Xrd online input delay

[burningvigor]

I wouldn't say that.... I would say it's meant for region specific netplay.....

Yeah.

 

Some Japanese player came to the SoCal rooms and was giving 20f lag.  Why would he go there....

[Grimsley-San]

I did some testing last night and for some odd reason I get godlike connections to people in New England and Mid Atlantic.  I have no clue how or why but I was in a lobby for Mid-Atlantic for 4-5  hours with 2-3F lag which felt like absolutely no lag or input delay.  I did notice that everyone was PS3 though and when one person who had PS4 played me it read 12F lag and have some small but very noticeable input lag.  I think that cross play might be part of the issue but not sure.

 

I have no clue what changed or why it's like this but now I know what lobbies to play in from now on.  I will also be testing out some more with PS4-PS3 connection just to see how it goes or if anything changed. One major thing I did notice when playing PS4 people is it felt like they had no input delay but I did...meaning they were playing just fine but I wasn't.  However it could have been the character, I mean they had Sol and were hitting combos on me but I could barely move.

 

 

 

Same goes for me...

I go against people in NA, I usually get shit connection, but when I fight someone from another country, the connection is godlike for me.

[Sigma]

Was having 4F (random bits of 5F) delay against a local player on netplay.

Asked them and found that they are using NA Xrd while I have been using JP.

Both are PS3.

 

Are people having issues with NA PS4 > NA PS3?

Expecting my stick/copy in soon, so I'm hoping not...

 

As a side note, I constantly monitor my latency in real-time to many different major hubs. There is no reason I should be getting 5F to anyone from Midwest to the East Coast, and did not before the newer patches.

[saber]

People need to realize not everyone has FIOS or high speed broadband. I think people forget NA has some of the most horrid internet in the world compared to other countries like HK,Korea and Japan. Hell a lot of EU has better net.Therefore we see more people with lower quality internet. They sometimes play and cause issues. That is their right. There are some people that play using wifi(these people are pure evil). Then there are causals that play with torrents or other things going on. That also can cause issues.

I for one think that the netcode is pretty damn good considering these factors. I also last night played someone with HK version and the are on the WC and I am pretty much MW. We both only had around 2-4 frames. I at this point can say that people who are having issues are having them for the reasons.Which I have stated above and not the netcode itself.

[Drake Aldan]

Can bars be a reasonable way of filtering connections, or should you just assume it's a lie and try everything out?

 

At this time of day I was looking at player/ranked match search... A lot of 0 bars, but a lot of 4 bars, too. Judging from the names, these were the Japanese... The thing is, I don't live in California, I live in Florida.

[Labrys31]

I really don't understand this netcode. Solid 2frame still feels like shit, compared to 3-5Frame that bounces frames. I load up games to find my opponent isn't moving at all and forced to go back to lobbies. 

[Sigma]

People need to realize not everyone has FIOS or high speed broadband. I think people forget NA has some of the most horrid internet in the world compared to other countries like HK,Korea and Japan. Hell a lot of EU has better net.Therefore we see more people with lower quality internet. They sometimes play and cause issues. That is their right. There are some people that play using wifi(these people are pure evil). Then there are causals that play with torrents or other things going on. That also can cause issues.

I for one think that the netcode is pretty damn good considering these factors. I also last night played someone with HK version and the are on the WC and I am pretty much MW. We both only had around 2-4 frames. I at this point can say that people who are having issues are having them for the reasons.Which I have stated above and not the netcode itself.

 

That's the thing though.

The 'poor' connections people have are poor because they lack bandwidth.

Games generally need miniscule amounts of bandwidth, especially fighters, and there isn't a connection today that wouldn't support it.

 

What matters is latency, and that should not be an issue in this country at all.

In fact, many data center hops have already been very close to the theoretical minimum ping for years now.

I would guess that the people I was connecting to had a latency of no more than 15ms based on their location.

Some of them are people I 100% know I have low ping to after doing direct tests.

 

The US internet infrastructure is not the culprit. The only case this will matter is from East Coast to West Coast, and even that should never be more than 4F.

 

The real issue seems to be the netcode in the JP version not syncing up with the NA version right now.

Aksys said that while they have deployed the new netcode patch for NA versions, ASW has not yet implemented the same patch for JP.

So there's the issue right there. We have two slightly different netcodes trying to communicate with one another, and even JP>JP version might be running into issues.

[GodPress]

Can bars be a reasonable way of filtering connections, or should you just assume it's a lie and try everything out?

