Unreal Pipeline

[von*ferret]

For my final I decided to open up and take on a group like the teacher wanted. So my question is how can I effectively have a pipeline for the Unreal2 engine. One girl is repsonsible for mostly static meshes concering foliage and skinning other models. But myself and this guy are working in both BSP and static meshes in the actual map. So how can we more easily communicate and keep both of our stuff up to date without wasting time or making mistakes. We only have 4 weeks to complete our level so its important we dont make silly little mistakes. If I explained this roughly throw some questions at me so I can clarify.

[TomWithTheWeather]

It's important that you block out the map in BSP first so that you know it'll play decently. This way you wont have to make drastic layout changes midway through your art creation process.

To avoid art problems, one of the main things you need to communicate to the other people is the size and scale of the models you need to have made.

One trick to making your static meshes scaled correctly the FIRST time is to actually build a very rough, yet correctly scaled/sized version out of BSP. Say for instance you wanted to make a house. Just build that house out of simple brushes, not even worring about texturing it, and play with it's size and location until you get it how you want it. To get scale right, have a "standard" smesh that is the same height as the player. I've seen guys who didn't do this and when they finished that super awsome looking doorway model and imported it, it was 3 times the size of the actually doorway.

When you're happy with your temp BSP house, select the brushes that it's made from, duplicate them, drag the new brushes up into the sky or at least somewhere where you have plenty of space around them, and rebuild. If you've gotten this far, you should now have a second house floating way up in the sky somewhere. Now take your red builder brush and completely surround the floating house and intersect it. Now your builder brush should be shaped like house.

Now you can export your house shaped brush into a format that you can then import into max and model from there. You probably wouldn't want to actually use that crappy imported model as a base for your nice house, but you can have it there for a side-by-side comparison of the nicer house you make from scratch.

Shoot, I've seen maps where the mapper blocked out the whole thing in BSP and then imported that complete BSP shell into max and created his smeshes to perfectly fit all the walls and such. As a result, his map was super clean and the vertex lighting actually looked very decent because he didn't have random mesh verts imbeded into the BSP.

[Kosmo]

One of the most important things is to have a firm concept of what kind of map you are making, once THAT is ready then make the changes accordinly, don't go for the direction of solo mappers and do changes as you progress on the map, it's not good when someone is making a staticmesh and you just say that you scrapped the idea for it.

Other thing is to define the overall looks/colors/fell and keep them unified, one of the biggest problems I have seen ever in teamwork is when the members of the team work on their own styles and then the final piece don't work when they are put together, and this is usually done on the end of the project (which means so time to fix your mistakes, bye bye project).

So I guess all I'm saying is, that make designs for the map, and give names to key areas, and stick to them, it is way easier to communicate when everybody know what thay are talking about, example. "What do you think about the corner on the left side of the map, you know where there is a long tunnel before the big drop?" vs. "What do you think about the corner on the 'twilight zone'?" yes I know, my names for certain areas are very very weird, but make up your own.

[Kedhrin]

yeah you guys definetly need a clear overhead.

for example, draw a layout of the map (top down) then mark numbers on it where key architecture features are and special events and areas.

for example

point 1- In this area it is a cylinderic area with stairs going down both sides of the wall to form a platform at the top. The stairs and the platform should go along the wall to form a half circle form. in the center of the half circle (between the base of the stairs on both sides) is a pillar with 4 head statues on each side of the pillar. look at for concept. across from the cylinder room is 2 tunnels comming in that lead to other areas. Since we are doing a stone/ancient complex make sure it smooths into the area with loose bricks and destroyed rubble. At the top of the platform is a spawn point for the flame thrower with some ammo for it lining the walls behind the spawn point.

basically, if you guys had everything planned out for the map

ie

Map type - DM

our general feel - stone ancient medieval castle theme

layout flow - double loop with a centeral point of combat in the center of where the loops meet.

etc.

Concept sketchings

basically complete a solid form of the map before you begin

and i mean SOLID a good thing to do to make sure everyone is on the same page is to create a little history of the place you are making. For example "This place was used by ancient Skarj as a battle proving ground before ascending in rank throughout their tribe. THe place is long and abandoned with the increases in technology the skarj are using now. Left alone for a long time the complex has succomed to weather and decay." - etc.

