Scalability of ND's Uncharted engine *spawn

I am not a tech-head, but as far as I know, games on the PS2/GC/Xbox, heck, even the PS1/N64 could have some expansive Vistas in the background, that looked good...They just weren't interactive, exactly as the case is with Uncharted 2

Even though I am not a tech-head, I am fairly sure that making an open-world game with Uncharted 2's engine, it wouldn't end up looking better than Cryengine 3 and ID Tech 5(running at 30 FPS in this case)
 
On the interactive point. If you can't be there how do you know whether you could interact with them or not? I understand the backdrop such as mountains etc, but what of the seemingly 3D objects further away?
 
Well, if you can't move it, shoot it or have it change in any way then it's not interactive. You said it yourself.. you can't be there.
 
Well, if you can't move it, shoot it or have it change in any way then it's not interactive. You said it yourself.. you can't be there.

Exactly, but in open world games its the same. You can't be there so would you say those far away places are also not interactive?
 
Exactly, but in open world games its the same. You can't be there so would you say those far away places are also not interactive?

Actually, I believe that in some "open-world" games, you may have geometry that's not intended to be accessible by the developers, but players have managed to get there. There's a term for it: "glitching"

Of course, you can do that in "close-world" games... but just to a lesser extent. Obviously.

If you can find a path/way around whatever invisible walls, etc. blocking a normal users progress, you may still be able to go to an otherwise inaccessible area. Of course, what happens when you get there is probably not very interesting since the game code most likely won't handle those cases too well, as it's a bug.

Also, "interaction" is a loaded word... you may be able interact with geometry, even without having a "character" colliding with that geometry. For example: lighting, shadows, etc are all examples of other *things* interacting with that geometry.
 
I appreciate Laa-Yosh's perspective, but I think he's perhaps not appreciating the generality of ND's engine. In the example atop the building in Nepal, where the player can see the entire city but can only go down one cable to another building, what would happen if there were multiple cables, to multiple destinations? With just one cable, the streaming can be easily hinted to preload the only possible destination. With multiple destinations, the streaming hints are very short notice (e.g. stream according to which cable the player is closest to). Will the streaming be fast enough to load in the higher levels of detail before the player arrives as the new destination? If it isn't, can we live with the detail pop-in, or will we want to spread out our detail budget between "current location" and "possible destinations" to help avoid pop-in? Whatever the choice, it's an existing dial in the engine/tool chain, not some re-architecting of the technology.

Personally, I would choose to live with the detail pop-in, and crank detail to the maximum. Detail pop-in is transient, whereas maximizing the level of world detail that the player appreciates as he's exploring is much more impactful.
 
I appreciate Laa-Yosh's perspective, but I think he's perhaps not appreciating the generality of ND's engine. In the example atop the building in Nepal, where the player can see the entire city but can only go down one cable to another building, what would happen if there were multiple cables, to multiple destinations?
But the argument is that there's only one cable for a very good reason - to enable streaming! Okay, it enables the pace and storyline too, which is more the point, but when you know the player has to follow a linear story arc with linear locales and dramatic set-pieces, you design your engine to suit (we hope!) and, knowing how much you'll be showing at a given time, created assets to max out the available space.

From the Nepal tower, if the player could go anywhere, assets would ahve to be paired back to fit more in at once, or stream in faster. Although with plenty of repetition, you may not have to give up too much. That sets me awondering, could the U2 engine cope with such a game, just in reduced fidelity, or would it need some reworking to change the lookahead and fetch code to be more adaptable?
 
Good question. The next would how adaptable would the engine's AI be to a sandbox world.
 
