Breakdown: Building Lived-In Stylized Viking Environment In 3D

Introduction

Hii! My name is Tulika Roy, and I am a 3D Environment Artist. Recently, I finished my MA in Games Art and Design at the University of Hertfordshire, and I made Outstead as my final major project.

I first became interested in 3D while doing my BA in 3D Animation & Visual Effects. We worked across different areas, including animation, VFX, etc., but I started to develop an interest in 3D modeling. I enjoyed making something from scratch and seeing it gradually come together as a whole environment or prop.

As I continued developing my skills, I became more interested in environment art. This then became the main direction I wanted to explore during my MA, through stylized environments like my final major project, Outstead.

Concept & Inspiration

I looked through a ton of different concepts because I didn't want to choose something that had already been done a lot before. After looking through different ideas, I came across this concept art of a Viking House by Skylar Zhang that I really liked.

I really liked how nicely designed it was, but it wasn't too complicated, which meant I could still make my own changes to it and add a few details of my own to fit my story behind this environment. I really liked the shape of the house and the slight curve in the roof, which gave it a lot of character.

Another reason I chose this was that I wanted to use this project as an opportunity to use hand-painted textures for the first time for a full environment. I had recently started learning hand-painting, but I had never made an entire environment using this before. I also started thinking about who might actually live in the house. I imagined it belonging to two old friends who had moved away from their more adventurous lives and were now living a much quieter life together. I didn't want the characters themselves to be the focus, but I wanted some of the environment and props to give a small idea of who they were.

I used a lot of references throughout the whole project, from the house and larger environment down to individual props and smaller details. This was important because it can be surprisingly difficult to create something completely from imagination and still make it feel believable. Having real references along with stylized references gave me something to work from and helped me make better decisions about the shapes, materials, and details of each asset.

Blockout & Composition

Once I had a solid idea, I started with a simple blockout of the house and the surrounding environment in Unreal Engine. I mainly wanted to get the scale and composition working before diving into the details.

I wanted the house to stay as the main focus of the scene, so I started thinking about how I could use the landscape and the smaller elements around it to guide the viewer towards it. I used things like the path, fences, and hills to help frame the house without making the environment feel too enclosed.

I also wanted the environment to feel like an open Nordic grassland rather than a dense forest. I used low hills to break up the background and stop the scene from feeling too empty, while still keeping the house as the main focal point.

Modeling Workflow

For the modeling, I mainly used Maya and ZBrush. I used Maya for making most of the assets, and ZBrush was useful for sculpting some of the finer details. I started with the house because it was the main focus of the environment. I wanted to keep the overall shape from the original concept art while also making some changes so that it felt like my own version of the house. I treated the house as a hero asset.

One part I spent quite a bit of time thinking about was the roof. I looked at a few different ways of making it, including using Substance 3D Designer, but I ended up deciding to model the roof itself. Because it was such a noticeable part of the house, I felt that modeling the actual roof shingles would give me more control over the shape and the overall look that I wanted. I modeled a few variations of shingles and used MASH and a lattice to shape the roof as needed.

For the smaller assets, I tried to keep the modeling simple. I also used references for each asset instead of trying to make everything from memory, which helped me make sure the proportions and shapes made sense.

Trim Sheet Workflow

This was my first time properly using a trim sheet, so I started by following a tutorial to understand the basic workflow. While working on it, I remembered another tutorial I had used before for a test, where I made a stylized wooden plank. That gave the plank a lot more shape and dimension, rather than making it feel completely flat. I thought I could combine the two techniques and use a trim sheet that has more shape and dimension.

I also made a second color variation of the trim sheet. I didn't want every wooden asset in the environment to have exactly the same color, as I felt that would make the scene look a bit flat and repetitive. The second version gave me a bit more variation without having to make a completely new trim sheet.

For the texturing, I also took some ideas from an 80 Level article by Alex Iveroth; I liked how they said to choose a local color, limit the palette, and then add a stronger color splodge in certain areas. I was really happy with how this turned out because combining the two workflows gave the trim sheet much more shape and made the wooden assets feel more dynamic, instead of looking like a flat texture placed over the models.

Texturing & Materials

For the texturing, I wanted to use hand-painted textures for the whole environment. I had only recently started learning how to hand-paint, so I was still figuring it out.

Something that surprisingly helped me with this was using acrylic markers in my free time. I started using them just to relax and make something creative outside of the project, but I found that physically layering colors helped me understand color theory a bit better. I started noticing how one color changed when I put another color over it, which gave me a better understanding that I could use while texturing.

For the wooden assets, I started with a base color and then built up the texture by adding shadows, darker areas, highlights, and lighter highlights. I also added small patches of stronger color to break up the surface and make it feel less flat.

I took some of this approach from Alex Iveroth's article, particularly the idea of keeping the color palette limited while using small areas of stronger color to make the texture stand out. I used a similar idea across the wooden assets so that they would feel like they belonged to the same environment.

This was probably one of the areas where I noticed the biggest change in my work during the project. I realized how much hand-painted textures can elevate an asset.

I also found a way to save quite a bit of time by turning my wooden trim sheet texture into a custom Smart Material in Substance 3D Painter. I already had all of my base colors, highlights, and shadows set up in the trim sheet, so I just converted that folder into a Smart Material that I could reuse on other assets.

