This might be a glaring hole in my self taught education, but is it ok or the norm to crash geometry? Does it hurt a portfolio piece if you do it?
I've been under the impression that models needed to be airtight. That could be because I read up a lot on 3D printing and I know its necessary there so I assumed it was across the board.
Replies
Generally speaking 3D printing and 3D modelling for games or film do have little to nothing in common besides ZBrush or other programms used. While 3D prints don't really care about topology, optimzed meshes for games or movies do very much. And while prints require airtight pieces, game and movies only care about what is seen (a very simplified statement ).
The short answer is that the method you will want to use is highly dependent on the use of the model and the goal of the project.
clean meshes are nice, but people don't really care about pretty wireframes anymore. is the art functional and technically literate within reason and the scope of the engine/project? great! chances are if you are making portfolio work that looks AAA in quality, the technical side of things will line up, so focus on making your work look as good as possible (in a realtime engine) first. If your unreal scene is running at 10fps...probably a bit of a red flag.
more importantly....does the final result look amazing and badass?? yes? but wait...OMG! an interpenetrating mesh?! not hired! - a scenario that will never happen.
On farcry 4 and watch dogs 2, all of our natural environments had a TON of meshes smashed through each other in what would be considered a very messy way. the same 2-3 rock models rotated, scaled and smashed into each other or half of the mesh sitting below the terrain and having polys that would never be seen. As long as there wasn't a major hit to performance, it was 100% fine.
open up any of epics demo scenes like the FX cave demo, its all the same stuff. meshes smashed around to create the end result, because having everything perfectly cut and welded is far too time consuming. It might have been important back on the ps1/ps2 where you were literally trying to optimize a few kb of data from the world but there is just so much leeway these days and other things you can do to quickly optimize besides mesh noodling.
I am a big fan of getting stuff done and looking good, and then tactically optimizing where needed, choosing where to invest my time once I have actually gotten something to a somewhat "alpha" state of production. sitting around and overthinking means I won't even start half the time.
@CrackRockSteady thanks for the link, a lot of good info in that thread. Is your username a Leftover Crack reference? If so, one of the worst shows I've been to.