View Full Version : Why 3d models should be combined into one mesh/object?

01 January 2009, 10:26 AM
Hi to the community,

This is such a newbie question I guess.

I'm learning modeling and texturing using Maya Software.
I've seen many tutorials over the net and I notice that the final complete model is always combined into one mesh (instead of keeping different parts like head, arms, shoes,...).
So I wonder what is the reason for doing this? I guess this technique is only good for game model in case of optimizing the models file size. But I may be wrong.
But what if some parts of the model need to be independently animatable from the other parts?
e.g. eyes for human model, wheels for car models,...

Thanks for your answers.


01 January 2009, 02:01 PM
Shape keys are one fairly simple way to get what you're looking for.

For example, suppose you've modeled a car and want the doors to open. Set your base shape key as everything being closed, and then set a key for each door, individually, being open. You should then be able to animate the doors being opened and closed, using the shape keys. The same thing should work for your other examples.

01 January 2009, 06:58 PM
Generally, when it comes to human (or animal) forms, you want to keep everything attached.
This is what rigging is for.
Obviously this can be dependant upon the needs of the object being modeled,such as keeping clothing separated from the mesh to run simulations on. But you're going to want to attach arms, legs, heads, and anything that is connected to yourself 95% of the time. This way you dont see a gap between say your arm and your shoulder when you animate, and people look and say 'His arm is broken off'.

As for cars, I really couldnt tell you why anyone would attach the wheels to the car itself, that would seem a bit strange :surprised But your eyes can move independantly from your head, and if they rotate in your eye sockets, generally nothing else visible will blantantly move along with them.

Hope that helped :)

01 January 2009, 09:06 PM
Thanks for all your answers.
But I think nmcelmury has more understood my request.
So I know now that I have to decide at the pre-modeling step if I must keep some elements independent or not . As you say, it dependents on the future use of this element (like eyes, clothes, car's door, wheels,...).
So as modeler I think then I must do some checklist before going indeep, like:
-does this part is animatable?
-does this it move independently from over parts?
But does this technique apply also to game modeling? I mean does today's game engine can also manage object with multiple independent sub-elements? What could be the constraints of today's game engine in terms of modeling(polygons constraints) and texturing (e.g. multiple textures for one object with some sub-elements)?.
In fact, I want to explore more modeling & texturing for game.

01 January 2009, 02:05 PM
one technique is to model the mesh so you can assign selection ID's before
'assembling' the mesh.

as an example, modelling the eyes separately, then moving them away
from the mesh before attaching them to the mesh (in 3D Max.) that way
you can select them, and either create a named selection set, or assign
a selection ID.

once the eyes are planted in the sockets, selecting them can be difficult.
but if they've been given a distinct polygon selection set or polygon ID,
then it's no problem.

eventually i will need to learn to do this in Maya, also.

01 January 2009, 03:03 PM
As a side note any reasonable game engine will support multiple seperate meshes in a single mesh file.

The standard game mesh file will generally look like this (hierarchical structure)

scene root
| | |______ mesh 1
| | |______ mesh 2
| |_____ mesh 3
| |_____ mesh 4
|_____ bone 1
|____ bone 2

01 January 2009, 11:18 AM
nice pic explaination GrogMcGee thank u

CGTalk Moderation
01 January 2009, 11:18 AM
This thread has been automatically closed as it remained inactive for 12 months. If you wish to continue the discussion, please create a new thread in the appropriate forum.