triangles
 

im curios to know what it is with you people and triangles WHATS WRONG WITH TRIANGLES any games ive extracted models from use all tris so is it just a game thing or are they idiots???

im so confused:confused:

I'm sure it's not because 'they are idiots'.

lol ya but everyone here goes out of their way to avoid tris and their just layin 'em down...

I think that a game engine converts the model to triangles, so when your exprorting them out that what your getting.

As for using Quads, the layout of the mesh is cleaner and I think that they are able to deform better when you setting up the character. Also if you go to sub d's triangles give some strange geometry and wont render in mantal ray (up to maya 7) (the subd not the poly)

If your making characters for games then the odd triangle wont matter, as its converted.

If I'm barking up the wring tree then please tell me as this is my understanding at hte moment..

i dont think your barking up any tree simply because people dont bark...

as for the game engines converting them: good to know now i can make some good use of maya and grand theft auto:attn:

also i didnt know that MR couldnt render sub'd tris... that blows...

if you smooth an object is it still a polygonal object or would it be considered crap by a game engine?

Not too sure about games to be honest.

I think that the main thing that you need to look into is the maximum poly count for the thing your importing, for example in GTA a car might be, say, 10000 polys max.

I think that mental ray in maya 8 can render tris. but prior to that it would ignore them.

How do you know I dont bark??

Yeah, gster hit the nail on the head every time there. The other thing about triangles is they're much, much easier and hence quicker to render. On a side note, it's not possible to have a non-planar triangle. Non planar polygons are a _real_ pain to process and light correctly.

There's just a hell of a lot less maths behind doing the calcs for triangles. As well as mental ray going a bit mental:D at some dogy subd topology converted from triangles, subd meshes are much 'lighter' and easier to manipulate and edit than ones that've been made from a mesh containing lots of triangles. Subds are also a much more powerfull way of editing, having the ability to affect areas with a weighting function, far better than say the sculpt tool does.

Trust me, write a scan-line renderer that will only handle triangles, then write one that will handle N-gons, with an upper limit of say 20 edges. As soon as you've done that, you'll realise just how lucky we are to have maya churn through almost anything you can chuck at it. Ngons are not fun to program for.

It can be well worth blocking out in quads, converting to subds to tweak and finalize before being converted back to polys and onto tris for a game engine. It's a very quick way of working. Just ask Kurt.


Simon.

Yeah Simon, forgot abut the render speed of tris, hence why there probably used on games, to squeeze those extra couple of FPS out ofthe graphics card.

everytime you use a tri, God kills a Puppy.

i assumed you dont bark...

ive seen cars people have made up to 30000 polys so i don think thatd be much of an issue

that means god took out the humane society for my mclaren...;)

thanx for the help everyone:beer:

I thought the reason was that DirectX and most graphics cards could only render tris, not quads.

Very well could be, hence why they are converted via the game engine.

I thought that the model was converted prior to being used in a game engine, as only few engine are capable of converting themselves, though I could be wrong.

100% of game engines require to have any and all geometry triangulated prior to it it being inserted as a game asset. This goes from simple objects like a box to the 80,000 poly cars we see in games on the Next Gen Consoles.

As far as I have read the reason for triangulation is to do with the speed of the processors math calculations. CPU and GPUs can calculate tris faster than quads.

I dont think there is any other reason behind it to be honest, its all to do with realtime calculations of the geometry by the GPU and/or CPU.

I think that GPUs only render tris becuase its cheaper to design, manufacture and test (tris need less math, less math less transistors and therefore less cost), not becuase technology can't go fast enough. Anyways, the first Nvidia graphics chip, for a Japanese game console used quads I think (can't remember).