Once you have this down, and you guys thuroughly decide what you want to do (and be realistic about your deadline as well) you go at the map full swing everyone knowing what has to be made.

for the bsp creation, lets say this map is a double loop and the loops connect at one center point.

Since it is your project i'd take Center point and loop side 1 and give your other guy loop 2. For your prefab person i would make a sketch of every object you want custom made and tell how it will interact with the world in the design document.

hmm

i'll think of more to do later

[von*ferret]

I've been working on the BSP brushes for a while now and have all the geometry for the most part exported into max for my team mates to start work on. Problem is I dont have ALL the areas fully blocked out and we've started creating static meshes. Forinstance I'm responsible for aproximately 80% of the actual level (yeah I have no life) which is mostly consisting of outdoor environments with little indoor connections. My team mate is responsible for working inside a large building which will act as the major "indoor scene." That area has one large brush that blocks it out, but isn't hollowed or anything. Should i Have him hold off as I get the layout in there set up? My third team mate is just working on generic foiliage meshes so I'm not really worrying about her at the moment.

My main concern is when we merge our work together we get errors or forget things because I can't handle wasting time on stupid little mistakes.

[Izuno]

This is a classic pipeline problem as you've so clearly communicated. Tom, Kosmo and Kedhrin all gave great advice and I would do what you can to follow it.

So it sounds like the critical issue is time. So here's a short primer on project management that is sort of basic that any good designer/producer will already know....but I mention here for anyone completely new to this stuff who happens to be reading. In any project, there are 3 major factors to balance: scope, resources and deadline. If your deadline is fixed, then you'll have to make compromises on your scope or resources (meaning "less scope" or add more resources/people). If your resources are fixed, then you have people work longer hours but that can often backfire on scope (a.k.a quality...read up on the EA labor controversy for some example stories).

So back to von*ferret's specific problem:

Your issue is making sure that your teammates always have something productive to do and aren't wasting time on work that will have to get thrown away because sizing/scaling is not accurate. If that's so much of an issue, you try exporting bits of your BSP to max for them to build the required models as you go, rather than have to wait for the entire BSP to be finished. That way they can get working on the meshes and get the sizing/geometry nailed and move on...handing the mesh to the skinner.

For that major indoor scene, are there any models that you can have leeway on size? Say there is some statue or light fixture whose size may not matter if it's just a bit off? If so, have them work on that like the foliage as you discussed. (I'm sure you've already come to that conclusion)

In general, do you have a priority list of what models that absolutely must be sized exactly? If it is not possible to get proper BSP layout to them for sizing, have them work on other things that will need to get done eventually.

Anyway, I'm not sure I can suggest much more than that, as I'm not very experienced in unreal2 engine. One thing to keep in mind, if nailing down the mesh sizes is a critical bottleneck, then you should immediately test them asap once a mesh is complete so that your mesher knows if they are correct. The faster turnaround on that part of testing will enable faster corrections and keep your mesher working on getting it to fit and waste as little time as possible. (Again you probably know that)

4 weeks is not a lot of time. I suggest your work your ass off this week just as long as you don't burn out or burn out your team. The more time you spend now the easier the final week will be.

Good luck and keep us posted. Listen to the other guys here, they gave really good advice.

Additional:

I think this is a "bite the bullet" type of problem. You're just going to have to suck it up and test meshes as you go...but to enable that you must hav those portions of the BSP they're fitting into completed. That's a bitch of a process given the short deadline, but you may not have any other choice, assuming the meshing/skinning is so time consuming and corrections so costly to schedule.

Can someone make a "bite the bullet" emoticon?

[kleinluka]

I am pretty sure Unreal 2 Ed allows you to scale a mesh inside the editor (unlike Hammer) so if something is out of scale, you simply scale it right in the engine.

[von*ferret]

I've exported the whole BSP brushes into .obj->.max formats to scale for the meshes to be created off of so thats not a problem scaling wise. Only thing I can imagine us scaling would be foiliage.

[Eliot]

I am pretty sure Unreal 2 Ed allows you to scale a mesh inside the editor (unlike Hammer) so if something is out of scale, you simply scale it right in the engine.

sure ued can

[kleinluka]

this thread = 4 months old

[Eliot]

lol didnt know wel i knewed it was abit old but not that -.-