Is there an easy way of creating LOD models in MAX?

#1

Hi…

I’m working on a PS2 game with an incredibly tight deadline. I have to create LOD (level of detail) models for the characters in the game, but have to keep the same physique wieghts, mapping co-ords etc

Is there a quick way to do this, or even better a plug in which will reduce polys and keep the UVW’s intact???

cheers

kenny

#2

Try the multires modifier. The geometry may get a bit ugly, but if the deadline is tight it might help. I doubt it’ll allow you to preserve the vertex weights for physique, but it should preserve the UVs pretty well. You can always save out the physique envelopes from the high res model and reload them as a starting point for skinning the new model.

#3

You could try the mesh-optimize modifier (Modifiers -> Mesh -> Optimize)

I just found out about this plugin today which may help you out.
http://www.npowersoftware.com/trans/trans.html

I always construct everything in NRUBS then go to mesh…but that’s just me.

#4

What engine are you using? There should at least be some basic lod algorithms built in.

#5

Yeah, I’d think LOD would be default in a game engine. I’d like to know if there is a way to set it up for rendering in 3DS Max. It would be easy to write an algorithim for one if one doesn’t exist.

#6

Yeah the LOD algorithm will be in the engine, but normally all that does is swap between separate pre-defined models depending on distance (this sounds like the case in this instance). You still have to make the models in the first place.

Dynamically generating LOD models is computationally expensive and rarely done on a platform like the PS2. If you’re using higher order surfaces like patches, then LOD comes as standard- they’re meshed at runtime anyway.

And PhantomDesign, there is a LOD utility for rendering in max, it’s under more in the utilities panel ‘level of detail’. I’ve never bothered with it though since it only works in renders, not the viewport.

#7

Dynamically generating LOD models is computationally expensive and rarely done on a platform like the PS2.

I am not so sure about that.

#8

Thanks for the information.

Dynamically creating LOD models would not only be heavy on system resources, but would have an increased probability of large visual errors.

#9

Dynamically creating LOD models would not only be heavy on system resources, but would have an increased probability of large visual errors.

Visual errors will be quite minimal if done right.

BTW, the whole point of LOD is to free up resources, not go heavy on them. So I don’t understand what your angle on that is.

#10

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.