Level Design in Unreal Engine 3

Mario Marquardt, who’s working on Star Citizen’s single player campaign Squadron 42, is studying Unreal Engine (UE3 to be exact). While this is not the hottest tech already, Mario shared some great insights on level creation and the way you can build exploration games.

Introduction

Hi, I’m Mario Marquardt and I’m currently at Foundry 42 working on Star Citizen’s single player campaign Squadron 42. In the past I’ve worked as a level designer for Climax Studios, Ovosonico, Lionhead and Teotl Studios.

The definition of what a level designer does can differ a bit from studio to studio so I’m fortunate enough to have done a lot of scripting, blocking out levels, optimization, a bit of 3d modelling, writing documentation, prototype development… you name it.

In my spare time I’m currently working on a yet to be named UE3 project that is first and foremost a learning project. While it’s mostly to improve my 3d modelling skills I also used this project to learn some basics of animation, sound work and AI.

Seeing that it is very much focused on 3d modelling practice the project probably fits in the ‘Walking Simulator’ or ‘Exploration’ category. There is an underlying story but not a lot of proper gameplay. I’ll probably keep that for the next project.

Level Creation

Usually my approach involves a lot of planning. For most of my levels I draw floorplans or specific key locations. Mood boards are also a great way for me to focus on and understand the atmosphere and overall style of a level.

I think it’s important for a level to have a background story. No place exists ‘just because’. This approach even helps with levels that don’t necessarily have a story in the context of the game – a level for a multiplayer arena shooter will still benefit from having a backstory.

I always try to give my levels landmarks that distinguish them from other levels. If players can identify my level with very few words like ‘the level with the clocktower’ or ‘the train level’ and others know what they mean, then I’ve most likely done a good job with this.

The level should make sense in the game world. Simply put, if you’re working on a medieval fantasy game, a high-tech sci-fi robot factory will probably strike the player as out of place. It’s easy to get carried away with cool ideas but it helps to every now and then stop and ask yourself if the thing you’re building into the level is plausible in the game world.

Interior and Open Levels

When building a large open level you can very quickly run into framerate issues because of the view distance and the amount of detail in the level. LOD settings on meshes and the culling of certain elements at a distance becomes very important here but at the same time you’ll want to make sure to avoid objects in the game world suddenly popping in.

While this is also true for interiors it’s usually not as much of a struggle because interiors are more or less confined spaces and you have doors and corridors that can hide areas that are currently loading/unloading content.

Depending on the project, lighting interiors or outdoor environments can differ quite drastically, too.  For an outdoor level having just a sunlight can already get you quite far whereas for example an underground laboratory can require lots of individual lights.

Unreal Engine 4

My current project is quite art-focused and therefore a bit of an exception to this rule but generally I try not to do art and decoration until I’m happy with the gameplay and flow in a level. Eventually that approach saves a lot of time that you would otherwise spend rebuilding parts of your level, ripping out art, remodelling meshes and so on.

If you’re able to automate certain aspects, go for it! UE4 in particular allows for some very efficient techniques using the blueprint scripting system. For example you can fairly quickly script an object you place in the world that automatically places a set amount of plants in an area, or elements of a fence that follow a spline and so on. With the amount of detail in today’s games this kind of thing is getting more and more important.

Wherever possible I also try to reuse scripts or use parameters to alter the behaviour of a script to save time.

For beginners in UE4 I would definitely recommend having a close look at blueprint even if it’s just for small tools like this. It may be a bit of a learning curve but it’s worth your while.

Production of Materials and Models

I’ve got some experience in 3ds Max and Maya but at least for personal projects I now only use Blender for modelling. For texture and material work I use a combination of Photoshop and Quixel Suite.

If I’m not working with modular meshes to build a level I often export the level geometry from Unreal into Blender to use as a scale reference for any new meshes I need. I do try to keep things modular and reusable because it saves time but I still find myself creating bespoke meshes and this approach simplifies my work immensely.

For materials I try to get them into the engine as soon as possible. Previewing them in different renderers is a good way to get a feeling for what they will look like but it’s not necessarily a guarantee that it’s going to look exactly the same in your level.

Triggers

Lighting is very powerful – warm light colours will usually attract players where as cold colours can be used to communicate danger which in turn can make players avoid an area.

Having the goal clearly visible in the distance is a great way to keep the player going. Obviously, this may not be possible in every level but you can still use interesting landmarks to attract players.

The shapes you’re using in your level architecture can help, too, by directing the player’s eyes to certain points in the level.

Smooth, round shapes are usually perceived as more inviting than rough surfaces with jagged edges which seem more hostile.

Perspective

While there are AAA games that I’m looking forward to, I think I’m most excited for some smaller games by individual devs or small indie studios.

I was sceptical about VR until I got to try it and now I very much expect it to make a difference. Initially maybe not so much in the way level and games are made but the possibilities for gameplay and even for use beyond games are incredibly exciting and I’d love to work on my own VR project.

Tips and Tricks

The most important thing is to have fun. If there’s a lot of pressure to become the ‘next indie sensation’, make your own game, sell millions of copies, land that particular job, chances are you’ll end up hindering yourself.

Share your work with others and seek feedback early on, be it on forums or in person, try requesting feedback via email, join Facebook groups…

Watch people play your level. There’s a lot to take away from seeing players move through your level, seeing mistakes they make, if and where they get stuck and so on.

I’d recommend starting with building levels for existing games like Unreal Tournament or any other game that has a solid gameplay foundation that you can easily build on.

Mario Marquardt, Level Designer

INTERESTING LINKS

BUILDING “ABANDONED STATION” IN UNREAL ENGINE 4

MAKING AN AWARD WINNING SCENE IN UE4

COMPLEX INTERIOR ENVIRONMENT DESIGN