I am an environment artist here at 343 Industries. I mostly work on hard surfaces like buildings and props for the human-based levels. I model, texture and sometimes place in some basic lighting.

I commonly get a rough layout from the designers. It is generally composed of squares and rectangles that indicate their needs for gameplay. I go in and give visual meaning to these forms. This is the fun part, because I can be looser and more creative. I start with a general massout, and start placing in what I would like the space to become. Sometimes, I have control in this area, other times there is already an idea of what the space is. I would then support this idea.

Then, I will break down what I have into secondary and tertiary shapes. This will be more time consuming for me because this is where I design my shapes and balance my proportions. From here, it is a matter of controlling where the player needs to go. As artists, we have a lot of control in this area. The fundamentals of focal points can be employed to great effect.

Just like in any other art discipline, starting large and working down into the details works best. It is always a good idea to have a solid foundation to work with.

It is generally a good practice to hide seams where the player will never notice them. For example, the underside of a pipe. Repeating textures can sometimes have edges in them that I can cut along.

The same principle can be used for geometry. Areas that the player may not see can have really low geometry. Any geometry that the player will never see can be deleted. If no one sees it, why render it? Also, the further something is from the player, the less geometry it can potentially have.

This can vary. Sometimes, I can use a texture that would support the detail I need, especially on flat surfaces. Other times, I will place a texture and cut into the model to bring out some surface variation and plane changes. This is cheaper than building a bunch of little objects. Objects that are in dark areas can have very little geometry. Same with far away objects. In this industry, less is definitely more. Details should be kept where the player will notice them.

We have amazingly talented artists in both our 3d and 2d departments. They can build off of what was already created and still bring something fresh to the table. From their work, guidelines are made for other artists to follow. Approved concept art can also help greatly. I like to use these as I would any other reference.

Creating the assets is usually the same process for me. Generally, people think that there should be a lower poly count, but this should be practiced in campaign environments as well. But I do try and use tiling textures as much as possible. The biggest difference for me is controlling the edges and silhouettes of objects. Since players are moving around constantly, sometimes using iron sights, then they aren’t noticing where they are going. We don’t want people getting stuck behind an object that they normally wouldn’t notice to begin with. Another difference would be the re-usability of assets. Multiplayer maps usually have higher constraints, so creating assets that can be used more widely can help.

Well, I can’t speak on this one. I haven’t had the chance to work on those races. But, from what I have seen, the same principles of level building are used. It is always a good idea to incorporate storytelling into a scene. Since we are an FPS, the toss up is when to use these details. The action is usually high paced. I notice these artists incorporating the culture of the race into the structures themselves, which is a very good choice.

With tech making larger advances, it is difficult to just take an asset and simply freshen it up. If the process isn’t too involved, or maybe it’s not a hero asset, then using these can save time for more content or more polish down the road. Or if the asset doesn’t look too dated, then we will use it as well.

I can’t comment on the remastering of ODST or Reach, but outsourcing is pretty common these days. Game projects are becoming more ambitious and that requires a lot of work. The most common challenge I see is how to get outside companies in tune with the style and quality bar. This is especially hard with sci-fi assets. I’m not directly involved in this process, however.

Absolutely. Communication is critical when making any type of game. Since I tend to get the rough layout of a level from the designers, it is important to maintain the same asset footprint, as well as metrics. If something changes, it is always a good idea to give them a heads up. Usually, discussions can happen at major milestones, like the end of a massout stage. Ideas can be thrown in or thrown out from everyone involved. Building levels is very much a collaborative effort.

From what I have seen, the biggest challenge is in the aesthetics. The principles of level building will remain the same. But the nuances of making a new culture can trip people up. Generally in sci-fi, it is hard enough to sell an environment without it seeming too modern or too utopian. I think that it is easy for the viewer to logically dismiss both. So, for me, the challenge is to use a different set of rules when designing alien worlds. It may mean going very organic, or very rigid. Combining this with good design and making it believable can be tough.

Keep creating! This is always thrown around, but it is also very true. You gain valuable artistic mileage for yourself, and you can begin to make libraries of your own assets. This way, you can spend more time telling a story rather than controlling an edge loop. No one sees the edge loop!