Normal Maps versus Low Res, Which first revisited.
polycounter lvl 18
I was thinking that maybe Low Res might be better to begin with if nothing else because you can get it in the game to see what does and does not work.
Versus if you start with the high, then make the low and find problem areas with the game engine. You then have to go back and add/subtract from the high, rexport and fix.
Seems a more streamlined process? Am I missing something other than a anal art director?
Versus if you start with the high, then make the low and find problem areas with the game engine. You then have to go back and add/subtract from the high, rexport and fix.
Seems a more streamlined process? Am I missing something other than a anal art director?
Replies
Of course if it's a brand new engine and you need to test out what works/doesn't work with it, then yes, building rough models to test would be a good idea... but usually the process is fairly well documented and predictable, I'd think.
i personally start with a midpoly mesh.. because i end up subdividing for the high anyway and i can can just subtract a little bit to get the low and reline up the mesh a bit.. all in all not much work at all..
as far as there being problems with the model in game after you make it.. unless your working in the dark ages there should be no problem.. most companies have what you see is what you get tools.. so what it looks like in the max viewport is how it will look in game.. at mythic we had a max plugin that actually started up the engine and you could check the shaders and animations and everything in engine before export.. most companies have tools like this.. there is no guess work anymore..
i personally start with a midpoly mesh.. because i end up subdividing for the high anyway and i can can just subtract a little bit to get the low and reline up the mesh a bit.. all in all not much work at all..
[/ QUOTE ]
Im going to second this, i usually start with the midpoly mesh and go up or down from there, it's just easier (personally) to be able to add detail where you need it and not worry about the constraints of a lo poly mesh.
max plugin that actually started up the engine and you could check the shaders and animations and everything in engine before export.
[/ QUOTE ]After export, before save, I think you mean.
Did they have it running in a Max viewport, or as a separate process? I'm curious because we're looking into the former.
Here's the thread link, maybe there's some information you can use there:
http://www.ogre3d.org/phpBB2/viewtopic.php?t=11520&postdays=0&postorder=asc&start=0
[ QUOTE ]
max plugin that actually started up the engine and you could check the shaders and animations and everything in engine before export.
[/ QUOTE ]After export, before save, I think you mean.
Did they have it running in a Max viewport, or as a separate process? I'm curious because we're looking into the former.
[/ QUOTE ]
no you didnt export anything.. you just had the maxfile open hit the button and it would open up the scene in a seperate engine window, with lights and particle effects or anythign else you had in the max scene at the time.. from there you could move around the file play the animations or what ever.. then you would export .. and that was a not so easy really long and contrived horrible process..
arshlevon, guess I'm being a bit overzealous but something has to be exported to make it run in another app.
I'm thinking the reason for the convoluted post-export is because things have to be compressed/trimmed down/swizzled into a game-friendly format, which may require artist decision-making. But IMO leaving it in a convoluted state is a bad management decision... though maybe they've seen the light since you've been gone.
on PS2, Xbox, etc. There is in fact an export, though. Much like Eric said, it needs to be exported somehow, in order for it to open in a seperate app.
The new engine we're getting though, has a similar feature to the Ogre3D engine that MoP posted. One of the viewport windows in 3ds max is actually the game engine, rendering the scene in real-time. The whole purpose of that of course, is WYSIWYG editing of your assests. It's really quite interesting, and I can't wait to get a hole of it
s far as there being problems with the model in game after you make it.. unless your working in the dark ages there should be no problem.. most companies have what you see is what you get tools.. so what it looks like in the max viewport is how it will look in game.
[/ QUOTE ]
Hmm.. Well maybe it depends on the circumstances. A character a High poly would work fine because a character is one mesh. However, a vehicles with alot of attachements and joints could end up having to be "rethought" engineering wise a few times after putting in engine (as has happened to me). It would make no sense to make a nice looking high res version, then animate the low version or other and find out you need to remodel/rethink areas the concept artists sketched.
Plus the amount of optimizations sometimes needed gets rid of protusions. Since these just look funny on a normal map, you then have to get rid of these on the normal mapped version and re-export.
There can and will be problems either way here and there.
Once you get used to the process, correctly guessing how something will translate into a normalmap is as simple as modeling itself, so, there isnt really much to think about. Whatever serves you better.
I always d te thing I find hardest to get it out of the way