Can literally be a difference of defending a low to not defending that low
Tekken 7 - PC/PS4/Xone, June 2nd 2017 - Page 3
Forum Index > General Games |
Foxxan
Sweden3427 Posts
Can literally be a difference of defending a low to not defending that low | ||
Spyridon
United States997 Posts
On June 21 2017 08:18 chocorush wrote: Tekken 7 is native to PC, and the input lag difference is even more if you disable vsync. Fighting game players have been playing on sub 16ms lag monitors ever since EVO switched to HD monitors. Input lag has always been a big deal, I don't know where this revisionist history is coming from that fighting game players didn't care about it before SFV. Well, it was around 2010-2011 when people first started making a bigger deal about input lag. But ever since CRT's started going out of style 10 years earlier, most people weren't even aware. Even the first few years of SF4 people didn't have any clue of what acceptable input lag was. You can even do a range search on google, and there's very few people that were even aware of the concept. Notice how all the input lag charts typically start on the LATE ps3/360 generation? Not even early in the generation? As mentioned earlier, I've personally ran in to many people who did not even know their TV's had a gaming mode, and they were playing with ridiculously noticable input lag, and did not know anything was wrong. More people whos houses I visited and seen their gaming setup DID NOT have game mode on their TV's, than those who did. Would you say these people should be complaining about the games input lag at all? When their TV's are adding far more lag? Unless you specifically purchase a gaming monitor and/or intentionally get one with low input lag, your likely going to be playing on a system that introduces a lot of lag. Out of the gamers those of you here are friends with, how many of them actually have a gaming monitor? Sure , tournaments may use it as standard, but if you are not playing on a low input lag monitor at home, your timing will be off playing at a tournament. All I'm saying, is people don't really have any room to complain about input lag that is natively in games, until you make sure your hardware devices and monitor are not adding any input lag. Because the native lag is there for everyone, but I would bet that far more than 75% of gamers out there who play fighting games have more lag from their hardware than the games itself. This makes it just seem like a crazy situation to me, when you have people making a shit storm for under 2 frames of lag introduced native to the game, acting like it feels completely unplayable and ruins gameplay, when the majority of those people have hardware that's adding far more than 2 frames of lag. Some TV's when not in game mode I've seen have almost up to 0.5 seconds of input lag. On June 21 2017 18:11 Foxxan wrote: 128 ms input lag is HUGE. Can literally be a difference of defending a low to not defending that low Another example of what I mean. This is very misleading. In reality, it's PS4 version is only 27ms difference from past Tekken games. People are acting like they have 128ms less time to react, when that is NOT the case. How many of those people with complaints have tested their hardware? Let's see how many of their setups do not have at least 27ms of lag coming from either sub-part input devices, and/or monitor setup. On another point related to this.... How many people in the past didn't realize that their sticks were adding more lag than a couple frames compared to the native controllers? Yet people still swore by stick. How many people are still using things like brooks converters? Even at tournaments how often do you see them? Right there they are adding nearly 2 frames of input lag. In the PS3 generation, most sticks were over 20ms natively, and converters were 25+ms. Can you honestly say these people have room to complain? And doesn't this at the very least show that it's not as big of a difference as they are making, if these people DIDN'T EVEN NOTICE A DIFFERENCE until it was pointed out to them? | ||
chocorush
694 Posts
Most FGC communities used CRT's well past Evo switched to the VH236H because it was expensive. We definitely knew when we played on laggy TV's and why the game wasn't playing as well. Our crew played on CRT's for a good chunk of Soulcalibur 4 and Street Fighter 4 because only one of us owned a VH236H. Marvel 2 was always played on CRT's. Smash players still swear by CRT's. Sure there are casuals that don't know there is lag. You can't just project that on the FGC as a whole, because that is patently false. | ||
Foxxan
Sweden3427 Posts
https://www.humanbenchmark.com/tests/reactiontime Go do this and then add a 128ms lag on top of it. Now this test might have lag cuz we have different internet and what not. But the point is simple, its a big deal. And as chocorush is saying i have also read that the technology has gotten worse in this matter as well. I am not expert or any kind, iam just saying what i have read somewhere on the internet. In reality, it's PS4 version is only 27ms difference from past Tekken games. People are acting like they have 128ms less time to react, when that is NOT the case. I dont understand? If you say its 128ms then its not 128 ms? | ||
