Fallout Tacticts! WOW!

Fallout's graphics did not have hundreds of thousands of polygons per object. Those renderings should be easily rendered in real-time on modern videocards.

Think something like Junktown in a 3D engine. I doubt such huge amounts of polygons could be rendered at a playable framerate at this time.
It can get close, but it will still be blocky.

The last time you played a 3D game was in 1990? Please go play Bioshock and then complain about 'blockiness' and 'loss of style'. Bioshock is the perfect example of a stylish game in 3D, a very Fallout-y style at times.

Ahm, yes, I just completed it today. And yes: blocky. Look at barrels, look at arches... well, look at every curved surface.

Urgh, what? Silent Storm had fantastic graphics, adequate for Fallout to keep it's style. Same for other modern games.

This Silent Storm?

http://img.photobucket.com/albums/v635/sfg/427962.jpg

Don't make me laugh. Van Buren looked better.
 
FeelTheRads said:
Think something like Junktown in a 3D engine. I doubt such huge amounts of polygons could be rendered at a playable framerate at this time.
It can get close, but it will still be blocky.
No it won't. It doesn't need to be, at least.

FeelTheRads said:
Ahm, yes, I just completed it today. And yes: blocky. Look at barrels, look at arches... well, look at every curved surface.
Que? Are you kidding me? What, did you play it at the lowest resolution and lowest detail? 'Cause I saw no blockiness, at all.

FeelTheRads said:
This Silent Storm?

http://img.photobucket.com/albums/v635/sfg/427962.jpg

Don't make me laugh. Van Buren looked better.
Don't be ridiculous.
 
Pre-rendering technology still remains FAR ahead of realtime consumer-level 3D.

If Fallout 3 were made today using pre-rendered sprite animations, those sprites would look photorealistic. You'd have to blink to make sure youre eyes are not deceiving you.

The difference between pre-rendered and realtime 3D will be obvious and pronounced, until realtime 3D manages photorealism that pre-rendered graphics are already capable of.

Meaning... I'd rather have sprites in my isometric and overhead games. Sprites still have their strengths.
 
FeelTheRads said:
This Silent Storm?

http://img.photobucket.com/albums/v635/sfg/427962.jpg

Don't make me laugh. Van Buren looked better.
Heh...
I played Silent Storm, I saw screens and I don't like it's graphics. They look artificial and a floating camera makes things even worse.
I just played Fallout and I can say that I definitely prefer Fallout's graphics to Silent Storm's graphics, despite that Silent Storm is newer.
So, either it's a bad example of 3d graphics or I simply don't like 3d graphics in isometric games.

About camera:
I prefer 2d isometric because the view feels real and immersive - it feels like a decent, stable perspective to me - in opposition to 3d isometric camera that makes me feel like I'm above the scene, floating around the fucking map - complex controls of camera don't help.
Again - I want to play the fucking game, not a simulator of a recon drone.
 
FeelTheRads said:
Right choose a screenshot that's taken at low settings.



Sorrow said:
I played Silent Storm, I saw screens and I don't like it's graphics. They look artificial and a floating camera makes things even worse.
Silent Storm's graphics were never going for realistic, how can they look anything other than artificial? The camera wasn't the best, but that varies from game to game, it's also essential to some forms of games. And complex controls? Give me a break,when was the last time you saw complicated camera controls?

Pre rendered graphics can go for a higher quality at a lower performance requirement, but 3D has a far greater flexibility. A role playing game might not need that flexibility but a tactical combat game does.
 
FOT was a poor tactical combat game, lacking many features that were present in other tactical combat games that predated it.

Things like destructable terrain, vehicles or fixed weapon emplacements offer plenty of tactical options and pitfalls. Options that are more easily achieved in 3d.

JA2 did just fine with prerendered art, but even it had it's limitations.
 
Please don't take this as me being a pain in the ass, but destructable terrain can be done in isometric 2D, and so can vehicles. It's all a matter of how good it looks.

And, in the end, how it looks matters less than how it feels. There's a reason why X-Com 1&2 are STILL being played by people around the world, including my own cousin.
 
shihonage said:
Please don't take this as me being a pain in the ass, but destructable terrain can be done in isometric 2D, and so can vehicles. It's all a matter of how good it looks.
I said it was more easily achieved, not that it couldn't be done. JA2 had destroyable buildings, and FOT of course had vehicles, but there's a limit to what can be done in 2d since everything has to prerendered. That was a reason for no motorcycles in FOT and why all the vehicles are covered, because otherwise you'd have to render the sprites with all the possible combinations of armour, squad make up etc. Where in 3d you just need to marry the models together and animate them.

shihonage said:
There's a reason why X-Com 1&2 are STILL being played by people around the world, including my own cousin.
That's because no one's done anything better since, not because 2d is better.
 