This was especially useful for the house, as I could apply the same material to things like the pillars, doors and gables and then add the details that were specific to each asset. It also meant I didn't have to keep going back to the trim sheet and checking which color I had used for a particular highlight or shadow. The colors were already there and stayed consistent across the different assets, which saved me a lot of time.

It didn't completely texture the assets for me, as I still had to add the individual details by hand, but having the main colors and layers ready to go made the process much quicker. It was probably one of the most useful shortcuts I found during the texturing stage.

Runtime Virtual Texture (RVT) for Grass

For the grass, I used a Runtime Virtual Texture (RVT) in Unreal Engine. The basic idea is that the RVT allows the grass material to read information from the landscape underneath it. In my case, I used it to pass the color information from the landscape into the grass, so the grass color could change depending on the area it was placed on.

I followed a tutorial for setting this up, which was really helpful because it explained the process step by step. The tutorial used Blender and Unreal, but I was able to follow the same idea while using Maya for my assets.

For my scene, I painted different colors onto the landscape and then used the RVT to make the grass material read those colors. This meant that instead of having one flat color across all of the grass, the grass could pick up the color from the ground underneath it. Areas with slightly different colors on the landscape would therefore also have slightly different colors in the grass.

This made the grass feel more connected to the environment without me having to manually create different grass materials for every area. It also meant I could make changes to the landscape colors and have those changes affect the grass automatically, which made the process much more efficient. I also kept the grass movement fairly simple by adding some wind to the material. This gave the grass a bit of movement without making it too distracting from the rest of the scene.

Foliage

Foliage was another area that was completely new to me on this project. I had never made foliage before, so I had to learn how to approach things like flowers, grass, and tree leaves while working on the environment.

For the flowers, I created the base color and alpha maps in Photoshop and then used simple planes in Maya to build them. I used a similar approach for the tree leaves, creating the leaf textures and placing them onto planes rather than modeling each leaf individually. This kept the assets simple while still giving me enough control over how they looked in the environment.

For the tree canopy shader, I followed a tutorial by Victoria Zavhorodnia, which helped me understand how to create a more stylized and fluffy-looking tree shader in Unreal. I also exposed some of the material parameters in the material instance so I could experiment with the look and adjust it for my own scene.

Assembling the Environment & Composition

By the time I started assembling the final environment, I already had a blockout and a main camera view that I was happy with. This made the composition much easier because I already knew where the main elements needed to go. I also used the original concept art as a guide, as I chose a similar camera angle so that the final environment would keep some of the same framing and feel as the original design.

I started placing the larger assets first, such as the house, fences, and other larger environmental elements. This helped me see how everything was working together before I started worrying about the smaller details. Since the house was the main focus, I wanted the surrounding assets to support it instead of competing with it.

I also tried to keep the story of the environment in mind while placing the props. Things like the bench, barrels, chest, helmet, and other smaller assets were placed around the house to make the space feel lived in and give small hints about the people who lived there. I didn't want to overcrowd the scene with props, so I mainly focused on the areas that would actually be visible from the main camera.

Lighting & Rendering

For the lighting, I wanted to keep the same peaceful and cozy feeling as the rest of the environment while still having enough contrast for the house and smaller details to be readable. I tried a few different lighting setups before settling on one that worked best with the colors and overall mood of the scene.

I used Unreal Engine for the final scene and rendering, and I used Ultra Dynamic Sky to help set up the sky and overall lighting. I also adjusted the lighting based on how it looked through the main camera rather than trying to make every part of the environment equally bright.

Time, Challenges & What I Enjoyed

Managing the project alongside my classes, part-time job, and everything else was already a challenge, and I also had an unexpected situation during production that meant I lost some valuable time. When I came back to the project, I found it difficult to get started again and had lost a lot of confidence in what I was doing.

One thing I learned from this was that sometimes the hardest part is simply starting. I realized that I didn't need to make something perfect straight away. I just needed to make something exist first, and then I could slowly build on it and make it better. Once I started working again, my confidence gradually came back.

I was also quite nervous about trying workflows I had never used before, especially foliage, which was one of the reasons I ended up leaving it until last in the project. However, it turned out to be one of the most enjoyable parts to work on. Being able to make something that I had never known how to create before and then actually see it working in my environment was really rewarding.

I could also see a big difference when comparing Outstead to some of my earlier environments. Seeing that progress and knowing that I had learned so much during the project was probably one of the most satisfying parts of finishing it.

Conclusion

Outstead took around two and a half months to develop, and looking back, I feel like I've improved a lot over that time. The project gave me the chance to work on areas I wasn't as confident with before, especially hand-painted texturing, foliage, trim sheets, and materials in Unreal.

One of the biggest things I learned was how important planning and preparation are. Having good references, a clear blockout, and a strong composition early on made the later stages much easier. I also learned to focus my time on the parts of the environment that mattered most rather than trying to make every asset equally detailed. I also became much more confident in making my own decisions, especially with hand painting.

There are still a lot of things I want to improve, and I want to keep developing my hand-painted texturing and environment art skills. Overall, Outstead gave me a much better idea of how I want to approach environment art going forward.

Tulika Roy, 3D Artist

Interview conducted by Amber Rutherford