Spyridon
United States997 Posts
On June 22 2017 07:40 Foxxan wrote: 128ms lag is a big fucking deal no matter how you put it. https://www.humanbenchmark.com/tests/reactiontime Go do this and then add a 128ms lag on top of it. Now this test might have lag cuz we have different internet and what not. But the point is simple, its a big deal. And as chocorush is saying i have also read that the technology has gotten worse in this matter as well. I am not expert or any kind, iam just saying what i have read somewhere on the internet. I dont understand? If you say its 128ms then its not 128 ms? TTT2 was 101ms of lag. People were not complaining about input lag. T7 is 27ms more. Now all of a sudden, "128ms is a big deal". This is why I mention hardware. It's pretty likely that the lag introduced by their setup exceeds 27ms. But they focus on the engines inherent delay, when their own is the larger issue (and the reason they can not block as fast). If 27ms was really so important, those people complaining should be throwing a few hundred dollars in to their setup. If they don't have a lagless setup, they have NO ROOM to complain. This is also discounting that engines tend to delay some inputs for smooth netcode. An example - GGPO adds delay to button presses, but has a REDUCED delay on blocking, in order for blocks to accurately seem responsive. Not sure if T7 (or SF5/I2) function in this way, but it's something to keep in mind - I think it's likely that T7 uses this method because even online I do not have many issues with feeling like I blocked something on my side but the game did not register. On June 22 2017 04:03 chocorush wrote: Sure there are casuals that don't know there is lag. You can't just project that on the FGC as a whole, because that is patently false. I'm not really saying the FGC as a whole. FGC was extremely niche (especially in the days you mention). Nowdays fighting games are having a resurgence and becoming more mainstream (regularly on first page of twitch). Out of all the players of SF5, Tekken 7, Injustice 2, GG, and so on. What % do you think actually have lagless setups? Even most players I know who attend tournaments have not invested in their setup. | ||
chocorush
694 Posts
If you have established a baseline, 3 frames of lag is when you start noticing it, and feels noticeably worse than 3 frames less. If you play on PC with Vsync off vs Ps4, you definitely feel the difference. If you're only playing offline, the game is objectively worse. Do you play a lot of tekken? Try reacting to things like Xiaoyu rain dance mixups where you have to attack as soon as you blocked to not get bs'd by her evasion. It has always been hard and is much harder with 3-4 frames of lag over the arcade version. | ||
Foxxan
Sweden3427 Posts
TTT2 was 101ms of lag. People were not complaining about input lag. T7 is 27ms more. Now all of a sudden, "128ms is a big deal". Stop comparing it to other tekken games. Its irrelevant. And your argument is flawed. If you see "complaining" as proof people are aware then you are wrong. Just because no one talks about it doesnt mean they dont FEEL IT SOMEHOW. They maybe cant explain it, maybe they play online and thinks its the network or just how the game is supposed to be played. I wont respond to this anymore, its pointless and you keep giving me flawed arguments and baseless crap such as "no right ot complain" TRUTH OF THE MATTER IS 100 MS LAG IS A BIG FUCKING DEAL, DONT TRY AND SPIN IT AROUND | ||
Spyridon
United States997 Posts
On June 22 2017 13:11 chocorush wrote: You need some input lag to allow for a buffer on commands so things like multiple button presses don't require perfect precision. This has traditionally been around 4-5 frames of lag, but tekken had chipped away at this for console releases because of netcode. If you have established a baseline, 3 frames of lag is when you start noticing it, and feels noticeably worse than 3 frames less. If you play on PC with Vsync off vs Ps4, you definitely feel the difference. If you're only playing offline, the game is objectively worse. Do you play a lot of tekken? Try reacting to things like Xiaoyu rain dance mixups where you have to attack as soon as you blocked to not get bs'd by her evasion. It has always been hard and is much harder with 3-4 frames of lag over the arcade version. Not sure where the 4-5 numbers are coming from, but both T6 and TTT2 are at 6 frames according to the tested data. Where did you see any data that suggests the PS4 version has 3 frames of lag difference? All the tests I've seen have a difference under 2 frames (1.7 IIRC). Although to be fair, PC version does not have solid data yet. These numbers seem to fit under the "3 frames of lag before you start noticing it" though...? On June 22 2017 20:45 Foxxan wrote:TRUTH OF THE MATTER IS 100 MS LAG IS A BIG FUCKING DEAL, DONT TRY AND SPIN IT AROUND All I have is 2 points for this: 1) About 40% of fighting games are ABOVE 100ms lag. 2) I'm sorry but, if your saying things like "100 ms lag is a big fucking deal" how am I not supposed to bring up that both T6 and TTT2 were both OVER 100ms lag....? That's not any spinning at all, it's the truth. Am I supposed to ignore the inconsistency in your argument if over 100ms was acceptable then but is suddenly not now? | ||
