Jump to content
Dustloop Forums

Recommended Posts

  • Replies 15.4k
  • Created
  • Last Reply

Top Posters In This Topic

Posted

Plenty of netcode works just fine on US infrastructure, and plenty of netcode doesn't shit itself over packet loss or delayed packets. Yes, physics says that there's a limit to how well you can do without compensation methods like rollback, but physics has -nothing- to do with whether your code is fixed delay or variable delay, how your code handles packet drops (spoiler: you can do it without the game shitting itself and dropping you to delay 10), how your code handles potential desyncs, or how your code handles spectating.

Literal doujin netplay code (see: CCCaster, IaMP RollCaster) does better than ArcSys code on basically all of these counts. There's no real excuse.

Posted

The best we could do is have 100% fibre optic connections around the world and I still think there would be some delay. 

 

For lagless online you need to find something in the universe faster than light that you can manipulate. Then you will be the greatest scientist/engineer of all time. 

 

Just popped in to reinforce this limitation. 19ms from LA to NYC in fiber optic. (http://www.wolframalpha.com/input/?i=new+york+city+to+los+angeles)

 

Still, more can be done elsewhere. This would just be an upper boundary.

Posted

That's basically just the speed of light within our air. The loss from Total Internal Reflection is negligible. I compared their results with the raw distance calculation and the difference could even be a difference in their displayed number of significant figures.

 

Net code and delivery would obviously determine how close to that number you can reach.

 

Edit: And by "net code and delivery", I am including all avenues between source and reception, not just the PSN netcode.

Posted

For those of you already calling doom and gloom over Xrd netcode, I question if you experienced the fickle bullshit that was AC+R netcode.  Hope you don't like FRCs!  

Posted

Sigh... Why did I hope even for a second than ArcSys would give two shits about the American market.

 

As Dusk points out, anyone who has ever played IaMP knows that netcode is a solved problem.  That ArcSys is still implementing variable delay with no rollback is flat out confirmation that they just didn't care.

Posted

Wtf is this depression syndrome. The game is HOURS away and you wanna complain? Can you at least wait until you have the game in your hands?

Posted

Seriously guys, come on now :T enough with the downer talk. We're not here to predict, especially of something you've never played.

Don't be predicting things you can't know for sure about.

Posted

theres literally 23 hours until the game comes out lol

 

just be patient yall, we'll find everything out soon enough

Posted

Humans can't adjust by afew frames. 1 frame is about 16 ms

 

I-No's 5K is 12 frames of hitstop and her 2D is 13 frames. I can clearly tell the difference between the two, as I have to adjust FRC timings for her horizontal chemical love (the big red beam) by 1 frame since I buffer the input during hitstop and I know they are not the same timing, so yes, you can adjust timings by as low as a single frame.

Posted

If only ASW used the magic that is GGPO.

I'm going to scream if I see one more GGPO begging post for any fighting game ever. GGPO is not a magical one size fits all band-aid, and requires a huge amount of fine tuning and testing for each individual game to get it right. Japanese companies that are not overly Westernized like Capcom will not use it because of the emulation/piracy stigma. Also, if you want to use any kind of rollback netcode outside of emulation, a game's source code has to be designed from the ground up to accomodate it. We've already been told that this is why the KOF UM games on Steam will not use rollbacks.

 

You should really be more happy that the developers have provided the means for us to play these games online at all. To put things into perspective, Chaos Code and Digimon All-Star Rumble are two recent fighting games that lack any netcode whatsoever.

Posted

Sigh... Why did I hope even for a second than ArcSys would give two shits about the American market.

 

As Dusk points out, anyone who has ever played IaMP knows that netcode is a solved problem.  That ArcSys is still implementing variable delay with no rollback is flat out confirmation that they just didn't care.

 

You do understand that the average casual player prefers slowdown/input lag to the teleport hell rollback turns into on bad connections (which America is chock-full of)?

Posted

You do understand that the average casual player prefers slowdown/input lag to the teleport hell rollback turns into on bad connections (which America is chock-full of)?

Lol I remember playing skullgirls on pc and the connection shat on itself for no reason. The game showed me getting a J.lk when I was already dead.
Posted

I-No's 5K is 12 frames of hitstop and her 2D is 13 frames. I can clearly tell the difference between the two, as I have to adjust FRC timings for her horizontal chemical love (the big red beam) by 1 frame since I buffer the input during hitstop and I know they are not the same timing, so yes, you can adjust timings by as low as a single frame.

 

I second this.  World of difference comboing from 5K/2D/5H because of the 12/13/15 frames of hitstop.  Anyway, as long as the netcode is stable, that's all I want.  I can play better in 10 frames of stable delay than in a game with the latency sporadically changing between 2 and 7 frames with random full-second lag spikes every 10 seconds.  Given the slightly increased buffer windows and the reduced need for hitting 1-2 frame cancel windows, the game will probably play better online than +R did.

 

Good thing I am gamesharing lol. There goes my limited edition

 

This is what being responsible enough to not spend money on the same game twice for an extra 2 weeks of fun got me.  Now I wont even get to practice before I go back to work.  :vbang: 

Posted

The version on the HK store is labeled as English/Japanese Ver according to the Playstation blog. Has there been any news on how much of that version is in English? Is it known if they have access to English menus and subtitles?

Posted

The version on the HK store is labeled as English/Japanese Ver according to the Playstation blog. Has there been any news on how much of that version is in English? Is it known if they have access to English menus and subtitles?

From what I've been reading, even the Japanese version is suppose to have English text patched in. It's likely that the HK version will have it. I remember when HK Demon's Souls had full English voicing and everything and gained popularity through importing. I'm sure someone here may be able to help you confirm that. 

Posted

Seriously guys, come on now :T enough with the downer talk. We're not here to predict, especially of something you've never played.

Don't be predicting things you can't know for sure about.

 

If you've watched the video recently linked, it confirmed the netcode was variable delay because of the behavior during the game intros (floating around from 3 delay to 4 then settling at 2 on what's presumably a jp to jp connection, meaning it should be relatively stable at that point). If it wasn't variable delay, it would've been two flat the entire time.

It's downer talk, but it's not speculation. The vid provided pretty concrete proof that it's variable delay, which is about as bad as it gets.

 

Posted

blazblue and p4a worked the same way though, the game would be notoriously slow as it synched or something, then settle on a completely expected connection

Guest
This topic is now closed to further replies.

  • Upcoming Events

    No upcoming events found
×
×
  • Create New...