I recently saw the other topic on tri’s, and I now know to try and avoid them, but i’ve downloaded some models off the internet and it seems that tri’s are extremely common.
So I began to wonder, if tri’s are so bad then why are they so commonly used? And also, being the noob that I am, how do you even model using tri’s? I only know how to model by subdividing or extruding, which both create squares.
Some 3d formats can only use trianlges. If it was a model ripped from a game it will most likely be triangles because that is what the game engine uses. Tt doesn’t mean that it was made with triangles only that it was triangulated when it was saved/exported.
Like I said in the other thread, it’s because quads are “only a concept”.
Every 3D model is made of triangles and quads are simply a representation of them so you see poly- and edgeloops better and that subdivision works like intended.
For instance .3DS models in the web will always be triangulated, it’s one of the file formats specifications, others simply have no idea what they are doing, and triangulate their .OBJ files on export, and some make it deliberately either to guarantee the direction of the diagonal on import, or to make it harder for people just to abuse their models… various reasons really.
And you don’t really want to model with triangles unless you absolutely know what you’re doing. You can always select one or several quads/polys in Blender and press <ctrl+t> and they’ll be triangulated. You can also turn the diagonal after triangulation with pressing <space> and looking for “rotate selected edge” but only in edit mode. And you can re-join 2 triangles to a quad with <alt+j>.
Triangle based modelling is, or better was usually used for really low polygon stuff, and I am speaking about applications where it counts if you have 10 triangles more or less to rasterize.
It became almost extinct, like the skill to create 2D sprites on the C64 for instance, simply because technology paced along so fast.
However with the recent retro-wave and OpenGL ES on mobile devices it became somewhat relevant again. Now as we enter the era of Tegra3, high performance OpenGL ES chips and Mobile Quadcore devices the need for real low poly stuff already fades again.
But a really optimized mesh which is deformed for games usually is triangulated and the diagonals are rotated to optimize deformation and geometry.
Ohhh okay that makes sense now. So they’re not generally modelled using tri’s, but they are just automatically converted to tri’s when exported.
arexma your answer really helped me understand the concept, thanks a ton guys