Creating a Black Myth: Wukong-Inspired Door with Lifelike Vegetation

Introduction

Hello, my name is Ugo Menchi, and I’m a 3D Environment Artist based in France. I’m currently in my final term at Supinfogame Rubika, where I’ve been developing my skills in environment art and game development.

Throughout my studies, I’ve contributed to various game projects across different genres – from adventure games to roguelikes and even mobile games. These experiences have allowed me to explore multiple roles within production, giving me a strong understanding of the entire game development pipeline.

Early on, I was passionate about every artistic aspect of making games. I initially specialized in VFX for games, but after my first internship, I realized how much I missed working on world-building and storytelling through environments. That moment was a turning point – I decided to fully dedicate myself to crafting detailed and engaging worlds.

What I love most about environment art is the ability to tell a story through every detail – from the textures of a worn-out wall to the way light filters through an abandoned door.

This year, I’m working on a strategy game set in Hell, where I’m responsible for the level art, as well as the mood and atmosphere of the environment.

Outside of school, I enjoy pushing my skills further by creating highly detailed, close-up environments. I also love diving into workflow breakdowns, optimization techniques, and new industry techniques – always looking for ways to improve and refine my craft. 

Inspiration & References

During the summer of 2025, I traveled to Indonesia, where I had the chance to visit temples and explore lush environments filled with incredible details. I took tons of photos, knowing they could serve as valuable references for future projects. Over the years, I’ve developed the habit of collecting references and organizing them in PureRef, separating them into categories like 2D concepts, 3D art, and real-world photography.

After revisiting my collection, I realized I wanted to work on a project with a Chinese aesthetic while incorporating some sculpting elements. Instead of creating a full environment, I decided to focus on a single hero prop – a piece that could stand strong on its own and allow me to fully refine every detail. Since I was also balancing my final-year project, I knew I needed to be strategic and focus on something that would showcase my skills without overwhelming myself with too many tasks.

Another key reason for choosing this project was my love for Wu Kong. I’ve always been fascinated by the game’s world – the environments are so grand and full of intricate details that they make you feel tiny within this universe of giants. When I saw the concept art by Zhu Weiwei and Yilin Dong on ArtStation, I immediately knew I wanted to bring one of their designs to life in 3D.

I decided to recreate one of the beautifully detailed door designs from Zhu Weiwei. The next step was to break the concept down into different parts, allowing me to analyze each element individually and gather meaningful resources for every section.

For this, I relied on PureRef, where I continuously refine my reference board, keeping only the most essential images. I organized the board by breaking down the door into its different parts and labeling each section based on the type of reference it came from:

• 2D concept art – to stay true to the original design
• 3D works by other artists – to analyze how similar elements are
  executed in production
• Real-world architectural references – to ensure material
  authenticity and realism

Luckily, the Black Myth: WuKong concept art was incredibly detailed, giving me a strong foundation for work. Usually, I struggle to find the right references, but this time, I had everything I needed to understand the structure, materials, and intricate ornamentation from the start.

Blocking

For modeling, I started by making a mesh in 3ds Max, and as soon as I got a basic shape, I put them in Unreal Engine, set up a camera angle, and cut the mesh into different parts. The nomenclature is very important in the process, naming each part precisely allows me to easily update each element, taking a new screenshot regularly to see how the advancement of the project was going, and to refine the most valuable part to the eye.

First, I started with the roof. I created a few tiles in 3ds Max and sculpted them in ZBrush. Then, I took them back to 3ds Max and mixed them together to have rows of tiles to be able to easily place them in Unreal and created enough variations between them not to repeat too much. In Unreal, I will later add or remove by hand some tiles to create this effect of damage even more. It was important to make them individually because I wanted to create this hole in the right part, like the concept art.

Then, I worked on the dragons. Since I had trouble reproducing the dragon there on the concept because I had trouble finding the shape perfectly, I dove into my references and started to sculpt. I mainly focus on shapes because the dragon at the top was creating a silhouette for my piece, so it needed to be well-cut.

For the dragon pillar, it was a little bit harder. Analyzing the concept art, I realised that there were very cool ornaments that represented intricate shapes of dragons.

