Designing Concepts & Props For Horizon Forbidden West Burning Shores DLC

Introduction

Hey, my name is Adriaan Stam, I’m a Concept Artist from the Netherlands working at Guerrilla as part of the building blocks and environment concept team. Most of my experience is in designing props, but recently, I’ve been doing more environment art as well. I’ve worked on the games Horizon Zero Dawn and Horizon Forbidden West, as well as their respective DLCs – The Frozen Wilds and, most recently, Burning Shores.

Becoming a Concept Artist

I remember always having a desire to create or visualize stuff that captured my imagination, whether it was through drawing, with clay, or LEGO or crafting things with bits of wood or paper. When I was younger, I drew a lot and would go through periods where I was into different, mostly historical or technical subjects.

The first obsession, which I can remember from when I was very little, was trains, then I got into ancient Egypt, the Middle Ages, and so on. Usually, those interests would be fueled by games I was playing at the time. Age of Empires was definitely my favorite and was one of the main reasons I became interested in history. Especially Age of Empires 2, where I spent countless hours playing on my own and with friends, and still do occasionally. Of course, those games also inspired many drawings. Here are some I dug up, including ‘fan art’ of buildings and castles from AoE and Stronghold:

I was also interested in the practical side or the inner workings of things. I’d read books on how medieval people built castles and cathedrals or would try to make a functioning trebuchet from LEGO. I think that a combination of interests, culture, and history, plus technology is a good fit for concept art and prop design in particular. Learning how and why people made things the way they did can help and inspire a lot when designing props. You’re essentially figuring out the same problems but within a fictional world.

I got to concept art in kind of a roundabout way. During high school, I’d stopped drawing but got into programming, and afterward went to college to continue in that direction. This turned out to not be a great fit, and instead, I spent more time active in the community for a Battlefield 2 mod called Project Reality, which I played a lot back then.

I got to know some of the people who were actually involved with the development of the mod, and they encouraged me to learn 3D modeling, something I’d been interested in for a while. Eventually, I ended up joining the development team, creating 3D assets for guns and doing promotional material. It was a fantastic experience, the team was full of extremely kind and talented people who taught me much about game art. Since it was a group from all across the world communication went through an online forum, but it already gave a taste of what it could be like to work in game development.

Trying to model weapons accurately to their real-life counterparts taught me to observe critically and develop a sense for volumes, but I did begin to miss doing more imaginative stuff. Around then, I watched Blade Runner and Ghost in the Shell for the first time and was playing the Metal Gear Solid series a lot. Being more aware of game and film production at this point I looked into the production design of those titles and found legendary artists like Syd Mead, Mitsuo Iso, and Yoji Shinkawa, among many others. Inspired by this, I started drawing again. What convinced me I might actually be able to do concept art professionally was coming across the FZD School of Design YouTube channel. Their videos clearly broke down not just how to make a pretty image but a good design and made it seem achievable even for a beginner.

I quit college and spent about half a year preparing a portfolio to get into the University of the Arts Utrecht, a Dutch school that offers game art courses. I felt more at home there and found people with similar interests. The classes were good but were more intended at giving you a broad skill set as a game artist, so I practiced a lot outside of school with friends as well. Finding a group of like-minded people and pushing each other was invaluable and definitely one of the best things to get out of it all.

Joining the Guerrilla Team

I first joined Guerrilla as an intern during my third year at school. It was my first experience of working in a studio and I was blown away seeing the quality and volume of art when I got there. The first couple of days, I was just getting familiar with the IP, looking through concept art, and reading style guides for all the different tribes. They were so extensive I remember thinking you could make an entire movie based on just one of those documents.

Making a portfolio for the internship was tough, and I wasn’t really confident I could make it. I doubted if my mechanical design concepts would be a good fit as Guerrilla had recently released their first trailer for Horizon, and my work was quite different from its tribal cultures or sleek robotic animals. In the end, it turned out alright, I think in part due to showing an effort towards rendering realistic material expression similar in terms of technique to what we’d be doing when making props for Horizon.

Along with the portfolio I sent a blog I had kept with sketches and writing. I’d seen someone else do this before and thought it was a great way to show more process without cluttering the actual portfolio. For the next half year, I had a blast designing props for Horizon Zero Dawn. I learned a lot during this time and it really helped shape my thinking. It was also very fulfilling to contribute to a real game and feel all that practice finally paid off.

But after the internship, I went back to school again to graduate. While we were busy with our graduation projects, me and my friend Tom Kolbeek (who was at Guerrilla doing his internship around the same time as me) got an email asking if we wanted to come back to the studio for a couple of months to help out with designing props for the Frozen Wilds DLC. We both decided to take the opportunity. It was an intense few months but well worth it. In the end, it led to both of us being taken on as Junior Concept Artists to work on Horizon Forbidden West.

Concept Art Pipeline Behind the Burning Shores DLC

After finishing Forbidden West there was initially just an outline of the story for Burning Shores, so we had a lot of freedom to do exploration for a while. We knew two main things: the locale was going to be a ruined and overgrown Los Angeles and the main tribe for this area would be the Quen.

The Quen are a seafaring people and kind of a unique case compared to other tribes in the game. They originate not from the Americas but from somewhere across the Pacific Ocean and sailed to San Francisco on a large expedition. A devastating storm caused their fleet to be split in two, however, with only a small part making it to its goal. The larger part gets shipwrecked against the flooded ruins of L.A. These are the eponymous Burning Shores, so-called for the lava flows there caused by centuries of tectonic and volcanic activity.

There were Quen designs to go off from the base game, but since they would be more prominent this time we now had to define their presence in the world on a bigger scale. Part of this effort was coming up with ideas for their encampment post-shipwreck. Since they would need protection from the elements as well as machines and potentially hostile tribes I wanted the location to look sheltered, like a cove. Initially they would have limited resources to build with so their best bet for safety would probably be to stay on top of and directly around the large ruined structures. 

Quen society is heavily influenced by digital information that their diviners (priests or scribes) gained access to from pre-apocalypse megacorporations through Old World devices they found. This knowledge made them more organized and technologically savvy than other tribes but also very hierarchical. It seemed logical that admirals, diviners, and others of the leading class would live above where it’s safest and they have a good overview of the surroundings. Soldiers and crew on the other hand live and work below and up against the high ground. Perhaps they could retreat to safety at the top if things got dangerous on the ground, kind of like a motte-and-bailey castle situation

Building on and with earlier work done by others in the team I did the overview shot above for the settlement trying to show these ideas. The base image was an early work-in-progress scene by a teammate who did a lot of the visual development work for the settlement. I basically rearranged parts of his early renders in Photoshop before painting and photobashing on top. 

Additionally, I looked at the shipwrecks and how the Quen might use them. The first idea was that they would just salvage the wood, reclaiming beams, planks, or whole chunks to use in building the nearby settlement. I worked on an image of a salvaging site showing a wreck surrounded by scaffolding and debris. I blocked it out in Blender using a mix of custom geometry and imported assets from the game to create the wreck and scaffolding. I knew I wanted to do an angle from inland looking out towards the beached hull so I mostly focused on detailing the visible side of the ship and the area in between as well as a few ruins visible in the background. Below are a few renders of the scene as I was working on it:

I tried to put in storytelling elements like wood processing areas, support beams to help prevent the wreck from collapsing or being carried away by the tide, and ramps built to make transporting material from up high easier. Unfortunately, it ended up almost making it look more like they were constructing a new ship instead of tearing it down. 

As the writing progressed, that actually turned out to be the direction the story was going so by luck it kind of worked out. To get the idea of it being a salvaging operation across properly though a less organized look would’ve probably communicated the idea better, without the cranes and scaffolding. At this point, we were starting to move out of exploration, and other work needed to be done. I did make a few sketches illustrating ideas for a construction site and a quick plan view of what the area layout could be like, roughly. Based on those, the earlier images, and some communication back and forth with the environment team, they could build the final in-game location.

