Best Practice / Suggestions for Laying out Many Non-Axis-Aligned Walls?

This is a simple question, but it requires a little bit of setup…

Say, I wanted to create a level with a lot of walls, but ones that aren’t axis-aligned. Take this first level from the original Deus Ex for example:

There are lots of walls, and at least half of them are not axis aligned.

Epic suggests (in their videos) to use BSP brushes to do initial level layout, then later export, model it externally for real, and then suck it back into the editor for final placement. This sounds great, but the problem is that BSP brushes apparently are very clunky and limited. The main problem in this example, is that they handle rotations poorly.

Say I create a simple chunk of a wall, and then rotate it. After the rotation, the transformation gizmo is still world-axis-aligned (and toggling to the local one does nothing), which makes it dang near impossible to position it correctly.


Since using BSP brushes to layout these walls is too painful of an option I’m left with the question of what I should do. I mainly see two options (with my related concerns):

[LIST=A]

Create a small chunk of wall as a static mesh (since I can easily rotate and position), and use many of them to create my long sections of wall?

  1. I’d have to design the chunk perfectly so that it effectively would seamlessly tile with adjacent chunks.
  2. Lots of extra geometry. (One long linear stretch of wall needs [at minimum] 8 vertices, not 8 times the number of chunks!) Perhaps there’s a way to merge them later…?

Model the entire level externally and import as one giant static mesh.

  1. Exporting and re-importing over and over is not anywhere near efficient.

[/LIST]

So what can I do? Option A seems like the only real choice, but I’m new to all of this, so I’m hoping there’s a better solution I haven’t seen yet.

Perhaps using a Blueprint spine is another option? Overkill? I’m not sure how that’d handle literal corner cases though.

Maybe Epic should fix the local space for BSPs so the widgets become useful again?

-Will

One can hope. I haven’t had any responses to my post on that, and I can’t determine the level of “priority” a brush overhaul has there. I’ve heard they’re “thinking about it” but that’s all.