Building Environments For AAA Games

I was lucky to meet Martin Teichmann at Gnomon this year, where he gave a wonderful talk about environment design. In our exclusive interview, we’ve talked about the main tasks of environment modelers in game development, discussed readability, use of color and optimization of 3d scenes.

Introduction

I am an Environment Modeler from Germany working at Naughty Dog since 2015. The last project I shipped was Uncharted 4. I started my career around 2007 working for smaller studios. 2010 I got the chance to join Crytek in Frankfurt Germany where I was working on Crysis 2 and Crysis 3. Later I took a job offer in London at Rocksteady and worked for more than 2 years on Batman: Arkham Knight.

You’ve been doing environment art for a while now. What do you believe are the biggest tasks of the environment artist? 

The main task is to create beautiful art, add meaning to the scene and enhance readability. That’s the obvious answer. But there is more to that. As an environment artist you are responsible for all sorts of gameplay mechanics too. Collision for example. You need to maintain all the technical elements that make sure the level plays as smooth as possible, even though you have to sacrifice your art in certain cases.

To tell a story is very important. Of course, color and composition is key for a great looking scene. On top of that I always love to create little story moments inside my environments. Who was living here? What is the purpose of that location? Did someone leave just before you entered this room? Why did he leave? Is he coming back? All sorts of questions I love to answer for my art and add hints and details to give the player an idea of that space.

How do you turn a simple block out into a comprehensive part of the game world? How does that process work?

The first step is always reference and concept art. Usually I start my modelling process by gathering reference material or get familiar with the ones provided by art direction. The second step is to look at the block mesh provided by the design team. They create the gameplay spaces before an artist is working on the level. The actual environment art job is quite straight forward. I work from the bigger to the smaller shapes. Usually I bring one or two areas to almost final quality. So I have an area representing how the level could look like when it’s done and to learn what I need to do to make it look good everywhere. Overall there is a lot of iteration and feedback involved. The game designer keeps iterating the gameplay and I get feedback from colleagues, leads and directors. Very important to that process is the concept department. They support me during the creation of the art. They provide overpaints, give ideas and also make sure we achieve a consistent look throughout the game.

Your environments look incredibly interesting, just begging to be explored. How do you manage to convey this feeling into your scenes? Is there some particular technique, how you test your scenes? When do you decide if it’s interesting and compelling enough for the player?

If you ask me, or really any artist, it’s never good enough and never finished. But release dates dictate a limited time frame. Usually, I am working on the whole level building it up gradually, adding details equally and make sure all corners are up to quality. If I don’t do so, the chances certain areas are not good enough are higher. For the overall picture I like to literally step back from the scene or squeeze my eyes to get the macro picture of the scene. Thumbnails are great for that too. If the thumbnail looks great the overall scene usually looks good too. It’s very easy to get into the micro detail and loose the overall picture and look.

Color and light are always big parts of any environment production. How do you use them in your scenes? What function do they serve in game environment creation?

In most companies there are specialized lighting artists doing exclusively lighting for the game. It is a very important part of art production, important enough to have an entire specialized team for that. With bad lighting it is very easy to ruin the best art. And with awesome lighting you can even fix average looking environments. Important for good lighting is it serves the gameplay as well as the artistic aspects of the game. I personally like to play around with shadows to create moody environments. Also I try to bring out the shapes and details using light. Color of course plays an important role too. It’s all about the mood and character you want to create for your environment.

Could you talk about the readability of the environment? How do you make scenes understandable to the player? How should he understand that this wall is claimable and this place is safe for example?

Color and lighting are very powerful tools to guide the player in your environment. It’s very easy to get lost in details and lose the overall picture for your scene. For example: it’s important to separate the floor and the walls visually to make sure the player can read an environment easily. Also gameplay elements need to be visible even from distances. Color codes can be very helpful in that case. For Uncharted 4 we had the rule that all the ledge grabs had to have a white line to make sure the player would know where to go. But said that: consistency is key. Ideally you want to create a set of rules the player can learn and use throughout the entire game. It can be painful sometimes as it can interfere with your artwork but good playability should always be first.

How do you work with optimization? How do you manage to make such rich environments, which are filled with so many beautiful assets and still keep it relatively cheap for the CPU? Maybe you could talk about clever ways you reuse the textures, assets, various vegetation and so on?

I try to stick to a set of assets for my levels. It’s easy to put all the cool stuff you find in one level totally ignoring that all the objects add on top of your memory limits. But it’s really more a feeling how much you can effort. Some levels can handle more than others. Depending on how well the streaming works in the area you work on. It’s always a trade of. If you have too many draw calls in your area you could combine geometry but that might cause a problem with your triangle count as you don’t have the LODs anymore. The main point I work with in terms of optimization is: “Does that detail add to the look here?” I try to add geometry and assets very cautiously. Nice details in the silhouette usually pay off for example. It can help to reduce draw distance in certain cases. But again it is a trade of: Visual fidelity against framerate.

Martin Teichmann, Environment Modeler at Naughty Dog

Interview conducted by Kirill Tokarev. 

Follow 80.lv on Facebook, Twitter and Instagram