Need some input for level design

hi guy's quick question, how do you design and prep your levels? would you model the scene in a 3d editor like 3d max or would you break it down into segments? i am a 3d artist but i dont often make my own scenes so i dont know what the best approach is.

for example if i make a indoor game level like a hospital, would you model the whole base environment (hallways,rooms) into one model? or would you break it into segments kind of like fps creator used to do.

thnx

 

I have been thinking about the same thing. I need a city. I have low poly models, but cannot draw to save my life! I was thinking maybe I build the city block by block in blender(I can still animate the windows and doors) then import each block into unity. I think having bigger "chunks" will be more performant when they are culled when out of view. But, I haven't gotten that far into my project yet. I'm still prototyping classes and working out on paper the client/server fun stuff. Since I am HORRIBLE at world design and all, I am kind of putting all that off until I HAVE to do it. When ever I get bored of flying around primitives, or driving around my terrain painted world. I am pretty sure I will want the race tracks as 1 full object that I can just plop down in designated area. Same with the airports and shipyard. I just lack the skills to lay it out in unity. If it was as easy as doing a finite state machine for AI, I'd be done by now! lol

 

I like to create small, modular objects in the 3D modeler (blender in my case) and assemble them in the Unity editor. For me, level editing is always an iterative process of testing and adjusting, so I'm constantly changing things at a broad scope. It's much easier to be able to do that directly in the editor.

 

For stuff like race tracks why not make a couple of base graphics and procedurally build a race track by defining a 3D curve?, kinda like Cities skylines does it. this requires a custom editor tool

maybe even a roller coaster tycoon style build tool, they actually have a racetrack there.

 

I also break my levels up into a few rooms each. I also separate places where the player and go and places that are further away (so that the lightmap resolution can be far lower for stuff in the distance.) For pieces that are especially intricate, I may have two models, one highly detailed one which doesn’t have any collision and a low poly version which is invisible and used only for its collider.

I use trigger volumes to load and unload areas in a script as the player moves.

 

Another advantage of making modular parts is that you can make several different buildings and other features using the same small set of pieces. So it can reduce your overall modelling time because you can reuse all of the elements but still create unique buildings. You can see countless examples of this on the AssetStore. This is just a random, simple one:

https://assetstore.unity.com/packages/3d/environments/roman-city-low-poly-pack-1-92702

In the image, you can see first the pieces included with the kit and then also examples of buildings that are made from these same pieces.

 

This is a interesting question.

To answer your question, usually in most cases,.....yes.
Stuff that don't move, eg. The entire level, floor, roof, rooms in single level, hallways, is usually one whole model.

Stuff that you want to move in the game, or stuff that can move, like tables, chairs, hospital beds, ceiling fans, are usually seperate props.

However:
If things like tables, hospital beds, bins etc. are not made to be moveable, you can still make them, as seperate props from the enviroment.

Another important thing to consider, is that according to the type of game you're making, and what you want to do in the game, that can also affect, how you go about modelling stuff too.

 

From my failures at trying to make an Open World Map (still failing) look right, i would say only this: WHY it is sooo difficult to make Level Design! lol

 

When I am on the can and I feel like drawing/doodling a theme and a level, I just make a general feeling to it and then, depending on the genre, I made it as separate objects coalescing together. I put them together in my game engine and have a solid base. Then, I add decorations.

Here's a thorn maze:


Here's an ice cavern:

 

It really depends on the overall game design. I realize the thread is old btw, just waiting on loading bars so I'm gonna bs a bit.

For open/natural world, I like to just throw down something and then start trying to find gameplay in it. Then I might stay pretty close to the original blotch or maybe I rework it dozens of times as I develop the gameplay.

For like, pretty much any other type of level design you probably wanna look up the term "modular environment design."

If you watch most enviro 101 tutorials you find artist usually work from a concept, they build a cache of the modular items they'll need to construct the level, then they place it all. As with most things, it makes sense to do a dry-run first. That is, make crude fast versions of your art first to block in the big ideas, then when you are certain you put the time to polish your work.

There is two big hurdles to figure out before you drop time into production. One is workflow. You need an assembly line setup so you can get your work done efficiently. Second is having clear design. You solve that by prototyping/greyboxing whatevere you wanna call it.

So once I know how I'm gonna do the work and what the work to be done is, then I'll actually make the proper art and set it up in engine. Overall, how you get started is probably a matter of user preference. I liek to toss paint onto canvas and let imagination flow from there. But of course you have some vague overarching goals so it's not like you totally winging it.


cool ice level @Volcanicus. needs some fancy refractive shaders!

 

@BIGTIMEMASTER thx. that is usually for the end game

Also, I want to mention that modular is best if you need to play test the nav-mesh for instance and need to change stuff:


In this point and click adventure I made, I had to move the couches several times to get the correct pathing i needed.

 

Looks Awesome!

 

It's not that old.