Setting Up the Props

Moving into production there were a number of props and building blocks to design that the environment team identified and blocked out based on our concepts as well as their own exploration. Our goal then was to brief those props for the asset creation teams, which included making concepts rendered or in linework with base colors depending on if the asset and materials were new or complex or not, and adding any visual reference material useful to the asset artists.

The ones I made included a few sets of old wooden beams and framework structures that needed retexturing to match Quen wood types. Here my goal was mainly to get them done efficiently with clearly rendered materials. I used the existing 3D assets to create base renders and then photobashed and painted over top for most of the material definition. Here is an example from one of those sets:

We use 3D all the time when doing prop concepts but it depends on the prop and artist to what extent. I like to use at least a base render of the speedmodel (blockout models from the environment team or if not available ones I make myself) with a basic material to make sure I get the proportions right and a decent ballpark for the lighting response. I went a little more in detail with the 3D of some of the more involved building blocks, working out their construction (which was fun to explain in true IKEA fashion) like with these ship beams:

In the end, however, most of the effort with these beams went into 2D, getting the right wood surface definition and wear/damage. The result left me a little unsatisfied; they were fine for their purpose but the rendering was a bit muddy here and there. I had recently done a personal project using mostly 3D throughout and wanted to apply a similar workflow to a prop to try and get a more crisp, high-detail render out of it, so that became my goal later on.

When designing props for Horizon we try to let the culture they originate from show through in everything from material usage to construction technique and decoration. Here too the Quen are different from other tribes in the game. Their material usage for instance was not dictated by the environment we encounter them in. Looking at the Nora tribe they’re pretty much locked into their part of the Rockies which directly provides them with the lumber and stone they use to construct their dwellings. The Quen’s ships and most of their equipment however are built back in their homeland and brought along to a completely different environment. 

The Quen’s design language wasn’t as well defined and documented as some of the other tribes in Forbidden West since they were a relatively small part of the game. Luckily there was enough there already to extract guidelines from. In a nutshell, those came down to something like this: the Quen use a great variety of materials showing their advanced level of technology. They have skilled metallurgists, particularly using alloys like bronze and brass. They rarely use robot parts unless specifically desired for functional or decorative purposes.

As expected from a seafaring tribe they’re excellent woodworkers and ropemakers and are capable of creating strong and intricate joinery from just these materials. Additionally, they can create refined glazed ceramics and embroidered fabrics. Depending on the use, objects will be highly decorated. Items such as tools may be less decorated but should still look sophisticated in their construction. As sailors, the Quen take much inspiration from the sea. They’ll use waves and foam-like patterns in carving, fabrics, and painting, and even materials such as mother of pearl and coral for more fancy items. These kinds of decorations are mostly applied on a smaller scale within an object whereas on a larger scale decorations and objects usually take geometric shapes and patterns, inspired by the Quen’s knowledge of navigation and astronomy. Below you can see a couple of examples of this in these props from Forbidden West:

Designing within its established guidelines is the first part of making an object feel right for a given setting. The second part for us is material definition. We try to show believable and interesting texture through color and material variation, dirt and wear, under neutral lighting to make it readable and reproducible for a 3D artist. Depending on what’s practical for the artist and the prop in question we use different techniques to accomplish this. Sometimes photobashing and overpainting are all it takes, other times a more 3D-heavy approach is practical.

One of the more detailed props I designed was the steering wheel for a boat which the player would be able to control. A gameplay mockup was already made using a rowboat and various props from the previous game to turn it into a trimaran. I dressed it up using more pre-existing game assets and did a paintover for the parts we could uniquely design and produce which you can see below. Those were the outriggers and engines, the rudder, and the steering wheel and its assembly. After that final design work was divided among a few people in the team with me taking on the wheel and its supporting column.

With this steering wheel prop, I took the opportunity to try going for a more 3D-heavy workflow and get a better final render. It would be used and seen both on the boat as well as repurposed in other places. The plan was to reuse previously made animations for player-usable wheels with a standardized diameter.

