. Personally we need another 20 years of advancements before we can get back to how the input was in the old days.

Also Tekken 7 lag has been reduced from 8 frames to 4/5 frames on PS4. This was done nearly a year ago, those issues with the Unreal Engine have been fixed.
From what I've seen, the built-in input latency from UE4 was reduced from about 4 frames of lag to 2 (for games running at 60 FPS). If the game itself is adding another 4–6 on top of what the engine is doing, the game still has problems. This might be the case with SC6, because it feels worse than 4 or 5 frames.
 

HappyColour

[09] Warrior
I personally like adding input delay online to add uniformity, I prefer consistency over responsiveness in a 3 to 2 ratio. I should say, I have been more successful online in this version of Soulcalibur than any previous version, so this format seems to be working for me.

That being said, I would be interested in knowing the difference of input lag between PC, PS4, and XBONE. A difference across systems would be unfortunate in my eyes...
 

MONEYMUFFINS

[13] Hero
A difference across systems would be unfortunate in my eyes...
IT'S ALREADY KNOWN THAT PC OFFLINE IS FASTER THAN PS4 OFFLINE.

SOME RETRO CONSOLES HAVE 3 FRAMES OF INPUT LAG. MODERN CONSOLES ARE PUSHING 8 FRAMES OF INPUT LAG, WHICH IS A JOKE TO SERIOUS PLAYERS. COMPETITORS DON'T SPEAK OUT NEARLY ENOUGH TO TOURNAMENT ORGANIZERS.
 

HappyColour

[09] Warrior
Would love to see some references on this MONEYMUFFINS. Any documented tests? Online vids showcasing frame differences? Etc...
 

damn-I-Suck

[12] Conqueror
I wasn't even aware of this. I don't even notice this and I'm on the ps4. I've been playing SC since 2 and this game feels no different in terms of input lag. I only notice a lag with GI, which I've kinda adapted to already.

Of course this is all anecdotal. But cmon, if we're talking about 8 frames were talking about 1/7.5 seconds, or .13333(ect) of a second. Slightly more than a tenth of a second. I'm surprised you guys can 'feel' that.

It is a legitimate argument though, esp for whiff punishing and such, all I'm saying is that It's not even noticeable to me and probably others as well.
 

b4k4

[10] Knight
8 frames is a pretty big deal for "reactable" lows being blockable/not blockable on reaction.

And I know this will sound silly, but I really can feel the difference in minor amounts of delay, like playing on a lagless monitor vs. ones with higher delay.

"Casual" players may not notice (and I'm not using that term derisively), but competitive players will definitely notice. That's why there are standards for competitive lagless monitors that get used by all major TOs.
 

Slade

[14] Master
8 frames is a pretty big deal for "reactable" lows being blockable/not blockable on reaction.

And I know this will sound silly, but I really can feel the difference in minor amounts of delay, like playing on a lagless monitor vs. ones with higher delay.

"Casual" players may not notice (and I'm not using that term derisively), but competitive players will definitely notice. That's why there are standards for competitive lagless monitors that get used by all major TOs.
Right. You might not notice the lag directly, but you'll notice when things that you can normally do consistently — hit-confirming strings, whiff punishing 2As, blocking certain lows, breaking command throws in reaction to being grabbed, GIing in reaction to your opponent's attack start-up — stop working when you switch to a setup that adds lag somewhere, be it a TV with high rendering latency or a console with high input latency.
 

Hawkfalcon1

[01] Neophyte
I knew I wasnt tripping. This game hasn't had that smooth feeling since release. Even offline it's just felt a bit off. I just thought the online was a bit weak but if it's an engine issue can this be fixed?
 

MONEYMUFFINS

[13] Hero
Would love to see some references on this MONEYMUFFINS. Any documented tests? Online vids showcasing frame differences? Etc...
SEVERAL PEOPLE HAVE TESTED CONSOLE LATENCY OVER THE YEARS. ORIGINAL NES HARDWARE IS THE FASTEST I'VE HEARD OF SO FAR.

HERE ARE POSTS RELATED TO THE TOPIC: