Shannow said:
DraQ said:
See DA2 non-essential NPCs for a glaring counterexample. If a game has patchy, crap/awesome appearance, it's jarring. You can't focus on well looking stuff, because every once in a while something horribly crappy floats into your FOV and demolishes the illusion of the world. And you can't adjust and come into terms with crappy graphics like you can in old games, because you're constantly exposed to high detail elements that disrupt your adjustment.
Also, it's worth of note that textures can confer more detail and more uniformly than geometry (that's their whole point, damn it!), even though they are less resistant to scaling (like due to perspective), so wrapping new geometry in >10y.o. textures doesn't work, while doing other way around might.
Finally, maps of all sorts are typically using textures, IIRC, so, while useful, they are not going to add any additional detail beyond that added by textures. If your textures are extremely low res, the maps will also be low ress and they will give your models distinct, half-molten looks of really low budget plastic props.
Yay for immersion, I presume.
Yeah, well, not knowing anything about the actual technical side of stuff
Well, you don't really need to be an expert when it comes to technicalities, being perceptive fellow with some minimal knowledge suffices.
Considering the the baby-step improvements in 3D games in the last 10 years, it'll take them at least another 10 years to consistently make games that actually look as good, as ToEE2 could have looked in 2005/6 if the industry hadn't given up on "dated" 2D graphics...
2 decades lost to stubborn ignorance
It's not that simple. First, there is a metric fuckload of things 2D graphics simply can't do. And it's mostly interesting stuff, although sadly underutilized in most games.
Second, modern 3D, even half assed, is vastly more detailed than 2D. It's just that the perspective tends to expose non-detailed bits more.
Third, 2D has peaked long ago. It wouldn't really improve because it can't really improve. Pretty much the perfection in 2D raster graphics is pixel art. The problem with pixel art is that development cost and storage requirements shoot beyond hilariously retarded and "are you fucking insane?" the moment you start improving resolution and colour depth. A bit worse in theory, but infinitely more practical and still fucking impressive in right hands is using some powerful raster graphics software like photoshop. There are still storage requirements and lack of reusability, though. To combat these, you can do two things - first, you can cut your assets into small, reusable pieces, second, you can apply all sorts of aftereffects, like having game calculate shadows rather than drawing them by hand, but then you run into a simple problem - whatever you do can be done simpler and better in 3D, because, for example, there is no way you can deform a shadow according to the shape of terrain if there is no shape of terrain encoded and it's all just a picturesque bitmap. And then there is pre-rendering stuff, which has the only advantage of being capable of displaying stuff modern hardware can't but at huge price of nearly zero flexibility. It was viable back when game farted blood when it had to display several textured, Gouraud shaded polys, but now the well made graphics can look really good, the problem is relative lack of well-made graphics.
Now the problem is not 3D, the problem is mindlessly applying hardware tricks thinking they will solve all the problems. They won't. Specular maps on everything just makes everything looks like cheap plastic casts. Likewise, parallax maps aren't fix-it-all when it comes to lack of detail. Bloom doesn't make everything look awesome nor realistic. And nothing, absolutely nothing will help if you have shit textures.
The first textures were just that - non-specific bitmaps that helped to convey what kind of surface you were looking at. Then, it was noticed that you can add a lot of detail to your 3D geometry if you paint it on as good old 2D raster graphics and textures in modern sense were born. Notice the keywords: 'lot' and 'detail'. The whole point of texture is adding detail, in other words it is only effective if it's much more detailed than your geometry. All the modern -maps are just an extension of the idea of texture, the may paint on additional spatial detail by working as a height map on the surface of the model, change reflectivity selectively and so on, but they are still stored in raster format, meaning big pixels in, fugly shit out. Take a look on wireframes of models in an old game, say Unreal. Now take a look at actual, textured models in game - see the difference? Now do the same in oblivious.
Lastly, there are well known tricks that some developers just plain ignore for no apparent reason, for example detail textures. Detail textures are pretty much going back to the root of the texturing - non-specific texture rather than details typical of some particular object. They are used by applying the same texture in smaller and smaller versions recursively upon an already existing one. It's a simple, yet effective way to prevent stuff dissolving into blurry mess up close. It was used to great effect in many games, like Unreal. It's still being used in many games, also to a great effect. But for some reason, it wasn't used in oblivious despite the game urgently needing this kind of attention.
There is a reason why I'm using oblivious all the time, this game was so badly executed in all possible ways, including graphics, that anyone defending it must be a moron.
