Alex Medina talked about creating weapons for Hunt: Showdown, explained the modeling and texturing processes, and told us about being a hard-surface artist for games.
Hi, 80 Level! I’m Alex, a 3D artist from Valencia, Spain. At the risk of sounding cliché, since I was a little kid with a Game Boy Color I was sure I wanted to make video games for a living. Alas, there were no game studios nearby or higher education options when I left high school, so I chose to study another discipline that mixed art and technology: architecture. Which I studied at the Polytechnic University of Valencia. During a brief stay in China in my final school year, I interned in a Chengdu urban planning studio, after which I decided architecture wasn’t really for me. Back in Spain, I enrolled in a master’s degree program in videogame art from Florida Universitaria and fell in love with 3D art.
Thanks to being recommended by one of my teachers, I joined the outsourcing studio elite3d in Valencia, where I worked for almost three years. There, I worked on Blood & Truth (PSVR), and Cyberpunk 2077 as a 3D artist doing props, vehicles, and environment sets. These days I work as a hard surface artist at Crytek (Frankfurt, Germany) on the game Hunt: Showdown.
Examples of my work at elite3d:
Joining the Hunt: Showdown Team
After Cyberpunk, I decided I wanted to specialize in weapon art and move to an in-house studio to have more responsibility and creative power over my work. I focused on bettering my weapon skills with a personal project (the PL-15 pistol) and then joined Crytek’s hard surface team in October 2020. The team, which is responsible for props, vehicles, and weapons, is comprised of about six people.
Having recently been on both sides of outsourcing work, I can lay out how being an in-house artist differs from outsourcing:
- There’s a bigger emphasis on design and creativity, more overall independence and trust in the artist, and less hand-holding by leads/producers. We get some degree of freedom to choose our own tasks, versus just doing what we’re assigned in OS. Schedules are also, in my experience, a bit looser.
- We’re open to taking different routes and making more high-level decisions. Not restricted by what the client wants, we focus on what benefits the game most, and also creates excitement and engagement in the community. Having an idea that solves a problem can lead to a discussion with the relevant department to get it in the game. There’s usually not such deep communication with outsource clients, and it tends to go through a producer/manager instead of direct to the team.
- There are a lot more meetings: team stand-ups, feature-related, art department check-ups, company meetings, 1-on-1’s with your lead, playtests...
- Projects are way longer in-house. It’s difficult to hop between different games inside a single studio, but it’s very common in outsourcing. That’s why OS is great to become acquainted with different ways of working, but in-house really lets you get deep into how every aspect of a game works.
Regarding weapon production, there are two main types of tasks for hard surface artists in Hunt: new weapons and legendary skins/variants. The first type is obviously not as common as the second, since a whole new weapon involves a huge amount of work for nearly every department: How will it be balanced along the rest of the arsenal to make it viable but not overpowered? How will it animate and sound? Is it operated in a similar way to any weapon we already have, or does it need completely new code and logic?
Still, Hunt: Showdown being a “realistic” game, a new weapon requires less preparation on the art side than a legendary skin. With the right references (YouTube and instruction manuals are lifesavers) and proper study of how they work, we can focus on nailing the model and then add character with textures. My first task here was the Terminus shotgun, and I’m really grateful for the trust the team put in me to bring a whole new weapon to the game so early in my new role!
Legendary skins, on the other hand, require way more focus on the design side. Not only do they have to be new, original, and look awesome, but also fit within the game’s art direction and lore, and not be too “crazy” to finish in a realistic timeframe. This creative process can speed up with the help of real-life references and the Narrative Design department, which feeds us ideas when we’re a bit stuck. We also have to be careful not to introduce materials, objects, or themes that could be too modern for the 19th-century setting – one example is our limited use of welding since it wasn't all that common at the time.
An extra difficulty is adding mesh attachments to skins because straps, trinkets, and cloth covers all need to fit the existing animations and avoid clipping or obstructing the sights.
In many games, weapon skins are absolutely bonkers, and that’s great! Here though, whenever possible, we try our best to keep them grounded in reality. While designing a new skin we try to make it “buildable,” which means no floaty bits, everything bound more or less realistically, and nothing that interferes with mechanisms or operation.
Finally, most of the weapons in Hunt: Showdown are modular, which means they’re built with interchangeable pieces. This makes it extremely easy to make new variants (longer, shorter) or add attachments (bayonets, scopes, silencers) by just replacing the barrel, stock, or handguard while keeping the receiver intact.
From the 2.8 release in 2019 onward, my main modeling tool has been Blender, and since then I’ve noticed more and more AAA artists in big studios adopting it, so its future looks bright! Regarding plugins, I don’t use too many, especially the ones that change how to use the software in a big way, because then it becomes a chore to switch between computers, keeping everything updated, remembering all the custom shortcuts… That said, self-contained add-ons like MESHmachine that help with very specific modeling issues or speed up certain tedious tasks are great, but vanilla Blender is definitely good enough to model high-quality weapons.