So, is there any possibility of using the flexibility of 3d, without a camera that makes the player feel like he's hanging above the scenery?
 
requiem_for_a_starfury said:
said it was more easily achieved, not that it couldn't be done. JA2 had destroyable buildings, and FOT of course had vehicles, but there's a limit to what can be done in 2d since everything has to prerendered. That was a reason for no motorcycles in FOT and why all the vehicles are covered, because otherwise you'd have to render the sprites with all the possible combinations of armour, squad make up etc. Where in 3d you just need to marry the models together and animate them.

Marrying the models together is not nearly as intuitive as it sounds, and it certainly requires more work. The end result is purely visual, not connected to gameplay.

In addition, Diablo 2 and Crusader series both used "combination sprite" method of rendering where they (especially Diablo 2) created the isometric player image by overlaying different parts on top of one another.

This eliminated the need to render all combinations of everything the way Fallout did. Sprite-based engines have a few tricks up their sleeve :)

shihonage said:
That's because no one's done anything better since, not because 2d is better.

I didn't say that 2D was better than 3D for gameplay purposes. However, to claim that 2D is no longer suitable for quality tactical combat in this day and age is not exactly an accurate assessment, IMO. I wouldn't exclude it as a contender, especially in indie titles of which more and more are springing up.
 
Face it, 2D is no longer suitable for this purpose. The reason being, as said, that each and every single animation frame has to be drawn and saved, needlessly expanding the installation space. 2D is not as flexible as 3D, is not as easy to modify, and overall does not offer as much flexibility and freedom as 3D.

Why should a developer bother with creating a 2D tactical game, which requires a lot more work to get it to behave properly rather than using a flexible 3D engine?

Of course sprites can be prerendered and then saved. But why should pre-rendered 3D models be saved to 2D if the high poly model can be switched to a low poly one whilst retaining the texture? The same technique was used for Doom 3 and Quake IV.
 
Mikael Grizzly said:
Face it, 2D is no longer suitable for this purpose.

I don't get statements like this. If someone wanted to do a 2D isometric strategic combat game right now, they could, with no problems, and it would look excellent.

The reason being, as said, that each and every single animation frame has to be drawn and saved, needlessly expanding the installation space.

In the world of today's hard drive sizes, you can't be serious.

2D is not as flexible as 3D, is not as easy to modify, and overall does not offer as much flexibility and freedom as 3D.

Well I guess if you're into modifying the graphics of every game you play, then 2D really presents a problem. However most people don't do that. They create scenarios, maps, scripting, but few actually dare to seriously mod graphics because that takes a lot of time and effort even for the 3D scenario.

Why should a developer bother with creating a 2D tactical game, which requires a lot more work to get it to behave properly rather than using a flexible 3D engine?

Actually with a 2D engine you would have a lot less graphical troubleshooting to do, as DirectDraw works the same across all cards. There's no such nonsense for DirectDraw as "Bioshock optimized driver" and "Quake Wars optimized driver" that came out from both NVidia and ATI as soon as these games did.

Of course sprites can be prerendered and then saved. But why should pre-rendered 3D models be saved to 2D if the high poly model can be switched to a low poly one whilst retaining the texture? The same technique was used for Doom 3 and Quake IV.

You're referring to normal mapping. Let's say you're doing a strategy game and you have 30 normal-mapped soldiers on screen.

Also the entire environment is polygon-based, normal-mapped, specular highlighted, and applied to all the other fancy tricks to, in realtime. Even with modern hardware, the framerate will inevitably slow down to a crawl.

With 2D, there's no framerate concern. You can have a battle of 200 people all shooting each other at once in an insane screen resolution without any slowdown on today's hardware, and they would all, including the environment, have extremely high pre-rendered polygon counts that would be downright impossible to achieve in realtime.

Mind you, I am not saying that everyone should go out and make 2D tactical combat games. What I am saying is that I don't see the reason to be so readily dismissive of such a concept.
 
Mikael Grizzly said:
Face it, 2D is no longer suitable for this purpose. The reason being, as said, that each and every single animation frame has to be drawn and saved, needlessly expanding the installation space. 2D is not as flexible as 3D, is not as easy to modify, and overall does not offer as much flexibility and freedom as 3D.