Fecalfeast
Canada11355 Posts
My steam id is 76561197979958800 or _pwnt_ I'm usually on in the evenings pacific time after 01:00 GMT (+00:00) thanks time tags! | ||
chocorush
694 Posts
We know the PC version with VSync on has about 6 frames of lag, at least one less frame than the PS4 version. Disabling V-Sync means inherently removing about 2-3 frames of lag depending on the methods they used, so 3-4 frames of lag is a conservative estimate of the difference between PC Vsync disabled and PS4. It's pretty obvious at this point that the PS4 version is playable. But the game is obviously better with less input lag. Objectively better, unless you're playing online. Why can't they have different latency when playing on netplay? PC games have done this since the 90s, and it is perfectly acceptable to expect the same standards with today's better hardware. | ||
Spyridon
United States997 Posts
On June 23 2017 05:44 chocorush wrote: Why can't they have different latency when playing on netplay? PC games have done this since the 90s, and it is perfectly acceptable to expect the same standards with today's better hardware. In these days rollback netcode is this used. Old-style PC netcode was input delay netcode. This means if the game lagged, your inputs would lag behind to, which made combos extremely hard to time correctly. Rollback netcode repairs this issue, but it could make the game very jerky with the auto-corrections. So they usually add a few frames of latency in order to make the auto corrects more subtle (or indistinguishable, under good network latency). Check any GGPO enabled game (which is typically thought of as the best fighting game netcode) and you'll see options for how many frames to delay. The reason they use the added latency on single player as well, is so the game does not feel completely different when playing offline compared to online. Although some (like Guilty Gear) have netcode that dynamically changes the delay as your playing to match network conditions. The only option that would be better than this, is to have official servers to host the game servers (rather than direct peer 2 peer). Problem is, this extremely expensive for a fighting game, and it could potentially cause even worse issues (if you lived far from the official server, even if you played a friend who lived down the block from you, you 2 would be laggy). But it should be noted, that it's a bit more complex than the above explanations. For example, I briefly mentioned above that the inputs are "selectively" delayed. If you hit a punch, that will be delayed. But blocking is typically NOT delayed (or at least not to the same extent). This allows players to "react" faster. (this is also why basically all input lag testing is incomplete - it is not factoring differences on which type of inputs). For the record, I am not sure if Tekken uses this method or not. But it is common in rollback style netcode. Credentials for talking about this type of thing: I created rollback style netcode for an unreleased 3rd person indie game a few years back. It was only used in tech demos, but had full server side projectiles & physics. It looks like a drastically different game (and it is), but the game still uses startup, active, recovery frames, block stun, and hit stun. Mechanically, it functions very similar to a fighting game. We did a lot of research on how input lag affects gameplay, and analyzed spreadsheets of the numbers in fighting game netcode to better undersatnd how they function. This is why I feel so strongly about common misconceptions about input lag. Link to tech demo: | ||
chocorush
694 Posts
| ||
Velr
Switzerland10414 Posts
Last week I finally started… Became Grandmaster with Asuka after ~25 games but now i am kinda stuck. I really suck at juggling-combos but from what i saw on youtube Asuka is luckily not really about that anyway . All these Chars that weren’t in 3 (or at least 4, but i played way less 4) often just run all over me because i have no clue what they can do :D - fuck you Dragunov, Claudio and „Girls that i can’t even identify by their profile pictures“. In general i can shit all over „Expert“ and lower ranks but greens beat the crap out of me the same way. Also, for some reason, i just get trashed by just about every Heihachi on the Planet :D. Fun fact: First I played a bit with Xyaou until I realised that I still have muscle memory from Tekken 3 and just can’t correct it, playing her whiteout first learning „correct“ juggling just feels pointless… /sad. | ||
chocorush
694 Posts
The general techniques used in combos in this game is the same for each character, so once you figure it out for one, it's not as hard to apply it to others. Just start with the easy ones, as the advanced combos tend to only do about 4 or 5 more damage (still good to learn eventually because they usually have more wall carry). | ||
