They didn’t say it wasn’t doable, just not something they’ve wrapped their heads around yet. Nowhere in that interview did they say it doesn’t work with 3D models, it just makes it more complicated.
: ( feels like he is prepping us for the bad news…
Indeed. The need to know about the online play is now more important than ever.
Game your talking about is the Dragonball Z arcade game that drops next month.
Too ass it up a little, their responses are what I have been saying since the beginning of November when it came up again. It’s good to hear the reps say the same thing >>
Maybe you missed the part where they said if you want ggpo then keep supporting it
From what I’ve read, it makes it more complicated for their current implementation. I’m completely assuming here, but it sounds like neither SF4 nor MvC3 were truly designed with online play in mind (SF4 and TvC were arcade games, after all), and, to make the game code “fit” into the GGPO paradigm would require a lot of time and effort, even if it is doable.
They really need to start designing fighters for online play. Period. This is not the fucking early 90s anymore. Arcades are not still alive, and Japan does not still account for half of all video game revenue. This is a Western world in 2011. We play online on poor Internet infrastructure, we number in the tens of millions, and we have gobs of disposable income. Western devs know this, designed their games around this, and have reaped the benefits. Your turn, Capcom.
quite the contrary, the heat should be turned up. companies hate any kind of negative responses about their products. conversely, they love it when people hype their games (for free) for them. players only hurt their own cause for better netcode by letting the fire die out. the fact they feel the need to perform PR damage control is a good sign. besides, keeping this thread open is more efficient than creating a new one once players play MvC3 online.
agreed, regardless of the reasons why they couldn’t do. the fact of the matter is they’ve failed to deliver on a critical aspect of the game. the whole point of voicing your displeasure is to let Capcom know the status quo is unacceptable and to light a fire under their asses to improve. if one goes into a high end restaurant and gets served a cold meal. does one seek to remedy the issue by complaining to the staff or does one say nothing, eat it, pay for it and let em off the hook? the response to Capcom over their failure should be no different than the former, like with any other company that fails with a consumer service issue.
Signed.
If you want them to improve the netcode, vote with actions that they track. You could not buy it firsthand, even if it’s just the first week, and then don’t play online even once. If the ratio of online players to sales is low, they should realize something is up with the netcode. Buying and playing games with good netcode online will also help. Personally, I’m thinking of buying MvC2 on the day MvC3 is released and seeing how badly I can get destroyed online. If you already have MvC2, you can get HDR or Final Fight: Double Impact to make a similar statement.
It’s pretty much pointless to expend energy arguing about this now. It’s over and done. Just have to wait until the game’s out and we can verify firsthand how adequate their efforts turned out on that front. I do not have high hopes though.
I still think their explanation is bunk (Sega-AM2 seems to have a decent handle on putting adequate netcode into their 3D arcade games after the fact; why not Capcom?), but I’m not going to shoot the messenger over it. Seth seems to have been about as helpful as someone in his position can be.
… I can only ever play my friends in the same town online without getting lag :(, and lag in mvc3 could probably lead to HUGE damage unlike ssf4 where nobody has any decent damage output.
GGPO requires two versions of the game to be running. In Marvel, this can be a disaster. 6 characters at one time, I doubt any netcode could handle that.
We already have rollback netcode for MvC2 so we know it works. If not GGPO exactally then something else that is still rollback.
Oh yes, because things that work well in 2D transition to 3D incredibly well without any unexpected hicups whatsoever.
sees SFIVs hitboxes
I really don’t get how the graphics have ANYTHING to do with the net-code. Keep track of X and Y axis for all characters and what they are doing. 3D graphics should not change this at all. It’s not like your X-box is rendering your opponents game off your disc over the internet.
There’s more to rolling back 3d models than just changing the positions, I think. They work entirely differently than sprites.
Calculating a 3D model that uses key frame skeletal animations: (generally the best choice for non-procedural animations for characters with limbs)
Step 1: Determine which keyframes are being applied to the character. This is the mostly the same as determining which frame of 2D animation to display, and usually uses a timer and checks to see whether the timer is between certain values.
Step 2: Interpolate the bone data of each key frame based on how close the timer is to each key frame’s time stamp.
Step 3: Apply the bone data to the mesh. (Techniques for this can differ, with common techniques being articulated meshes, CPU skinning, and GPU skinning. MvC3 likely uses GPU skinning, as it generally balances the load between CPU usage and GPU usage better, and looks better than articulated meshes.)
Step 4: Draw the mesh.
The only thing of the above that the rollback mechanic would have to manage is the animation timer. All other data needs to be calculated on a per frame basis, anyway, except the key frames, which stay the same for the life of the program.
A major problem is the particle effects. Generally particle systems have a heavy random factor to them and it is hard to rollback randomness and have it look at all similar to how it did before. So, they would have to make their particle system completely lack a random factor, find a way to roll back their random number generator so it gives the same data back as before, find a good-looking way to transition to the new rolled-back particle state, or some other programming voodoo. Alternatively, they could just not do rollbacks on particle effects besides completely destroying them, such as a case where a fireball got stuffed during a rollback.
The other issue is more logistical. Generally when a game is made, different parts of it are given a memory budget, like, say, 16MB for sound, 64MB for graphics, 64MB for physics, etc. If they didn’t account for the added memory usage that rollback netcode will require, they’d have to redo their memory budgets late in the project to include it, even if the impact would be relatively minimal. What we need to demand is that companies keep good netcode in mind from the beginning, budget memory for it appropriately, and make sure all the systems play nice with it.
i agree. seth is a helpful guy, but it’s his job to be helpful. if part of his job entail hyping Capcom’s games 24/7 and singing his bosses praises. conversely, part of that job is also taking a bullet for Ono and Niitsuma when they screw up. besides, if he’s such an advocate for good netcode as he claims. it’ll give him ammunition he can take to his bosses to try and convince them the company really screwed the pooch on the netcode and need to improve in future titles (ie: 3S Online and SFxT). OTOH, giving out free passes to will ensure nothing gets done.
the conditions you stated applies to the PC not console games or as Caim stated in this thread: ***“GGPO needs to ultimately run two instances of the game being played because the games are emulated; ponder has no access to the actual source code for any of them.”, “Capcom has the source code. They have the specs of the hardware. They don’t NEED to run two complete instances of the game”***.
That guy doesn’t know what he’s talking about and likes to insert subjective opinions into his baseless argument. I really don’t care what he thinks of the SFIV models, a model being “ugly” doesn’t make it easier to run.