Why should a developer bother with creating a 2D tactical game, which requires a lot more work to get it to behave properly rather than using a flexible 3D engine?

Of course sprites can be prerendered and then saved. But why should pre-rendered 3D models be saved to 2D if the high poly model can be switched to a low poly one whilst retaining the texture? The same technique was used for Doom 3 and Quake IV.

You still have to explain me why a player should care about all this?
And installation space? Ahahaha. You did notice that most games now are around 6GB right?

And by the way, I'll be damn if I understand why most of the time. Take Jade Empire for example. The way that game is made is incredibly retarded. So, it has around 4GB of movies. Movies, that are rendered with the game's engine, but saved in a movie format at a very, very low quality. But the game also has animations rendered realtime with the engine which look much better. So, I ask, what the fuck is the point of having those 4GB of movies if they could've just made them directly ingame and thus save a lot of space and have a superior quality?

Of course, this is unrelated to the topic. Just noting again how retarded Bioware is.
 
Mikael Grizzly said:
Face it, 2D is no longer suitable for this purpose.
I think feasible is a better word

Mikael Grizzly said:
The reason being, as said, that each and every single animation frame has to be drawn and saved, needlessly expanding the installation space. 2D is not as flexible as 3D, is not as easy to modify, and overall does not offer as much flexibility and freedom as 3D.
It is not as flexible as 3d in the way that you would have to re-render all the frames if you so much as changed one texture on a character. One thing though is that all game details should already be planned in this stage, so they shouldn't have to change anything. (in the perfect world of course)

Mikael Grizzly said:
Why should a developer bother with creating a 2D tactical game, which requires a lot more work to get it to behave properly rather than using a flexible 3D engine?
I'm not sure what you're aiming at here. "using a flexible 3d engine" sounds like that a developer would buy one pre-made. I'm pretty sure there's flexible 2d engines out there too which behaves properly.

Mikael Grizzly said:
Of course sprites can be prerendered and then saved. But why should pre-rendered 3D models be saved to 2D if the high poly model can be switched to a low poly one whilst retaining the texture? The same technique was used for Doom 3 and Quake IV.
Resource handling maybe? While there are techniques for using very detailed objects in game it gets pretty expensive to have that kind of thing running on every single object in large environments (which is usually the case with birds view). Oblivions environment very pretty amazing in the amount of detail they portrayed, but they did spend a tremendous amount of work on getting it that way while still enabling it to work in real time. With 2d you would use all the bells and whistles for the render, but only have to plan for effective image handling.

add: yeah, basicly what shihonage have already said

So, my conclusion is that if you're going for a 3d engine and want to do the fallout thing, lock the camera to a few angles (if needed at all) and use a lot of prerendered materials for static things together with 3d for movables and destructables.

Ragnarok online is a pretty cool example of a combination of 2d and 3d. They even have a solution for different clothing on 2d sprites. I think they have seperated the different bodyparts into chunks which can be exchanged easily. The camera is locked in a span but you can rotate and tilt freely in this span. Also, the 2d is drawn! Not pre-rendered!

http://img.ragnarokonline.com/game/screenshot/ayo_dun_kid.jpg

Note in the picture how light conditions don't apply that well to the pre-rendered stuff. I don't have any good arguments here. I guess it's the big flaw of it all
 
shihonage said:
Marrying the models together is not nearly as intuitive as it sounds, and it certainly requires more work. The end result is purely visual, not connected to gameplay.
More work for the programmer maybe, but that's not something the player sees. With 2d everything has to be done, even using tricks if it's not been rendered then you can't include it. If a new animation is needed to circumvent a problem in a level it's a lot less work to add a new animation than animate the model and render the frames then add them to the sprite. 3d does offer more flexibility, to the level designer to the player. Take Silent Storm, yes JA2 had destructible walls and scenery. But nothing on the scale of Silent Storm's, how much extra work that would take in 2d, if it was even possible? Plus 3d allows for rotating and zoomable cameras which give you a far greater tactical overview than transparency bubbles or wireframe walls.

shihonage said:
I didn't say that 2D was better than 3D for gameplay purposes. However, to claim that 2D is no longer suitable for quality tactical combat in this day and age is not exactly an accurate assessment, IMO. I wouldn't exclude it as a contender, especially in indie titles of which more and more are springing up.
Quality of games has nothing to do with the graphics, FOT had far better graphics than JA2 but was a lot more mediocre in gameplay. JA2 is in many ways still superior to Silent Storm, but not in it's actual level design. The first 3d game to marry the flexibility of 3d with the tactical options present in JA2 would produce a game that no 2d production could compete with.