I appreciate Laa-Yosh's perspective, but I think he's perhaps not appreciating the generality of ND's engine. In the example atop the building in Nepal, where the player can see the entire city but can only go down one cable to another building, what would happen if there were multiple cables, to multiple destinations? With just one cable, the streaming can be easily hinted to preload the only possible destination. With multiple destinations, the streaming hints are very short notice (e.g. stream according to which cable the player is closest to). Will the streaming be fast enough to load in the higher levels of detail before the player arrives as the new destination? If it isn't, can we live with the detail pop-in, or will we want to spread out our detail budget between "current location" and "possible destinations" to help avoid pop-in? Whatever the choice, it's an existing dial in the engine/tool chain, not some re-architecting of the technology.

That one is pretty easy to answer and ties in with my former posts (and others too): You don't put multiple cables but perhaps a path where you can climb down to different locations, again, limiting the pace of the game by the speed at which you work yourself to different locations. You use cables and quicker 'transportation methods' where you can and know that the engine can handle it.

If you have different paths to different areas would already come off as open-world enough; e.g. being able to move to anywhere within the city where you wish. At this point, it's not much different than AC and just as open-world at the same fidelty (assuming again, limitless budget). If you want more freedom in game-mechanics, then sure, you'd have to start making trade-offs. That doesn't say it couldn't handle an open-world setting however.


BTW: an older quote but quite relevant:

Laa-Yosh said:
I'd also like to add that if Naughty Dog does not take advantage of the player's limited freedom, then they're very far from making the most of the system's resources. Considering their technical abilities I have to assume that they squeeze everything out of the PS3 that they can, by making compromises that other types of games can not do.

That's all relative. I bet Naughty Dog could have made the game even better looking if they had chosen to use cinematic sequences to preload next assets - which for the most part they did not, since all the pre-rendered cut-scenes are skippable. So if they're not streaming while playing pre-rendered cut-scenes, they're doing it before during gameplay, hinting that the engine is in fact a lot more versatile than one would suggest. Sure, smokes and mirrors is a factor... just like in every other game as well too, open world or not.


You really can't back track very far in Uncharted 2. Almost very larger open area you have to drop into from a ledge so you can no longer get back. Or for the same purpose, a zip line, a door that closes behind you, some scenery gets blown up behind you.

There are countless little tricks in Uncharted 2 that control how far players can get back, go forward, and when they can go forward. Always allowing Naughty Dog to know where the player is, without the player ever knowing it within the gameplay.

That's true, although I'm not sure how that invalidates what I posted in reply to Laa-Yosh. Laa-Yoshs point was that all the areas where you have to wait for your AI companion to drop down the ladder is where the engine streams the next part of the game, hinting that these little sequences are necessary.

Even while you're waiting for the ladder to come down, at this point, you can still travel back into the battle field. I actually tried doing this a few times, when I decided to go back looking for more ammunition. I'm actually pretty certain a good place to test this is the battle-area at the beginning of the chapter "Desperate Times". Sure, there might be areas that are shield off eventually, but that doesn't necessarely negate the option to have a more open-world setting if the areas that do fit into memory are large enough and the pace of the player slow enough to stream in new areas fast enough.
 
The game does not immediately unload the area that you're in while waiting for the next to load.
It probably unloads the previous area that you've already left many minutes ago.

I've already suggested that they're probably only streaming in a few dozen MBs of data at the same time, and that only takes about ten seconds if they cache it on the HDD. In-engine cinematic sequences are about that long, so it's obviously not a coincidence.
Prerendered cinematics at HD res usually require some bandwith to read the video frames and the audio - which is why they are not streaming any data at those times.

An example of loading when backtracking would be the Halo series.

Also, the whole open world discussion is b***shit because you guys keep bending the meaning of the term.
 
The Desperate Times level actually proves the theory very well. You drop from a ledge into that area so you won't be able to get back. And you can't leave that first open area without help from an NPC. And it isn't necessarily that little ladder sequence where they stream in the next part of the level, but rather I think it happens during the whole fight since you won't be able to leave that area during the fight. After that you quickly reach an elevator which gets stuck, once again blocking you way back.