My workflow involves using live booleans but keeping their number of sides non-destructive by using modifiers. For example, with three edges and a screw modifier, we can create cylinders, use them as boolean cutters, and then change the number of edges depending on the step we are on. Same thing with boxes and bevel modifiers. This way, we don’t need to retopologize: we can use the same live mesh to achieve both a dense enough high poly and a low-res enough low poly using the same procedural boolean shapes. This is also how the Boxcutter add-on operates.
This method is great for complex high poly pieces, but its destructive nature means it’s slow to iterate on. For this reason, I use whatever method is best or fastest depending on each individual piece. Simpler pieces might be faster to model with support loops or bevel modifiers than going through the whole remesh/polish pipeline. Afterward, though, we need to take care that their smoothness doesn’t look too different between pieces. It’s easy to get very soft edges in ZBrush, but then we can’t have sharper edges everywhere else or they’ll look out of place.
I have to add that Blender is slowly catching up to ZBrush in the remesh+polish workflow. There’s the “Mesh Filter – Smooth” sculpting tool that’s equivalent to polish in ZBrush, although it leaves some noticeable shading artifacts on the edges, even with a very high polycount.
To avoid this, we can use the “Edit Face Set – Fair Tangency” tool, which makes perfect bevels from an existing face set. This one is easier to control, we can separate bevels and give a better result than ZBrush’s Polish, even though it’s a bit harder to set up (shoutout to Ron Frölich for teaching me this method!). The missing link right now is Blender’s performance when dealing with big high poly meshes.
Here are some step-by-step instructions for both of these methods in a small Blender guide and notebook I keep in this Google Doc.
As a final note on modeling, there’s a modifier that comes in very handy when modeling mesh attachments for existing weapons: Shrinkwrap. It lets you project new meshes/splines into existing geometry, saving us a lot of time when adding straps, strings, or cloth to weapons. Of course, the straps don’t take one another into account, so we need to fix the clipping later, but it’s still faster than doing everything from scratch.
Retopology and UVs
Unwrapping in Blender works pretty well, but it is kind of bare-bones and that can make certain tasks (like straightening UV islands) slow and painful. I use three add-ons for UVing that make my life easier: UVToolkit 2.0 (commercial), UVPackmaster PRO (commercial), and Texel Density Checker (free). These add the tools to work fast and avoid boring, repetitive stuff.
When UVing in Blender, try to work as much as you can on the 3D space, marking edges as seams and continuously unwrapping with the U key, and trying different unwrapping methods – they yield different results! I noticed conformal tends to keep the object proportions a bit better in hard surfaces so I tend to use that one. Once a UV shell is “finished,” with no big distortions, I straighten it, align it, and pin it (using P) to keep it as is, and then I move on to the next one. It's often worth it to introduce a little distortion in the UVs if it means shells are straight: this will help avoid aliasing (texture stepping) issues and make texturing easier.
This advice is more general, but try to keep the low poly mesh’s shading as clean as you can. When baking, the Normal Map has to “bridge the gap” in shading between high and low poly meshes. Because of this, a “dirty” low poly shading means the normal map has to work harder, introducing gradients that make it more vulnerable to errors during compression or downsampling (which is very common in games). Also remember: hard edges require UV cuts, but UV cuts don’t require hard edges! Please don’t put hard edges in cylinders!
As for optimization, FPS’s guns have to be optimized less aggressively than other parts of the game, since they’re visually always present, display up close, and are thoroughly inspected and operated. It’s not unusual to employ three to four texture sets and a huge amount of polygons to keep all the small details and silhouettes as pristine as possible. That said, we do work with a big amount of LODs to aggressively lower the fidelity once the camera is far enough.
Hunt: Showdown has been in production for so long, the first weapon models in the game were textured in Photoshop! These days of course we use Substance 3D Painter. I don’t think there are many “tricks” involved: just look at interesting references, then look back at your textures and try to bridge the quality gap. If they’re not close yet, then they need more work! I can say this though: not all references are “good” just because they come from real life. Real-life objects can be very ugly and poorly designed, so actually choosing the reference well and adapting it to our needs is as important as any technical step.
There are two main materials we use for Hunt: Showdown’s weapon art: wood and metal.
For metals, the spec/gloss workflow allows us deeper material customization (especially when adding dirt and wear) compared to just metalness. It’s in the middle-ground materials between metallic and dielectric where the specular workflow shines. Personally, I had not used this workflow before coming to Crytek, but I’ve come to enjoy it a lot, even if it isn’t as easy to work with (at first) as metalness/roughness. A breakdown layer by layer would be:
- Base metal: it must have proper PBR values (special attention to the specular color and levels!). Since the project has been going on for years, we already have a library of very cool substance materials made by Crytek artists, with many exposed parameters to adjust depending on the case.
- Treatments: blueing, case-hardening, anodization... be careful, they often change the surface into a non-metallic oxide.
- Finishes: milling, polishing, engravings, inlays.
- Rust. Use sparingly!
As for wood, this is one of the trickiest materials to get right but it’s also extremely versatile. The range of colors and finishes we can achieve with it while sticking to realism is huge.