This one also had to be based on those measurements and the outer ring of the wheel needed to be mostly smooth to avoid clipping through the character’s hands. After some rough sketches, I started with a 3D blockout to lock in a design. I like Blender for everything else but am still most comfortable with modeling in 3ds Max so that’s what I used here. I always keep a human scale model in the scene to make sure the proportions look right, and in this case, I also imported a model of another interactive wheel to maintain the correct dimensions.

The design wasn't too complicated so modeling all the individual parts was doable. For most parts, I made a low poly shape with hard edges and then added a subdivision modifier to make the topology denser without smoothing the hard edges. Then on top of that another SubDivision modifier, but now ignoring hard edges, causing the mesh to smooth less extremely due to the denser topology below. It’s dirty modeling but it’s fast and allows for easy editing of shapes without having to worry much about support loops or edge creasing. Any smoothing errors were minor enough to paint over later.

The roughest-looking area was the fabric wrapping around the center which was made by extruding geometry along splines and pushing it around using basic sculpting tools. I knew we had good photo material of the fabric I wanted here so I didn’t put in effort to make it pretty in 3D. In hindsight it would have been nice to do so as photobashing it took more time than I thought. 

Initially, the ropes weren’t part of the design but while modeling it felt a bit empty, lacking some layering or shape-variation. Luckily someone in the team suggested adding rope between the spokes which was a great way to add more material variation and also make it feel ‘more Quen’. I tried a few different patterns seeing what would best complement the existing shapes and settled on this four-pointed star tied around the smaller spokes. It seemed like a good fit for the Quen too, visually hinting at a compass star and navigation at sea. 

Lastly, I modeled out these rope sections by twisting multiple cylinders and path-deforming them along the splines I’d used to make the pattern:

With the blockout completed I took it into Blender to apply materials and render out passes. I auto-UV’d everything, straightening out the sides of the curved outer pieces so the texture’s wood grain would follow the curvature. When assigning materials I alternated light, medium, and dark wood tones which is something the Quen do as a decorative touch with many of their multi-part wooden constructions. It also creates more visual interest and shows off the segmented construction.

When designing props for Horizon we generally slightly overdo construction details. This shows an object’s handmade character and helps emphasize each tribe’s unique ways of building things. You could consider it a sort of stylized realism. This is what the materials ended up looking like rendered in the viewport with Cycles:

An advantage of fully modeling the design instead of doing most of the work in 2D is already having a backside designed. When modeling the blockout I thickened the rear of the spokes where they meet the axle to strengthen the center connection and slightly offset the wheel from whatever it attaches to. It also makes the side silhouette less flat.

What worked well for the rope material was to use a Normal Map of rope strands on each of the coiled cylinders, adding the final level of detail without having to model it all in:

Additionally, I put small, loose threads on the rope using the hair particles modifier in Blender to break up the clean 3D look a little. Those two things made the base render of the rope detailed enough to avoid having to photobash over it later on, which can be very tedious to do with rope.

I rendered out passes including direct lighting/shadows, ambient occlusion, and gloss as well as a ‘clown pass’ for masking out materials and sections of the prop. These were mostly useful for selectively adjusting the lighting response in specific areas and bringing back or enhancing lighting on top of where I’d paint over the base. The initial base render and material mask looked like this:

Unless you have a detailed sculpt the surfaces of a blockout like this usually look too sharp or too smooth to feel handmade and the materials can look a bit washed out. So a lot of the photobashing and overpainting I do is done to add more detail and reduce that 3D look.

I made minor changes to construction details in 2D, like making the outer wood pieces overlap where they join. Then I started photobashing the fabric since it was the roughest-looking part straight out of the 3D render. The wood material already had some detail and adding more to it was quick in comparison, cutting and warping bits of photo material onto it. We keep a large, shared library of photo references in the team covering everything from props and architecture to individual materials organized per tribe and type. This cuts down on time spent searching online for that one perfect ref.