Velr
Switzerland10414 Posts
(I hope my Input writing is correct, never used it or learnt how to read it until... Yesterday :p). After a launcher (atm mostly f/d+2 and d3+4) I just go for a "2,1,1+2" or "2,1,4" (iirc) and then unblockable (if they stand up) or cancel it and just kick em if they stay down. In General i have problems getting something more after they fall down the first time (i think its called "Screw"?). All else I try to counter lows with "d3+4". Try to stay at midrange and punish misses/counter with "b3, db4,4,4, 3" Poke with "1,3 or 4" and "b1+4, (3)" Sometimes put a f,f 4, 3+4 or f, f, 3 or f,f,f 3 or 4~3. Main Problem is lacking knowledge about other characters, especially the not really popular ones. Its a pure guessing game how their combos work and when i can jump, duck or parry punish them. Asuka also doesn't seem to play too well on offense? Well.. Need more play Btw: When it comes to lag/delay. Small diffrences can matter. Back in the day a ping of ~100 was deemed okay for online (egoshooters) and felt totally playable, but being over 150ms made everything feel bad. These diffrences matter on how you feel while playing. Its only 20 ms more isn't really an argument because after a certain point you REALLY notice it because controls start feeling sluggish. Its like the fps discussion. Compare 30 fps vs 60+ fps. 30 is totally playable but if your used to more you'll notice it... Now go down to 20 and even a total casual feels that its not playing how it should anymore. | ||
chocorush
694 Posts
Character knowledge is definitely something that comes with just playing more (I find unless you have someone helping you directly, it's almost impossible to figure it out on the fly other than going to practice mode and looking up frame data for each character). In general, you need to watch out for these against pretty much any characters: Hopkick, anything that launches - Jab punisher, 12 frame punisher. Shoulder attack - jab or 12 frame punisher. Low sweep that launches - Crouching launcher Hellsweep, or any low that causes them to stagger - Crouching launcher Fast low hit - 11 frame while standing move (usually WS 4). If they ever whiff while you are doing a backdash or side step, try to do your generic launcher (d/f + 2 in your case). The punishers get a lot more complicated when you go into character specific stuff, but these are universal no matter who you are playing against and which character you are playing. Edit: On top of that, launchers after strings are often launch punishable with d/f +2. Think Xiaoyu rain dance kick, or Law's flip kick strings. Just get in the habit of pushing something after you block these and try launching if it's something that looks very dramatic, and you'll get a feel for what's launch punishable. | ||
Velr
Switzerland10414 Posts
Just to understand you, I punish the laucher after the combo, not between last normal hit and his launcher? If so, yeah, thats how i win games... If not wtf, i'm worse than i imagined . I tend to get kicked in the ass by stuff that "looks" punishable but isn't really (at least not with the stuff i try to use). The revenge option is glorious btw., the amount of games I turned after "getting it"... Too bad it doesn't work twice, going 1-1 just feels bad. And getting your ass kicked twice but less bad the second time also feels strangely rewarding. | ||
chocorush
694 Posts
There are quite a few where you can duck the second hit in the middle and interrupt the third hit (sometimes with a launcher). Many things are not launch punishable (in high level play, people intentionally choose these moves because they are safe(r), while others are just swinging). Most launchers are between 12-14 frames punishable, so they can bait you if you tried to launch punish). As a visual cue, if a launcher has evasive properties it is more likely to be launch punishable. | ||
Spyridon
United States997 Posts
On June 27 2017 00:16 Velr wrote: Btw: When it comes to lag/delay. Small diffrences can matter. Back in the day a ping of ~100 was deemed okay for online (egoshooters) and felt totally playable, but being over 150ms made everything feel bad. These diffrences matter on how you feel while playing. Its only 20 ms more isn't really an argument because after a certain point you REALLY notice it because controls start feeling sluggish. Its like the fps discussion. Compare 30 fps vs 60+ fps. 30 is totally playable but if your used to more you'll notice it... Now go down to 20 and even a total casual feels that its not playing how it should anymore. Just want to point out a couple things based upon the 2 points above, since I've worked on both. 1) Lag/delay. Ping <= 100ms. This was told as a rule, but most people don't understand the rule, or even "why" it was a rule. The little known truth of the matter: Whether you had 10ms ping, or 100 ms ping, it did not make a difference. And if you had over 100 ping you might as well have had 200 ping. This is the reason it went from being playable, to