 

At this time of day I was looking at player/ranked match search... A lot of 0 bars, but a lot of 4 bars, too. Judging from the names, these were the Japanese... The thing is, I don't live in California, I live in Florida.

The bars are currently a horrible way to filter connections.

[GaijinMo]

Both the bar estimate and the delay number are rarely accurate. I get a 4-bar estimate to people in Japan and South America in player/ranked even though I know that's nonsense. After the latest patch, the match will often show 1F delay at the versus screen and then shoot up to 12-19F once gameplay starts. And even when it's steady, sometimes 5F feels better than 2F, etc.

 

In my opinion, people should take better advantage of the region/district system and make more lobbies since it doesn't rely on the connection indicator. Lobbies take most of the guesswork away because you can't argue with geography. It's not like the districts are running out of space, either. Maybe the better players are already using lobbies, but they sure as hell looked like ghost towns anytime I've been on.

[GSD-SPL]

What's the natural frame delay for XRD offline? I'm mostly wondering because someone said in another topic that XRD has 2 more frames of delay than AC, but didn't specify on the numbers for each.

 

Different games and all, but Training Mode Delay 2 feels noticeably worse than playing KoF XIII Steam with my buddy in the next state over, which doesn't seem right to me at all. Online feels all over the place regardless of it saying 2 frames.

[Dime_x]

Was waiting till i got some online matches when my connect was at a decent speed (12 mb/sec) to really judge the game.

Played a guy about 2k miles away, a distance that i play people in skullgirls in all the time and have good connections with despite 90-100ms.

Well, this guy and i went at it in around 7 frames of variable delay. Its completely unplayable and feels like more than 7 frames of lag, and inputs only come out about half the time. Wanted to see whats up and went to training mode and played at 7 frames delay... Pretty bad, but somewhat playable because of no dropped inputs.

Once i get some locals to play against i will test again with people in my city, but so far xrd gets a VERY miserable F for online play. Its about as playable as old cvs2 online or laggy sf4. This is probably only "good" netcode to people that are within 500 miles of each other and have over 30 mb/sec imho... In other words... Bad.

[Gramas]

thanks to the patch the netplay works much better

 

I played with people and had 3-4 frames. Now I play with the same people and im having 1-2

[Sigma]

What's the natural frame delay for XRD offline? I'm mostly wondering because someone said in another topic that XRD has 2 more frames of delay than AC, but didn't specify on the numbers for each.

 

Different games and all, but Training Mode Delay 2 feels noticeably worse than playing KoF XIII Steam with my buddy in the next state over, which doesn't seem right to me at all. Online feels all over the place regardless of it saying 2 frames.

 

2 more frames than what version of AC? It really doesn't feel like that, and until I see any actual tests, I wouldn't trust anything anyone says.

The typical input lag of ASW games (and most games) is 2F with vertical sync enabled. This is the case with Naomi GG games and BB on the Taito Type X2. It appears to be exactly the same on the Ringedge 2 as well.

I'll assume that's what you mean by the 'natural' delay.

 

360 ports of arcade games seem to not run vertical sync, and have lower but inconsistent frame delay compared to the native arcade versions. If the comparison is to the 360 version of AC, it's invalid.

 

 

 

Also clearing up the actual cause of the higher frame delay in matches:

 

[Praise Be]NerdJosh @NerdJosh

Yo @aksysgames is this for US version only? I dont see it?? If it fixes it enough I'll buy the game a 5th time tis the season.

 

Aksys Games @aksysgames

@NerdJosh The patch is for the North American version for now. The Japanese patch should be coming soon, but not sure when.

11:43 AM - 24 Dec 2014

 

 

 

Chibiaya @AyaimmortalVF

@aksysgames the Japanese game is still as laggy as ever. Looks like no online for me...

 

Aksys Games @aksysgames

@AyaimmortalVF The patch is for NA for now. The J-version should be coming...soon. That's up to ASW. not us.

 

12:51 PM - 24 Dec 2014

 

JP and NA versions are using different netplay patch versions.

[Epic]

JP was patched after Christmas. So unless NA got another new patch in-between, then the versions should be exactly the same.

[Sigma]

JP was patched after Christmas. So unless NA got another new patch in-between, then the versions should be exactly the same.

Just got the patch this morning, though I didn't check yesterday.

Seems to have fixed all the delay issues I was having. Getting 1-3F typical after 1.03.

Hopefully that's the end of the issues.

