Question, Aardappel?

by ChesterHQ on 08/24/2004 21:03, 5 messages, last message: 08/29/2004 21:40, 1099 views, last view: 05/03/2024 00:57

Aardappel,
will you enable halflife 1 models in the future in cube..that would be much easyer to work with then quake2 models..

i think it would be cool if it works with both .MDL(hl) and .MD2(Q2)

thanks for replying.. btw nice engine :D... im not that good at coding otherwise i'd add it myself ;)

   Board Index   

#1: Don't bet on it...

by Pxtl on 08/24/2004 21:24

The reason MD2 is popular is because Quake 2 is opensource, and because they're simple. MDL models don't have hte advantage that the main renderer of them is opensource, plus they have the added difficulty of skeletal animation. As I understand it, Half-Life supplies the skeletal animation and the models simply follow the animation, as opposed to the per-vertex work of a Quake 2 model. This has many obvious advantages:
-modelling is easier
-models are smaller and easier to download
-models are more extensible if you need to add new animations.

The problem is that it creates a lot of work for the platform developer as the platform has to have animations available separate for the model and run the process of applying the animation to the model. This is much nastier than simply saying "render model X frame Y".

As I understand it, there's more support for looking into MD3, as MD3 allows for more personality in the models (they don't all follow the same animations) and works similar to MD2 with the simple "render frame X of animation Y" instead of worrying about bones.

But Aard can correct me if I'm wrong.

BTW, this is seriously not the right way to post this - does not need it its own thread.

reply to this message

#2: Re: Don't bet on it...

by _Aardappel on 08/25/2004 00:34, refers to #1

modelling is identical for either of them. infact, modelling for vertex based systems is usually easier, since a program like 3ds max takes care of all the transforms, whereas most skeletal systems have limitations in terms of the number of bones being blended etc.

The advantage of using bones is that in theory there is less data to shove around, though somehow these fileformats manage to be bulkier than .md2 (md2 is compressed to a single byte per vertex component).

I could easily add skeletal animation to cube, as I have the code for it lying around, but there is little point. Maybe for Sauerbraten.

reply to this message

#3: Re: Don't bet on it...

by _Aardappel on 08/25/2004 00:35, refers to #2

oh and... on modern GPUs skeletal animation is a LOT faster, as all transforms can be done in the vertex shader.

reply to this message

#4: Well...

by Pxtl on 08/25/2004 13:57

What I meant by "easier" was, if you're using drop-in animations provided by the host program, modelling for skeletal systems is simpler because you don't have to provide the animations.

And in half-life's case, it is an exception to the "bulkier" rule - Half-Life has very compact models. I'm not much of a graphics person - I didn't realise that's what vertex shaders were for.

To me, the main appeal of skeletal animations in an engine is extensibility. If you want a character to be able to swing a sword or drive a car, you can simply make new animations on the old skeletons and apply it retroactively to the models. Plus, the smaller filesize is good too (I love how compact Cube is).

reply to this message

#5: anim8or

by entrunner7 on 08/29/2004 21:40

has anyone here tried anim8or? (http://www.anim8or.com) It's a small, open source 3D modeller that uses bones. It has very efficient models size wise, and is remarkably easy to learn. It also has coloring and texturing built-in: seeing as the present models have only one skin, that could be counted as a good thing.
I had something else to say about it, but I've forgotten what.

reply to this message

   Board Index