Is This Possible In Any Way?

[IchI]

I was thinking about how normal maps work and how a 2d texture can react with light. This made me think about how u see a game when you are playing it. When u play a game it is initally flat. But its prossed in 3d information. What if u had 3d information but it was prossesed as colour. U would make the game in a 3d design tool. But when its actually prossesed why does it actually have to be 3d? If u think about it when u play a game its only a image anyway. Its sort of an interactive image. What if as a starting point. U had a player model. This player model was a deforming 2d plane. On this plane it had all the information for your 3d model. When it movies it works like an interactive movie. But its only a 2d plane. You could actually have real graphics in games if this was anyway possible and also the need for very powerful graphics cards would not be a question. You would only have to have a specialised texture possessing gfx card. What do u think? Am i talking out of my ass?

 

[SLH]

Interesting post, rather wierd though. When making a game to be percieved as 3D, it's a lot easier to think about it in 3D when designing.

Not sure if i'm understanding you correctly at this point...
I also think that while the amount of processing power needed would (i think), if done correctly would go down, the amount of data required to store the model would be huge. When you concider that you'd need every possible position of the model stored as an image; in each direction! Thats 360^3 (46,656,000) images for 1 model (i think)! I think you get the idea.

 

[dawdler]

I cant understant a thing of it

What I do know is that there is a reason for doing stuff the way 3D engines do... Its simply the way to do it.

And 3D isnt a image... Never... Its coordinates. Just a bunch of vertices. Its never an image. What you see if merely the second layer that dresses these vertices with textures. Rotating and moving something becomes very easy this way. A picture needs everything prerendered. And that would take up so much space its insane. You couldnt even fit one HL2 level on 3 DVD discs (at equal quality as 3D).

I still dont have a clue whether I'm even close to discussing the same topic though

 

[Nostradamus]

Originally posted by dawdler

I still dont have a clue whether I'm even close to discussing the same topic though

think you are

a good idea, but as the 2 others said, it would require a stupid amount of images...

ow normal maps work ... hmmm

not a clue, but they look cool

 

[Samuel]

Why not just use a shader that makes things appear flat and cartoon like? I think that's what your trying to do..?

 

[IchI]

no.... I am not trying to do anything. I am saying is that why do 3d games have to be 3d. As long as it looked 3d it wouldn't matter if it was 3d or not. Think of it as a sort of like.... urr..... well when u watch a movie its not 3d is it. Its a flat animated image (video). Why do games have to work in 3d enviroments. If the computer knew how to transfer 3d data into 1 flat plane, so lets say u did what I am saying on a character. This character is 1 flat plane. On this plane it would have something like human model data onto it. It would work like your everyday model and would need animating and would actually be made in 3d. Its the way the games engine would transfer that data into a image. Sort of like... how can i explain it. Its so hard to explain.

Hmmmm......

Its kinda....like southpark. Well character wise anyway. Lets say u have a flat peace of paper (2d plane) and on this peace of paper u had cartman from southpark, as he turns around 180 the peace of paper (2d plane) stays where it is but it projects as if cartman is moving around. If it was in a games engine the 2d plane would morph to follow the shape and edge of cartman.

did that make sense? lol soz its so hard to describe

does that make sense?

 

[dawdler]

Originally posted by IchI
no.... I am not trying to do anything. I am saying is that why do 3d games have to be 3d. As long as it looked 3d it wouldn't matter if it was 3d or not. Think of it as a sort of like.... urr..... well when u watch a movie its not 3d is it. Its a flat animated image (video). Why do games have to work in 3d enviroments.

Because its so much faster than pre rendered! Thing is, you NEVER know how a player will move. NEVER EVER. In a movie this isnt the case, as you know exactly how the camera move. 3D in a game is what the computer is good at, crunching numbers. Even fake 3D (ie sprites) is numbers.
But the difference is this, as an example:

Take Daggerfall, one of its NPCs. Its got what, 1 head on, 4 different angles? And the same for the backside. That's 10 images is total. Now, all this have to animated. And you have to check at what angle the player and NPC is at to display it correctly. Fairly basic.
In true 3D however, you have a animated polygonal character. And that's it. No angles, no need to edit multiple images, and less to keep track on. The engine then keep tracks on all the vertices, and draws them accordingly, to form the character in 3D space.

See the difference?
1 animated polygon.
10 separatly animated and drawn images.
Its easy to guess whats the easiest and most effective way.

The way you are thinking would need like those 46,656,000 images like SLH said to create something of equal quality of a polygon.
Sure, it IS faster to draw it. But its gonna take a day or so to load them all
Which is best? 1 polygon or 46,656,000 images? Tough call eh?

 

[IchI]

u still do not get what I mean at all. It would still run on a 3d based enviroments and also 3d models. Its just the way the imformation is prossesed.

Right this is as simple as I can put it and it still doesn't really describe it very well. U have a game like, lets say half-life1. When u play it and move around u are inside a 3d enviroment. There is a player infront of u he is an actual 3d model modelled in an exsternal program. The model starts to rotate at 90 degrees slowly. As u watch it looks very normal and just like a 3d model would work. But what the actual engine for the game is doing, is sort of recording his movments, lets say with your everyday camera. then playing them onto a 2d plane. This 2d plane obviasly is very very low in polgons and would morf to the outline of the model. Everything is STILL! 3d its just what I was thinking is maybe u could change the way a 3d engine prossess information. Basiclly why do u have to prossess such a complex 3d polgoned model as a fully 3d model. When at the end of the day its only a place on your screen filled with thousands of indervidual colours.