The design had two brass rings going around the entire wheel that were too small and hard to see initially. Part of the feedback on my first pass was to make them larger and stand out from the wood, so I increased their width and did a region re-render of the side of the wheel where they were visible. I just pasted that on top of the layer stack in Photoshop and did a fair amount of overpainting to deepen the color range, add in small dents and imperfections on the brass, and darken AO in seams.

I tried applying various painted patterns along the outside as suggested in the sketches. The turquoise or cyan colors balance out the warm tones of the wood and ropes and complement the blue elements in the center. The first few patterns I had going around the entire wheel, but this turned out noisy and it worked out better to instead focus it at spoke connection points.

The pattern was painted with a textured brush on blend-if layers to make it look a little roughly applied, keeping it out of creases in the wood. I also clipped Diffuse, Gloss, and AO passes from the render onto the paint layergroup to get light and shadowcasting back on the areas I covered and have the wood grain show through. 

Finally, I applied more dusty patina in and around crevices to make the prop look well used and soften some color and shape transitions. It’s hard to show the 2D workflow in detail as it’s a lot of small, loosely grouped actions, but these comparisons hopefully give some idea of the changes made along the way:

I’m happy with how it turned out, especially the definition and variation of materials that came out the way I had hoped them to. It’s quite detail-oriented work, and for a prop this size, took more time than usual, about four weeks from start to finish, but it’s very satisfying to get to design something down to this level sometimes.

And here it is presented in context with the other pieces I designed for the boat:

The Challenges

It was great to get to do more environment concept art this time around with the Quen. I’m happy with the result but I have a lot more to learn in terms of composition and scene setup in 3D. The changing nature of the story during early development makes it very handy to be able to change things around quickly, and while I made an attempt at organizing objects in the scene that was still very improvised 

With the steering wheel prop, I wanted to push my rendering to be as crisp as possible. With other props, I occasionally resorted to using lower-resolution pictures to photobash with or painting over it too much. Overall I think it came out well here, although I think the fabric strips could’ve been done better. The photobashing on it is a little warped in some parts and some of the painted shadows don’t look quite right. I could and probably should have used cloth simulation to get a better result but I knew we had good photo reference of the kind of fabric strips I wanted that I could photobash with. It ended up taking more time than I thought and the result isn’t as clean as cloth sim with texture and a little overpaint would probably be. So if I’m doing stuff at a similar level of detail in the future that’s the approach I would take instead.

Advice for Beginners

I think it’s good to always approach designs within the context of the world they exist in, and to make one up if there isn’t any. Think about what rules the world has that objects need to adhere to: what are the physics, is there gravity to take into account for construction and does it have weather that could wear on materials? What materials are present and what kind of tools are used to work them? Does a prop have a specific function or idea it communicates and does that come across without having to explain it? And does your design hit those points while also being interesting to look at? I think that’s the art in concept art, to mold a convincing design into a visually interesting shape.

It can be a good idea to gear your portfolio towards the type of concept art and style you know a studio you’re interested in is doing a lot, though I think knowledge of design usually carries over regardless of style. In my experience, the most important thing is to know for yourself what you’re most passionate about and go in that direction. That will keep you motivated to keep going and improving.

Dealing With Anxiety

I think making any kind of art causes anxiety and the best thing to do is to accept and use it. For example, comparing your work to others you look up to can cause anxiety if you feel like your skills aren’t where you want them to be. But critically observing differences between them is also an important way to learn and see where to improve as well as see what you’re already doing right. To me, the anxiety I experience then is the desire to become better. Of course, at some point, it can become more stressful and damaging than useful, so you have to set limits for yourself. Then it’s better to disengage, for example, I’ll sometimes avoid looking at much other art if I’m working on a difficult piece and I already know what I want to accomplish. Or if possible just do something different altogether for a while to relax.

It’s good to have stuff besides work to channel energy or creativity into whether it’s art or something unrelated. It’s good to be invested and passionate about work but it’s also in the nature of concept art as a job that designs don’t always end up the way you would’ve liked them to. At those times having a personal project or anything else that’s fully your own thing can help a lot to stay happy.

Adriaan Stam, Concept Artist at Guerrilla

Interview conducted by Arti Burton