completely unacceptable, very quickly. Doesn't make sense? Well, let me go in to the technical reason of why. Netcode is NOT a "real time" function. It's not 100% continuously sending data non-stop. It's on an interval. This interval has had a standard. Want to take a guess at what the typical interval is? 10 per second. AKA 100ms. This means that every 100ms, the server sends new data out to all connection clients. No faster, no slower. The netcode also has prediction, it predicts the changes between the last interval and current one, and it uses this to calculate where the character "should" be. This is how the game can look completely smooth on your side. If it detects differences between what is sent from the server & the prediction, it does a "correction". It smoothly blends them together to be least noticable as possible. So what happens if you had an amazing 10 ping? Well, you would be sending data to the server more regularly about what your doing. But you still won't see anything that anyone else until you reach 100ms. What happens if your at 150 ping? Well, your slower than this interval. Not only is the server receiving your movements AFTER the earlier interval, but you are also receiving data from the server at a minimum of 200ms rate. What happens at this point, is you see many more "corrections". Now a part that won't make much sense - even most fast paced games do NOT use intervals much lower than 100ms. Although nowdays you occasionally see one that updates 15 per second. But typically never more. For the reason I just stated above - it basically makes a "standard" rate you want your ping to be at, and at a rate higher than that, it becomes obvious if you are having more "corrections". Note, that fighting netcode is a slight bit different, as they "intentionally" add a delay, where the game will receive your input, but hold it back for a certain amount of time. Some inputs have a higher "priority" where they will update the host immediately (such as for blocking), some are held back (such as attacks). Another note - this is completely different from how input delay works. Input delay is received on the next frame, and is not polled. How the game reacts to the input, is a whole different story - and THIS is what you are seeing when people say "X amount of frames input delay". This isn't JUST the hardware latency - some of the input delay is intentionally held back. In other words, the sluggishness mentioned in the comment I quoted isn't really comparable, because it's received on an interval rather than ASAP, and it gets exponentially worse after the interval. A more comparable situation that you are talking about is if you are using a TV that is not on game mode. That's an input delaay that's more similar to games input delay. Although as mentioned, games usually prioritize certain inputs to feel more responsive. Next, on the subject of 30 vs 60 fps. This is a strange subject. To be honest, much of the game development community has been forced to go for 60 fps as a benchmark, because users get upset if you are not 60 fps. But from a design or artistic perspective, it's sub par. Again, this probably makes no sense. But there's a good reason why. And it's related to the same exact reason that hollywood movies are not filmed at 60fps. There's a strange phenomenon with how our eyes work. If an image looks "too perfect" the mind subconsciously starts looking for flaws. 60FPS is great if you are going for a fantasy art style, but if you are going for a realistic art style, the same EXACT artwork has been tested, and the mind sees it as looking more photo realistic at 30fps, rather than 60fps. Remember when Hobbit was the first movie to attempt going past 30fps? Some people thought it looked great, but a major complaint was that it looked "fake". Why did 60fps become the standard? Well... these days it's mostly because of the rivalry between BF3 and CoD. That brought the attention to this mainstream. Activision pushed the belief that 60fps made their engine superior. After this, the newer DMC had a similar situation. Both had a huge backlash for not being 60fps. The belief became mainstream, and that was that. But sad as it sounds, 30fps actually DOES look better artistically. Whether it's better for gameplay... that's a different story. But many of these games would be intentionally going 30fps if it wouldnt cause as much backlash. This, btw, is part of the reason why some console exclusive games actually look better than PC games. Not simply bc their optimized for the hardware, but if their going for a photorealistic look, 30fps actually is interpreted better by the mind | ||
chocorush
694 Posts
And even then, input delay on movement is also terrible in a game like tekken where you need to side step and duck on reaction. Extra delay on attacks is terrible because there are so many strings that you need to interrupt and punish. You've given no reason to explain why this delay makes the core experience better (offline, competitive play), yet continue to bash people for not understanding why the hardships of developers. | ||
| ||