Page 1 of 1

Choosing the exporting pipeline

Posted: Sun Dec 11, 2011 1:25 pm
by progmars
I have read so much about various exporters for Ogre 3D, it seems each of them have some problems, so it is hard to choose the right (and mostly free/open source) tools to get things done easy.

My game contents can be described as following:
- game characters - humanoids only
- animation - bones only (also for facial animations)
- materials - nothing complex, I still feel like a noob in texturing...

But there is "a catch": I need also some predefined vertex animations (morphs) and recolorable materials.
From the user perspective, the user will be able to choose the base mesh and deform it a bit (using pre-defined morphs, like adjusting the "fatness" of the character, and also some facial features). After those deformations, user clicks Apply button and I need to save the modified mesh as his avatar mesh (thus avoiding using vertex animations later in-game).
Also the user should be able to take some base texture and fill the colors with custom values. As far as I know, Sims3 works like this, and I have seen some posts here about recoloring materials in Ogre3D, so this should be doable.

As I am mostly a programmer, my 3D exporting pipeline idea is pretty vague now. The Blender exporter seems still a bit tricky to use, but I guess, there is no better free alternative, so I'll have to be patient and get used to Blender exporter.

If you have a stable exporting pipeline (preferably, using only free/open-source tools) for humanoid animations, I will be glad to hear about your experience - what 3D software/exporter combination has proven to be the most hassle-free for you?

It would be especially great to have a good rigged and well-tested low-poly human model (with high-res also available, if I need to tweak something) with bones for facial animation, so if you have used such a model in your games, it would be great to know. I have seen some pretty good free rigged models, bot they usually do not have facial bone rigs :( .

Re: Choosing the exporting pipeline

Posted: Sun Dec 11, 2011 3:41 pm
by metaldev
I think you are stuck with blender, because all the other 3D software either cost money (maya/max/ etc) or are not going to fill your needs (sketchup/ etc).

Re: Choosing the exporting pipeline

Posted: Sun Dec 11, 2011 6:43 pm
by progmars
I found a thread which is very similar to my problem:

Blender exporter is evolving, so it should be stable enough soon and I can wait, I have much coding to do until then. Or if it does not reach the stable state, I can for some time use 3DS Max educational version (at least while I am connected with my local university).

But the second issue for me is evaluating free resources. I could choose a 3D rigged human model which initially seems fine. And then I try to add a facial rig to it, following some of the tutorials, and I discover that the mesh topology is not appropriate for skeletal animation. Or maybe the rig is not adaptable for the free motion capture data from CMU. So I need to start looking again. Trial-and-error hell.

I have watched many 3D modelling/animation tutorials, but they mostly were specific to their own tagret audience (animators, renderers, modelers). It is hard to find videos which show how to create and animate a high-res mesh with skeletal facial rig and then export it to a low poly version for some game engine.

I have no experience to evaluate: this model is good because it has xxx, but this will not work for me because of yyy. I would trust any experienced 3D content creator, who would suggest me a good model for my needs, but I think, experienced 3D content creators create their models from scratch :lol: so they have no time to explore and evaluate freebies.

Re: Choosing the exporting pipeline

Posted: Sun Dec 11, 2011 11:02 pm
by metaldev
I have time to give you a small advice:


If you have no knowledge of art pipelines, why would you begin by making a high res character, presumably baking down normals to a low poly mesh, blended weights etc. ? motion capture?!
I don't mean to discourage you, but IMO, that's sort of like someone's first programming project being a networked 3D game.
Let me relieve you of your delusion that motion capture will somehow magically integrate with a skeleton: motion capture is no panacea and you need someone who knows their sh*t in animation and is comfortable with rigging and character animation curves - or you are just entering a world of pain.

Why not make more representative art?
Here are 4 examples of simple characters that could work successfully in a project but wouldn't require a cutting edge character pipeline:

Otherwise if this is not possible, I might recommend hiring or partnering with a professional artist.
(is this not a similar advice you would give an artist who wanted to start learning programming by making a networked 3D game?)
That said, if you are just trying to learn, have at it - I'm sure you will learn a lot from trying. But if you are doing this for a client or something... beware.