Biting off more than I can chew.
polycounter lvl 6
Hi forum, my last few threads have been a bit varied asking for help. And I think I've just gotten carried away. So mainly what I'm trying to figure out is a good solid workflow.
Id search it here however I'm not really sure what to look for. Anyone have some tips? I mainly want to make game ready assets, and texture them.
I know a fair bit now about modeling in 3DS max. So here's what I think is the order of operations. Feel free to correct me. Object is a standard non moving prop, as an example.
Get concept/reference -> Model high poly -> convert/remodel to low poly ->smoothing groups -> UV unwrap -> bake textures -> paint textures -> done.
Would love to hear what techniques you guys/and gals have to share. And thanks!
Id search it here however I'm not really sure what to look for. Anyone have some tips? I mainly want to make game ready assets, and texture them.
I know a fair bit now about modeling in 3DS max. So here's what I think is the order of operations. Feel free to correct me. Object is a standard non moving prop, as an example.
Get concept/reference -> Model high poly -> convert/remodel to low poly ->smoothing groups -> UV unwrap -> bake textures -> paint textures -> done.
Would love to hear what techniques you guys/and gals have to share. And thanks!
Replies
Ok awesome, good to know im sort of on the right track. Im making stuff to learn so I dont really have any set direction or guidelines. Is that something I should have?
That's the gist. Surely, it's not as easy as it sounds and there are tons of issues that you need to be aware of when making an asset. It all comes with experience, really.
And I should base the uv's off the smoothing groups? >.>
Pretty much. I made a small poster on my wall with this rule.
If you have a smoothing split you must have a UV split and padding, the inverse is not true for UV splits.
This is how the handplane developer described it any way and it's worked for me in my older projects. Either way it's a good rule of thumb.
On the other hand, Epic Games state in some of their documentation that their method is to use 1 smoothing group on the whole mesh and bake. It's quicker but you can end up with some tangent gradients that might cause artefacts, it's not incredibly likely though.
Honestly if it's not some mesh that's meant to be used often I'd go with Epic Games method. If it's causing issues, then I'd switch to handplane's method.
I'm not entirely sure if this is what they do in the industry, Epic probably have methods of reducing the artefacts without resorting to re-doing the unwrapping. Probably by adding edge loops etc. to control the vertex normals more. Earthquake has a fantastic post about skewed bakes in 3ds Max that touches on the issues.
I think you will be interested in reading my A Practical Guide On Normal Mapping For Games, you can download it in my signature. It covers pretty much everything you need to know about normal maps.
Actually I have seen it in my vast search for doing normal maps correctly. Very long though so I dont remember if I finished it. >.>
Would really love to see from start to finish how someone models a game ready asset. Like the ENTIRE process. Many digital tutors videos Ive watched they leave a bunch of stuff out to shorten them. and its always the stuff I dont know how to do. >.>
As far as smoothing groups/hard edges go, you'll hear a lot of different things from a lot of different places. It really all depends on what you are doing and there are 101 million different factors to which is right for you, but fortunately, unless you are doing work for a studio(in which case they would most likely provide bake specifications) you really only have to worry about what looks best.
Just as long as you aren't getting into the practice of painting out baking errors, then you're okie dokie artichokie.
I've heard it's normal to do that in studios when stretched for time. Obviously try and avoid it as much as possible but it's quicker to spend 5 minutes painting something than it is 5 minutes fixing the geo and then go rebake etc. 5 minutes would be real fast too.
My tutor has talked about artists having to do that a few times at his old studio. Honestly I can't imagine being an artist at crunch, I would hope priority assets would have been finished way before deadline, I guess it's fine to use a rushed approach to some of the smaller stuff. Even so, when quality gets sacrificed for quantity you know you're in trouble haha.
Any ways, my point is I guess it's not so bad to practise painting out an error here or there, but if you're doing it for every object you need to revisit your base skills. Anything for a portfolio? Make sure that bake comes out great.
Alright awesome Ill check that out too. Thanks everyone
Often, you'll get a better looking result if you bevel the harsh edge instead of putting a smooth group on it.
This adds polygons, ofcourse, but it's not as bad as it would seem. Really, the thing that matters for game engines is vertices, not polygons. When you put a smooth group on an edge, it actually doubles the vertices along this edge. At that point, you might as well bevel; you'll end up with the same vertex count.
Not to mention, polygon budgets have gotten so steep in recent years that usually you have enough to bevel as much as you want, anyway!
Ok I suppose I can try that. Read a bit of that in my research. So the only engine Ive any intrest in is Unreal engine 4. I dont know the first thing about it yet but thats what Id like to try my stuff in. Is there a general set of do's and dont's for that engine?