I have an intresting question...

[joz]

We all know that to create a large object, we simply fill it whith lots of smaller ones.

Why not just create one large object and have it stored in a local buffer, instead of lots and lots of little triangles re-created every frame?



Are the graphic accelerator creator companies pulling our legs? in the direction they want?

[Miker]

We all know that to create a large object, we simply fill it whith lots of smaller ones.

Why not just create one large object and have it stored in a local buffer, instead of lots and lots of little triangles re-created every frame?

Are the graphic accelerator creator companies pulling our legs? in the direction they want?

By smaller objects? you mean polygons?

[cvasquez]

Not sure what you mean bro...be more specific as opposed to have us guess what you are trying to say? I am assuming you are talking here about graphic cards.

[Goliath182]

Because from different viewing angles the needed triangles are changed for example you have a 3-D object that is in the shape of a triangle. The triangles that make up this triangle on a flat side would have a flat side facing the outside of the shape. The triangle on a end of one of the protruding sides would have a point facing out. So if you tried to move this object you would have to move the placement of the triangles to have the same object. You may say well isnt this all rendered? The part of the object that you see is rendered so the individual triangles change as you see a different side of the object. Too get more in detail if you moved this object and had a protruding side facing you you would just see a different size triangle. That is what polygons are for. But polygons are limited by how powerful your PC is. So you cant make everything in polygons. Point in case, textures. It would take a whole lot of processing power to make all the rocks and other objects in a game have a unique shape, so video game designers use textures that make a object look like it has "bumps" when in fact it is a simple polygon.

Lot there lol hope you get it.

[joz]

Yes, but what if you are viewing an object continualy from the same angle and distance.

It just seems to me that we are getting screwed out of alot of graphics power. I mean, for the ammount of time im sniping from a position looking at the spawn points on top of big, ass objects, wouldnt it be easier to just render it a few times, texture it, and save it? instead of re-rendering it every frame?

I mean, it seems to me that the current method of triangle rendering kinda sucks,

[Miker]

What makes a GPU better at graphics then a CPU is that is it doesn't multitask like a CPU. All it knows it doing math, fast to produce polygons. Making it put a picture to texture would just slow it down. And no game has 100% still parts, everything moves a little, or you would just be looking at a picture. Did you ever play Resident Evil 64?

[Goliath182]

You could probably save some if you didnt continuealy re render but the GPU wouldnt do anything else anyway so you might as well have it keep on rendering.

[Frag Maniac]

Yes, but what if you are viewing an object continualy from the same angle and distance.Then it would be more of a sprite than an object, and sprites are generally only used at farther distances to reserve resources for more noticable objects in the foreground. Even with sprites one has to be careful what objects are used to avoid the cardboard background look. Typically trees are reservered for sprites, and as mentioned, only distant ones. http://en.wikipedia.org/wiki/Sprite_(computer_graphics)

Besides what has been mentioned about the need to make a 3D shape appear realistic from all angles, you have the need to gradiate color for shadow and lighting effects too. Essentially what you're propsing would take us back to the 2D stone age of computer graphics.