3D for Games
polycounter lvl 4
Hello,
I have just started getting into 3D and have some questions that I hope somebody will be able to answer. (they might be realy stupid but please be nice T.T)
So i've noticed that in games, surfaces seem to be very jagged, for example a circle is never really circular.
I understand that it would create a very dense mesh if you tried to make it an actual circle and imagine that would be too harsh for a game.
This branches into 3 questions:
1. So to make 3d models for games should I not use turbosmooth (for example) and maybe use bevels and things to give it the appearance that it is less jagged?
2. Why don't games use turbosmooth on surfaces? Or something similar. (I am under the impression that it would make the jagged edges smooth without adding in extra geometry).
3. So i've heard that video games use geometry made out of triangles. Do people model things in quads and then somehow they transform it into triangles? If this is the case, how important is it to model something with correct topology or for example use bevels in edges if its going to be transformed into triangles?
Thank you for putting up with my weird questions.
I have just started getting into 3D and have some questions that I hope somebody will be able to answer. (they might be realy stupid but please be nice T.T)
So i've noticed that in games, surfaces seem to be very jagged, for example a circle is never really circular.
I understand that it would create a very dense mesh if you tried to make it an actual circle and imagine that would be too harsh for a game.
This branches into 3 questions:
1. So to make 3d models for games should I not use turbosmooth (for example) and maybe use bevels and things to give it the appearance that it is less jagged?
2. Why don't games use turbosmooth on surfaces? Or something similar. (I am under the impression that it would make the jagged edges smooth without adding in extra geometry).
3. So i've heard that video games use geometry made out of triangles. Do people model things in quads and then somehow they transform it into triangles? If this is the case, how important is it to model something with correct topology or for example use bevels in edges if its going to be transformed into triangles?
Thank you for putting up with my weird questions.
Replies
1 & 2 - when making assets for games it's about keeping the mesh optimised. Making it look good without going too high-poly. Saying that, game engines and hardware these days can handle a lot more than they used to, so a suitable poly amount is always changing. Turbosmooth increases the poly count massively once you've collapsed the modifier stack which is one reason it isn't used often.
You don't necessarily need to use turbosmooth or meshsmooth, as you said you can use bevel or you can use the smoothing groups to help make a mesh look more circular/smoother. If you did use turbosmooth, you can always remove any loop edges once it's applied.
3 - Engines will, if they need to, convert your quad mesh into tris. Unity for example does this when it loads in an FBX file. And as far as I've seen, they can take very complex meshes. Everytime I use model in Max I will do so in quads. Saying that it's still worth keeping a clean mesh, it makes it easier to optimise and alter later on.
http://wiki.polycount.com/wiki/Rendering
With films I'm not too sure. From what I understand however is that the only limitations they have are time and the financial budget. As films are pre-rendered I would assume there isn't much limit to their poly count or texture size limit. With films it's all about making it look as good as possible. This might explain it a little better - http://blog.digitaltutors.com/whats-the-difference-a-comparison-of-modeling-for-games-and-modeling-for-movies/
I wouldn't know about working for both industries as I'm more involved the games and simulation side, but the modelling is essentially the same. Some of the workflow might be a little different though with texturing for render engines for films.