Having more time to stream in the next part of the level won't allow them to make the graphics better beyond a certain point. That is as soon as they fill up all the RAM available. And if they create multiple paths through a level that are large enough to required separate parts to be streamed in, the area that allow you to choose those paths will be limited by having to fit the area your currently in, all the next paths you can choose, and if you can get back the previous area you came through, all within the available RAM. They can probably do it if they keep those areas small enough, or anything that doesn't take up a lot of memory, but it might not feel like an open world if they don't present that choice frequently. And every time they do it will require smaller areas.
 
cornsnake said:
The Desperate Times level actually proves the theory very well. You drop from a ledge into that area so you won't be able to get back. And you can't leave that first open area without help from an NPC. And it isn't necessarily that little ladder sequence where they stream in the next part of the level, but rather I think it happens during the whole fight since you won't be able to leave that area during the fight. After that you quickly reach an elevator which gets stuck, once again blocking you way back.

I know it does, but the area is also huge and the point was more to show that the "ladder sequence" didn't do much in this area besides holding the player back until the battle has been won. I agree on the loading bit and would go further to say that the loading is done while inside the building and being pre-ocupied with climbing up around it.



Laa-Yosh said:
Also, the whole open world discussion is b***shit because you guys keep bending the meaning of the term.

I find this remark a tad bit ironic and annoying given that I asked more than one time for a definition of what kind of open-world games we're talking about. Suffice to say, you never really gave one, so how are we bending it?

The point at which you entered the discussion was the argument used, that the games linearity gives them the opportunity to deliver the imperssive graphics that they are, implying that if they didn't stick with the linear gameplay, the graphics would surely take a hit. I agree, that linearity does give benefits that can be exploited one way or the other - and that a team like ND would be silly if they hadn't exploited it in any way. No argument there, really.

On the other hand though, I don't see it as guaranteed proof either that ND used this solely on visuals or that this somehow negates the engines ability to offer a non-linear gameplay experience at these visuals. We, who have played the game, have offered views and named specific parts in the game that show that despite all the smoke and mirrors (which is present in all games), that there is a lot of streaming going on that isn't necessarely confined to intelligently placed sequences. If anything, it's more down to excellent level-design that makes the progress of the game predictable.

Now maybe if accoarding to your definition of open-world-games, there must be helicopters, cars and the option to be anywhere at any given time in a split of a second, then sure, you're absolutely right that the graphics would take a hit. No one ever really argued this though.


If the game offered Tibet as an open city to explore with various routes to chose from at the same time, I bet you could keep the same visuals as long as you confined the gameplay within the bounderies of the engine. This includes putting in some "climbing elements" where you'd need to slow down the player to preload the approaching area etc. Smokes and mirrors - but what is so new about that, as long as the game offers more than a fixed linear path?
 
I know it does, but the area is also huge and the point was more to show that the "ladder sequence" didn't do much in this area besides holding the player back until the battle has been won. I agree on the loading bit and would go further to say that the loading is done while inside the building and being pre-ocupied with climbing up around it.

I suppose all that's left now is to define "huge".:p
 
While I do find all this talk of an "open world" Uncharted game using the U2 engine rather riveting and informative, having had my mind blown out the back of my head by playing U2 this weekend, I'd be very uninterested in anything other than the very story-driven formula with awesome set pieces that we've all come to know and love.

If ND did decide to make U3 more "open-world", i highly doubt it would detract too significantly from the fomula followed in the previous game (why change what is already AMAZING?!?)...

In terms of game design, It's very likely in my opinion that ND wouldn't want to make U3 as "open" as say a GTA game, as it certainly wouldn't feel like an "Uncharted" game any longer.

I agree with the sentiment that a more possible suggestion for an "open-world" Uncharted game would simply consist of a very tightly controlled semi-linear course offering the player multiple paths towards a common "end".