[Soviet Bear]

thanks to the patch the netplay works much better

 

I played with people and had 3-4 frames. Now I play with the same people and im having 1-2

 

This has been my experience as well. Works great.

[Teyah]

2 more frames than what version of AC? It really doesn't feel like that, and until I see any actual tests, I wouldn't trust anything anyone says.

The typical input lag of ASW games (and most games) is 2F with vertical sync enabled. This is the case with Naomi GG games and BB on the Taito Type X2. It appears to be exactly the same on the Ringedge 2 as well.

I'll assume that's what you mean by the 'natural' delay.

 

360 ports of arcade games seem to not run vertical sync, and have lower but inconsistent frame delay compared to the native arcade versions. If the comparison is to the 360 version of AC, it's invalid.

 

 

 

Also clearing up the actual cause of the higher frame delay in matches:

 

 

 

 

JP and NA versions are using different netplay patch versions.

 

+R on Ringedge 2 is most likely 4 frames native delay, given how people complained about additional input lag and how Arcsys slowed down reactable moves (dusts and other overheads) by 2F universally to compensate for the increase from delay 2.  4F input delay is common to the great majority of PS3/360/TaitoX2 fighters like the SF4 series, BB series, KoF13, etc, so this makes the most sense.

 

AC versions have been tested here: http://www.dustloop.com/forums/index.php?/topic/6004-ggxxac-input-lag-testing-results/

 

Xrd delay is unknown at this point.  As soon as my PS4 copy gets in, I'm running tests on PS3/PS4 versions of Xrd using arcade sticks.  I did a crude test using a dual shock 4 pad which showed 4F delay: 

[Sigma]

What kind of display is that?

Is the 'half-screen' transition between frames due to the refresh of the display or the camera's sensor?

 

Have you considered using a high-precision timer next to the display along with a higher framerate capture as opposed to using the camera frames as reference?

Wouldn't the PC #R and Naomi AC versions be technically lagless if the input can only be registered on the next frame and the input display shows it registering at frame +1?

 

Not disagreeing with your findings, just curious what methods you used.

[Killnetic]

Don't forget people have TFT panels that have plenty of frame delay as is. Unless you bothered to research your panel before you bought it, big chance you have a shit one. http://www.tftcentral.co.uk/

 

There is absolutely no reason to post a video showing your perceived frame delay without including your panel's make/model/sn

[Teyah]

The TV details are posted in my original comment in the Xrd thread as well as the linked thread above, and it's pretty obvious to see it's a CRT from the type of screen refresh the pictures show. I'm not getting into a technical discussion here, I'll save that for when I get real results out

Yes I could use a higher fps camera; but I'm not buying one specifically to perform one test where a 60 fps camera and a large number of trials + averaging will suffice.

The input in my pictures occurs on 0, so for it to be lagless the full output would have to be on +1. I suppose you can consider PC #R and Naomi XX series to be lagless in a sense, but since there's always some time where the camera will be out of sync with the CRT, I've used the result where the full image is displayed instead (+2).

[Sigma]

The higher framerate capture would only really be a suggestion if the camera you already had happened to have a HFR capture option, just due to the difficulty syncing capture with a CRT as you mentioned.

 

Nice to have someone actually investigating the frame delay.

I'm interested to see the Xrd results; from a stick delay perspective as well.

[Dusk Thanatos]

Teyah has always been pretty on top of this.

iirc, he's the one who quantified the differences between PS3 and 360 +R for us.

[Drekko]

So after a few days of trying the new Patch 1.03 for the JP version of the game.... my verdict.... it's decent........

 

The reason why I'm saying this even though the netplay has improved greatly, matches against NA version players are still giving me roughly about 5-8f while JP version players are 2-4f.  

 

I have run into a new issue now with the implementation of 1.03, but I don't know if it's just me or everyone else is also getting it.  I'm getting moments where my character will just stand there for a half a second, or it wont read inputs for moves and just does a regular attack.  Originally I thought it was execution error on my part so I went and polished my execution and timing, even after that I'm still having the problem.  I still thought it was execution error, but after having some matches offline with no issues the other day, I'm scratching my head.

[Justice7541]

I've actually been playing a guy on US West Coast from EC and only been getting 4f delay. The netcode is actually amazing, just seems lke some people are having issues with specific things. And yeah, he's JP and I'm US.

[MISTER_BUSH1D0]

Is bbcp dead online or should i go straight to ggxrd