- Wood pattern. The most important part! Sometimes taking it from an appealing photo or reference is a good idea: in this case, the pattern would be a black and white mask driving a Gradient where we could tweak the wood colors.
- Grain. The smaller lines and bumps caused by the direction of the wood fibers. Height/gloss variation and usually color darkening.
- Gradients. Color variations and darkening, especially in the parts connecting to other materials like metal. If in doubt, always add more color variation, even with a very low opacity it will help the material feel more unique.
- Varnish. Top layer that amplifies the color saturation and glossiness. It doesn’t have to be present everywhere of course! Removing it is as important as placing it.
Finally, we have the common layers:
- Material-specific wear and tear (edge wear, scratches, bumps, discoloration). References are a must here: each type of wear tells a different story. Maybe the gun is a prized family heirloom in which case there won’t be any scratches.
- Gunpowder burns and deposits.
- Stains (oil, paint, grease).
- Dust and dirt.
Ornaments and Engravings
As for engravings and ornaments, we have to think about how they would be achieved in real life: a delicate silver inlay won’t have the same footprint on the object as a rudimentary wood carving. Artistically speaking, it’s important to keep some balance even in the most ornate weapon: not everything should be intricately decorated or we would run the risk of making the gun extremely noisy. Composition rules apply: we should keep some rest areas with less decoration, a change of style, material – whatever we can do to break up patterns.
On the technical side, there are many Substance Painter tools to speed up texture production: filters (transform, mirror) and anchor points to make duplicate engravings faster, a tri-planar mode for when UV’s don’t really match or are distorted.
Cool use for anchor points (to make, for example, a stamped serial number on metal) is instancing a mask to create text and also the stamping mark (base mask + blur filter + different height value). This way, any modifications done on the base mask will also reflect on the instanced one, and the Painter file will be a bit lighter because it uses fewer masks.
It’s also important to keep things non-destructive wherever possible. One way of doing it is using fill layers instead of paint layers for specific details, like lines or logos, so they can be moved or adjusted, instead of having to paint and delete strokes.
Something that’s unique to Hunt: Showdown is the gameplay-driven dynamic mud masks and clumps. Every time a hunter gets downed, their weapons get progressively dirtier. This is achieved with a mud mask, essentially a mask that blends the weapon textures and a tileable mud material. The mask’s levels go up after each death, resulting in a smooth “getting dirty” effect. After a couple of deaths, the mud clumps (small meshes that are mapped to the same tileable mud texture), which were previously transparent, appear. The clumps make this progressive mud way more realistic, and help dirty weapons pop!
Approach to Art
I do think that textures can make or break a project, and this is especially true for weapons. An average model with amazing textures will always get more attention than a great model with just OK texturing. It’s indispensable to treat texturing with the love and attention it deserves, and at Crytek, texturing usually takes about as long as all the other steps combined.
3D weapons must achieve a balance between high-frequency details, which look good on close-up shots; while maintaining a good enough low-frequency variation, making the weapon not look too “flat” from a distance. This is difficult to achieve, and it’s approached differently depending on the game type:
- Third-person shooters need weapons with big, defined macro details, easy to read from afar even when the weapon is very small on-screen.
- Simulators, weapon disassembly games, or games with a heavy customization component need to nail the small microdetails because players will zoom a lot into each and every part of the gun. Sometimes, bigger textures will be used to achieve enough micro quality.
- First-person shooters need a nice balance between the two, and special consideration to the upper area around the rear sight, since it will be extremely close during ADS
One tip that comes to mind is don’t get tunnel vision. You should change the camera angles, lighting, distance, and other variables to check that the model looks good in all conditions. Most important of all: it doesn’t matter at all how the model looks inside Substance 3D Painter, only how it looks wherever you want to render it, be it a game engine or a standalone renderer like Marmoset Toolbag. Export your textures constantly and check, check, check!
Challenges and Memorable Moments
For me, the best part of working on Hunt: Showdown is being a part of the dev team and the community at the same time, engaging with other players, seeing them react to our work... It's priceless, and I feel very fortunate to be in this position! We have a very dedicated community that really appreciates the work we do.
I’ve been a Hunt fan since Early Access – not just of the gameplay, but of the art direction and atmosphere too and it’s been a dream come true to see my work in the game and collaborate with such a talented team. Hunt: Showdown is a very special project, made by a small but passionate team who work tirelessly to make it the best it can be, and I think it shows! There might be a bit of bias here though…
It’s definitely been a challenge to join a live game so late in its development, with many of its pipelines and quirks already set up with years of changes and updates. But it has served as a huge opportunity to see how a game is maintained month by month instead of for a single release day: it’s a marathon, not a sprint. Another challenge was to beat impostor syndrome and slowly introduce my own designs into the weapon skins instead of blindly following references or concept art. Design is a whole new beast to tackle, but I feel more and more confident with it every day.
Finally, I want to thank the hard surface team and especially Alex Asmus and Matthias Wagner for teaching me so much and being overall awesome people. And also thanks a lot to you, reader, for reading this whole thing: I hope you enjoyed it and learned something useful!