Game Theory: Making a gun

I just want to get this straight: how to make the most effective gun for BGE

  1. create a quad model
  2. unwrap gun and pin verts to image using “p” in image editor
  3. use triangulate modifier and apply
  4. unwrap again
  5. texture gun

Is this process a recipe for creating the most efficient and effective gun?
Or is there another better way that game industries use…?

Thanks :slight_smile:

This is the method that I use:

1)Model the gun: I just model every model poly by poly.
2)Unwrap the gun: I use seems, even for structures.
3)Make the textures for the gun: diffuse map, normal map(optional), specular map(optional) and I use GIMP with some photos, a image of the UVs and some references.
4)Texture the gun: setup the materials(normally i use only one material, except for structures), setup the textures(diffuse map, normal map, specular map), ect.

Triangulate modifier? if you know what are you doing, you don’t need modifiers (except for mirror and armature).

When you start to model something, you need to know the max number of polys, if you pass the max number, you don’t need to use modifiers, there’re tools that can help you, like edge loop, or just delete some faces.

I apply all this to every model, I don’t care if the model is a gun, it can be a barrel, a car, a person, etc.

Ahh ok thanks :slight_smile:
Someone told me that triangles are the most efficient poly to use for game engines though

I’ll try and just use textures, normal maps, and AO maps to cut down verts

What do you mean with “most effective”?

Most realistic (I know that’s opinionated, and I mean best quality results)
Most game-efficient and uses least resources. Ex: a 12,000 poly VS 1,200 poly gun
best collision bounds

best results in general

Why a gun need collision? if you talk about guns on the floor, that use a low poly mesh as collision mesh. About the collision bounds, i don’t know if tris are better than quads, because I normally use a low poly mesh as collision mesh using convex hull.

When someone model a object, normally make it with quads(i think nobody want to model something with tris, it is difficult) and after finish the model, it is converted to one triangulate mesh. For example, Unity not support quads, only tris, so if you import a quad model, Unity automatic convert the quads to tris, and i think AA games use tris, but the models are from the beginning made with guads.

Not all models are made just with quads, some have tris, but but in small quantity, example: cars.

BGE support quads, so is not necessary convert it to a tris, and if you want to make it, use the triangulate modifier (apply it).

You don’t ever need to worry about triangles when making game assets. Most engines will just convert quads to tris (actually it’s more like it handles them as tris) automatically.

When some of your quads are non-planar auto-triagulate produces different results depending on which way it will be split and most likely the other will look significantly worse than the other. These cases you should split manually as you model or model so that all your quads are (more or less) planar. N-gons should be split into quads / quads & tris.

When you make real time assets it’s mainly just your everyday modeling. Just remember:

  • pay attention to polycount
  • faces are (usually) only visible from the side their normal is pointing
  • understand manifold (closed) meshes, how to make one and when you can make non-manifold parts
  • analyze which parts of the mesh are more visible and which not perhaps visible at all and add detail accordingly