 

[Samuel]

Originally posted by IchI
no.... I am not trying to do anything. I am saying is that why do 3d games have to be 3d. As long as it looked 3d it wouldn't matter if it was 3d or not. Think of it as a sort of like.... urr..... well when u watch a movie its not 3d is it. Its a flat animated image (video). Why do games have to work in 3d enviroments. If the computer knew how to transfer 3d data into 1 flat plane, so lets say u did what I am saying on a character. This character is 1 flat plane. On this plane it would have something like human model data onto it. It would work like your everyday model and would need animating and would actually be made in 3d. Its the way the games engine would transfer that data into a image. Sort of like... how can i explain it. Its so hard to explain.

Hmmmm......

Its kinda....like southpark. Well character wise anyway. Lets say u have a flat peace of paper (2d plane) and on this peace of paper u had cartman from southpark, as he turns around 180 the peace of paper (2d plane) stays where it is but it projects as if cartman is moving around. If it was in a games engine the 2d plane would morph to follow the shape and edge of cartman.

did that make sense? lol soz its so hard to describe

does that make sense?

South Park is actually created in Maya now, this is really evident in the movie when they mix that rendered fire in with the rest of the scene.

I think I understand what you mean, and it's already been answered above. So games do use pre rendered sprites that look 3D, take Fallout. But even with Fallout the amount of tiles required for each actor was huge and approaching the memory reference limits of 32 bit computers.

 

[Incitatus]

Originally posted by IchI
I was thinking about how normal maps work and how a 2d texture can react with light. This made me think about how u see a game when you are playing it. When u play a game it is initally flat. But its prossed in 3d information. What if u had 3d information but it was prossesed as colour. U would make the game in a 3d design tool. But when its actually prossesed why does it actually have to be 3d? If u think about it when u play a game its only a image anyway. Its sort of an interactive image. What if as a starting point. U had a player model. This player model was a deforming 2d plane. On this plane it had all the information for your 3d model. When it movies it works like an interactive movie. But its only a 2d plane. You could actually have real graphics in games if this was anyway possible and also the need for very powerful graphics cards would not be a question. You would only have to have a specialised texture possessing gfx card. What do u think? Am i talking out of my ass?

Can someone translate that into English?

Someone mentioned 3d being coordinates. From what I think I understand of the first post, that seems to be the proper rebuttle (<-spelling?)

 

[dawdler]

Originally posted by IchI
u still do not get what I mean at all. It would still run on a 3d based enviroments and also 3d models. Its just the way the imformation is prossesed.

Right this is as simple as I can put it and it still doesn't really describe it very well. U have a game like, lets say half-life1. When u play it and move around u are inside a 3d enviroment. There is a player infront of u he is an actual 3d model modelled in an exsternal program. The model starts to rotate at 90 degrees slowly. As u watch it looks very normal and just like a 3d model would work. But what the actual engine for the game is doing, is sort of recording his movments, lets say with your everyday camera. then playing them onto a 2d plane. This 2d plane obviasly is very very low in polgons and would morf to the outline of the model. Everything is STILL! 3d its just what I was thinking is maybe u could change the way a 3d engine prossess information. Basiclly why do u have to prossess such a complex 3d polgoned model as a fully 3d model. When at the end of the day its only a place on your screen filled with thousands of indervidual colours.

Hm, so you mean basicly rendering it normally in 3D, then grab a picture of it, and render it in 2D, but realtime? That wouldnt be very good at all... It would be like running two engines at the same time. Doubt any computer could run it. Not to mention, its gonna be VERY hard to figure out what is what, how do you know what the edge is of a character? Or objects? There would be a really big problem creating a depth in the screen.

 

[synth]

Castle Wolfenstein and DOOM were advanced 2d giving the illusion of 3d. (Quake was the first fully-3d game). I think that's what you want, but with better graphics.

 

[SLH]

Originally posted by IchI
Right this is as simple... <SNIP!> ...thousands of indervidual colours.

So, you mean as if a light behind the model is projecting a shadow onto the users screen, in colour??

 

[IchI]

sigh..... Its so hard to describ what I mean. Not to mention how shite I am at english.

 

[dawdler]

Originally posted by IchI
sigh..... Its so hard to describ what I mean. Not to mention how shite I am at english.

Explain really slowly

 

[Dan]

I think most people here understand what you mean, and they've responded to it pretty well. The problem is that even if you don't want to actually render 3d, you're gonna have to save an image of that somewhere on your harddrdive, and you're gonna need millions of images to equal the possibilities with one 3d model. I saw you refer to the normal map of the cliff. Well that 2000 poly cliff was indeed saved as a single image and then displayed on a flat surface. But imagine if that cliff were say a player model, you could have a nice pretty looking normal mapped player model that would look good from straight on, but it wouldnt' be able to move or anything unless you also normal mapped every single frame of every single animations from every single angle. Thats a lot of normal maps.

 

[babywax]

I thought he was talking about heightmaps...
Basically the same way that battlefield stores it's information.
The only problem with that though is that video cards thing in triangles, each "poly" is a triangle.

 

[coax]

Ichi, I think I get what you mean now.
Most engines don't actually render the sides of the 3d models that you don't see.
Like, the character has his back turned to you, the front of the char is not rendered when you move.

Storing the polygons that aren't rendered isn't a big job for the computer nowadays..
When you lose FPS in games, it's because things have alot of polygons, and alot of detailed textures that the graphics card has to render..

To combat this many engines have a distance meter, so to speak, it doesn't render things far away to make it easier on the graphics card and get higher fps.