Regardless, I think the current debate is dealing more with the ability of ND to produce any "open-world" game (not necessarily an Uncharted game) with the current version of their U2 engine.

Having seen for myself those vistas... those beautiful beautiful vistas... I think i'm a believer that with clever level design, a possible slowing of player mobility, and further relevant tweaks to the streaming engine and LOD systems, I'm entirely convinced that ND could do it!
 
I find this remark a tad bit ironic and annoying given that I asked more than one time for a definition of what kind of open-world games we're talking about. Suffice to say, you never really gave one, so how are we bending it?

Open world still means that you're - theoretically - completely free to traverse the enviroment as you want to, a good demonstration is the ability to use flying vehicles in GTA games, AC2, and some Halo levels.
The fact that you may not be able to use such a vehicle, or can fly only through some cheat, does not change the underlying technological differences between a linear and a completely open game world.

If a game only offers you to chose from three fixed paths, it is not open world yet, it's still linear, but with multiple branches. Imagine a tree or a graph structure describing the options for traversal of the world - the options are still limited.

UC2's engine will not be able to implement even this severly restricted approach just because some people here say so. That's something only the ND devs can tell for sure - all we can do is make up theories and present some explanations to back it up. This reasoning is what's completely missing on one side of the argument.
 
Having seen for myself those vistas... those beautiful beautiful vistas... I think i'm a believer that with clever level design, a possible slowing of player mobility, and further relevant tweaks to the streaming engine and LOD systems, I'm entirely convinced that ND could do it!

Sure, but like I probably mentioned before in various threads, I'm hoping they'll do something like the first Jak & Daxter on PS3 instead. I loved that.
 
Open world still means that you're - theoretically - completely free to traverse the enviroment as you want to, a good demonstration is the ability to use flying vehicles in GTA games, AC2, and some Halo levels.
The fact that you may not be able to use such a vehicle, or can fly only through some cheat, does not change the underlying technological differences between a linear and a completely open game world.

If a game only offers you to chose from three fixed paths, it is not open world yet, it's still linear, but with multiple branches. Imagine a tree or a graph structure describing the options for traversal of the world - the options are still limited.

At the end of the day, it's all smoke and mirrors though - as you yourself pointed out. If the streaming is already present, the more branches you have, the more open-world it becomes. Where do you draw the line? If you take the city of Tibet for instance: If you take away the rubble and open most alleys, it would already give enough freedom to pass off as an open-world game, even if some platforming is still necessary to reach some areas.

Perhaps to turn the argument around - if Rockstar decided to make a GTA game with an ancient Rome as its city during the roman empire where there were no helicopters or aeroplanes and the pace of the game was limited by how fast the player could move around (e.g. much slower than in a modern setting), I'm sure the engine would be capable of offering more variety.

If ND choses to limit their game to the pace of walking speed by foot while offering a non-linear game, surely that's no less an open-world game by its definition then. Of course not with the same amount of freedom, but as I said, no one ever argued they could pull of GTA4 with these graphics nor was that ever the discussion...


UC2's engine will not be able to implement even this severly restricted approach just because some people here say so. That's something only the ND devs can tell for sure - all we can do is make up theories and present some explanations to back it up. This reasoning is what's completely missing on one side of the argument.

...

I agree that Naughty Dog is the only one that can give a clear answer to the limits and bounderies of their engine, though that surely also includes your remark about what the engine apparently can't do. ;)

We're all just making reasonable guesses and sharing ideas and thoughts, and perhaps trying to point out some things that you may have missed, given you evidently have not played the game for a reasonable period and are judging a lot on videos posted - which is why I took the liberty to point out some things that at the very least raises a few questions here and there as well as hinting to some of the possibilities and the potential of the engine.
 
That's exactly what I was talking about when I said that you keep bending the meaning of the term 'open world'.

If this, and that, and that too - then it's an Open World Game! Why not call Super Mario Bros or even Space Invaders one as well?
 
Back
Top