making near and far objects for games like doom 3?

[SIM-O]

Hi, I've made this high res model for doom 3. i figured that since ID soft made low res and high res, or far and near models, i should too. is there a way of making near and far objects without starting again each time? (plus keeping the uv map for obvious reasons) eg, could i just reduce the vertex manually?

[shashanks18]

hey SIM...well the thing with games is Game Engine.....game engines are programmed to automatically reduce or increase the resolutions of the models as well as their textures.....modellers dnt hv to care about all tis....jus make a good model.....wthin a good poly-limit ofcourse :P

can i see the model??? thx :D

[SIM-O]

it's got a little bit to go, to be finished, so I'll just finish it up and show you.
it's currently 1478 vertexes. it's a player model, but not a lot of detail needs to go into it. it's the player character from System Shock 2. I'll show you the artwork I'm basing this off.

System Shock 2:\ Multimedia Databank

System Shock 2:\ Multimedia Databank

System Shock 2:\ Multimedia Databank

[SIM-O]

Hey, I'd still really like to know how to make near and far objects, i want to know the WHOLE character model process, because the far objects are tied in with the high res objects, so I'd like to work this out so it works well in doom 3.

ok, here is my basic mesh without uv maps, smooth groups etc.

it's 1467 vertex

[milki]

Okay, so:
i have 1,5year of professional experience in game character modelling.
The near and far objects you are talking about are named LODs. lod - level of detail. level 0 is highest resolution (the closesd object to camera)

There are 2 approeaches to do LODs (there are many algorims, but in general theere are 2).
1. You have to make every lod manually by yourself.
+faster for engine, because it dont have to coumpte it
+quality is based on your feeling, not some "stupid" algorithm
-larger memory consuming
-time consuming for artist
2. Let computer do lods.
+no modelling time
+smaller memory consuming
-you cant control geometry of LODs, so you can see mistakes, that you don't want
-its expensive computing

In that company we had to make 1. choise because of ram on ps3 (terrible low number)
But in my opinion... it was better choise anyway. The models looked great at distance too with even more reduction, than engine will compute. But ofcoarse... it took whole day to make 4level LOD.

And last thing to add:
if you like character like in doom3 and later... you need to make highpoly (really high poly (in z-brush like 5mil)) and bake that model to normal map of you game model. Without normal map... models will look flat for today expectance.

Hope i answered all you wanted.

[SIM-O]

Okay, well, if you know how to make doom 3 compute lower polygon models, don't be shy ;) or even if you know a source of information, please tell me.

if you like character like in doom3 and later... you need to make highpoly (really high poly (in z-brush like 5mil)) and bake that model to normal map of you game model. Without normal map... models will look flat for today expectance.

i had already started working in zbrush when you posted this :P I'll show you when it's done. Or maybe I'll even post it properly, as a good high resolution zbrush work usually deserves a thread.

If if do make the LOD 1 (as there will be 2, high and low) should i leave it until after making the displacement map?

and what characteristics shall i share with the 2 LOD's? should i animate them exactly the same? anything else should worry about?

[milki]

well, i don't know the algorithm in doom3 engine
my guess is that, they made lod manual way (just reducing, collapsing, deleting.... polys)
there are few "lod" makers on net like Polygon Cruncher etc.
If you are programmer you can make one yourself

I was just trying to introduce you to the principles of LoD making.

[SIM-O]

well thanks anyway, it's become a bit more clear to me now.

[fantasia421]

Man is the artificer of his own happiness,rmt iroiro.

[SIM-O]

bugger off!

[mBrunetta]

Also, check out the Multi-Res modifier.

From the Help File:
"The MultiRes modifier reduces the memory overhead needed to render models by decreasing the number of vertices and polygons. This is useful not only within 3ds Max Design but for content creators who export models for use outside of the program, such as in Web-based 3D applications. MultiRes offers several advantages over the Optimize modifier, including faster operation and the ability to specify reduction as an exact percentage or vertex count."

[SIM-O]

will do buddy, thanks.

[SIM-O]

WOW it works great, thanks!

[mBrunetta]

Hehe, never used it myself, but it seemed to be the kind of thing you were looking for...

[milki]

But there are some limitations you need to know (this why i did not told you before about this )

Especially in characters its tricky tool to use, man.
When you do nondeforming geometry (like gun, or enviroment) there's almost no problem with it. But according to fact, thet you are modelling character, you have to think about what happens, when you start to animate this guy.
You can end up with banging your head to the wall because the geomtry is not deforming like you wanted (something will be more stretched than other etc.)
And as i can see you probably learn "how to" on this project, so my advise to you, that you make the LP or LOD by yourself.
And as experienced character modeller (worked for 2kCzeh for 1,5 years as ingame character modeller) i can tell you, that noone from character modellers in our company ever used tool like this (on characters (on nondeforming geometry we used polygon cruncher)).
The very, very big advantage in making LP by yourself is that you can get
a) better shape of model with same polycount according to automatic tool
b) same shape with less polygons

but really for you to make it manually it should be fact, that it will be good practice in making LP and modelling in general.

Hope i did not dissapointed you