shihonage said:
Well I guess if you're into modifying the graphics of every game you play, then 2D really presents a problem. However most people don't do that. They create scenarios, maps, scripting, but few actually dare to seriously mod graphics because that takes a lot of time and effort even for the 3D scenario.
Actually I've seen a lot more graphics mods than new scenarios, especially for games that have some form of multiplayer. About the only games that have more levels made for them than new graphics are those where editing the graphics is impossible or a lot of hard work i.e. 2d games like FOT.
 
FeelTheRads said:
You still have to explain me why a player should care about all this?

I don't understand your point.

And installation space? Ahahaha. You did notice that most games now are around 6GB right?

Is this a reason to needlessly expand the needed space by having to store all the animation frames instead of just the models, textures and animation files?

I think feasible is a better word

May be. Although I think it's no longer suitable, as the 3D engine is just better for tactical combat games. Just check Silent Storm (again) and bullet physics.

It is not as flexible as 3d in the way that you would have to re-render all the frames if you so much as changed one texture on a character. One thing though is that all game details should already be planned in this stage, so they shouldn't have to change anything. (in the perfect world of course)

True.

I'm not sure what you're aiming at here. "using a flexible 3d engine" sounds like that a developer would buy one pre-made. I'm pretty sure there's flexible 2d engines out there too which behaves properly.

Easier to use and easier to operate on. Anything pre-rendered is a pain to modify, as you said.

Resource handling maybe? While there are techniques for using very detailed objects in game it gets pretty expensive to have that kind of thing running on every single object in large environments (which is usually the case with birds view). Oblivions environment very pretty amazing in the amount of detail they portrayed, but they did spend a tremendous amount of work on getting it that way while still enabling it to work in real time. With 2d you would use all the bells and whistles for the render, but only have to plan for effective image handling.

You forget about LOD adjusting. It's very common in 3D games and with it you can have all the bells and whistles at a reasonable cost.

Unless it's something like Morrowind, which renders everything, even if it's far away or not seen.

I don't get statements like this. If someone wanted to do a 2D isometric strategic combat game right now, they could, with no problems, and it would look excellent.

At the cost of much more work and time consuming operation on an engine whose features might as well have been done faster and easier with a 3D engine.

In the world of today's hard drive sizes, you can't be serious.

Sitting through half an hour of installation as all 2D animation sets would be copied isn't my brand of enjoyment. As is filling my hard drive up.

Well I guess if you're into modifying the graphics of every game you play, then 2D really presents a problem. However most people don't do that. They create scenarios, maps, scripting, but few actually dare to seriously mod graphics because that takes a lot of time and effort even for the 3D scenario.

a. It's easier and faster for the developers to use and correct any mistakes.
b. It's easier to use for any modders. I modify textures a lot and it doesn't take much effort, only patience and dilligence.

You're referring to normal mapping. Let's say you're doing a strategy game and you have 30 normal-mapped soldiers on screen.

Not just normal mapping. Also textures. A common trick nowadays, models are very simple and textures very complex.

Also the entire environment is polygon-based, normal-mapped, specular highlighted, and applied to all the other fancy tricks to, in realtime. Even with modern hardware, the framerate will inevitably slow down to a crawl.

My C&C3 and WiC think you're wroooong :)
 
Mikael Grizzly said:
You forget about LOD adjusting. It's very common in 3D games and with it you can have all the bells and whistles at a reasonable cost.

If I'm not incorrect here I would say that LOD is used where the distance between a camera and the objects varies when moved around, which is not the case in fallouty isometric style. LOD works wonders in a morrowind/oblivion style where the camera is aimed at the horizon and objects far away simply does not need a great deal of detail. If the camera is at the same distance to the object all the time LOD wouldn't change because that would be apparent to the viewer.

What I want to say is that if you make a game with 2d sprites you won't have to think about polygon restrictions (which is very important in the examples you mentioned) and the hassle of optimization of the rendering process, because it is done already. Both "methods" have their pros and cons and it somehow feels is up to the project planner to make that decision based on team experience and what suits the game idea I guess ("do we want in game cut scenes?" contra "Why the hell would we want in game cut scenes for?!". "Is the camera fixed, or not?" etc). At least how it feels after this discussion.
 
Back
Top