The philosophy I used was that I didn't aim for the best sculpt ever but for the sculpt that would create me the best texture maps by sculpting nice volumes, some grains, and interesting details to have the best-looking material to use in UE.

I used a different workflow depending on the shapes I wanted to do. For the side, I used two alphas that I replaced using space and selection inflate to have a nice base to start with. Then, I mainly used the TrimSmoothBorder and the ClayBuildup to create refined shapes and create some damages that would look good with baking. Quite a long process, always having to refine edges.

For the dragon, I wanted a strong shape in the middle, so I sculpted outside of the plane and then used DynaMesh to merge it with the flat rock behind. I did the same with the waves and the root around it. As I planned this part to be full of moss, I left some blank space to let the moss grow around it.

I made a first version of it, and started over because of the wobbly parts, and I wasn’t happy with the background maps. The main issue was I went too far from the original concept and the corner didn’t look Chinese statue enough, the sculpt were too small and noisy for this kind of piece. When I realised that, there is not a thousand way to turn it around, redoing it was at the end the best way to improve the overall piece.

Unwrapping & Texturing

To unwrap my assets, I used ZBrush Pro Unwrap. First, I did a retopology pass to reduce the polycount while keeping a clean silhouette. Then, I applied a Pro Optimizer to check if the unwrap was efficient. Once that was done, I moved the model into 3ds Max or Blender, where I optimized the geometry further, grouping objects with similar materials to unwrap them together.

For consistency, I established a texel density of 20.48 tex/cm across the project. This helped maintain even detail distribution and a coherent look. I followed Anthony O’Donnell’s tutorial to refresh my workflow and ensure I applied these techniques properly. Even though I didn’t use an atlas, I applied this method for my generic textures.

Since some parts of the mesh required vertex painting, I kept specific edge loops to allow better control over material blending.

Each object in the scene has two UV sets:

• First UV set → Used for baking high-poly details onto the low-poly model and
  generating an RGB mask for texture variation (dirt, leaks, moss, etc.).
• Second UV set → Used for tiling generic textures (plaster, wood, rocks, etc.),
  ensuring seamless blending and a realistic finish.

Material Workflow & Optimization

Since this was a side project, I used Quixel (now Fab) textures to speed up production. By refining my master material, I was able to blend different textures efficiently while maintaining high visual quality. 

For this project, I reused the master material I originally created for my Iron Maiden scene. This setup allowed me to control textures directly in-engine, making it easier to iterate and avoid excessive texture sets.

I worked with four main material types:

• Wood
• Bricks
• Plaster
• Rocks

Additionally, I included two moss variations and one dirt layer for extra detail. At first, my master material might look complex, but it follows a simple and structured logic that gives strong control over texturing.

Each material is built on three texture maps:

1. Base Color – I added adjustment parameters to tweak values and ensure proper color balance.

2. ORD Texture – This packed texture includes:

• Occlusion (Red Channel)
• Roughness (Green Channel)
• Displacement (Blue Channel)

3. Normal Map – Blended with sculpted normals to preserve high-frequency details (grains) while keeping medium/large sculpted forms (damaged areas, ornaments, etc.).

To control the blending, I used a set of Lerp nodes with parameters to fine-tune the influence of generic normals vs. custom sculpted normals.

Since I wanted two different generic materials on the same object, I used vertex colors and Lerp nodes to blend them smoothly.

The masking part of the material nodes is used to control the application of moss, dirt, or color variations based on the masks I created in Substance 3D Painter. By leveraging RGB masking, I can assign different materials to specific areas, allowing for precise texturing directly in the engine.

For example, on the sculpted pillar, I wanted to introduce more variation in the moss. I used two channels (R & G) to create subtle color shifts and different roughness levels depending on the surface’s position. Additionally, I used the blue channel to add dirt variation.

To apply my red mask, I followed a similar approach to my generic materials. I created parameters for UVs, Base Color, Normal, Roughness, and Displacement to maintain full control. Then, I add some parameters to the red channel of my mask texture to fine-tune how it is applied to the mesh. 

Even though I was happy with the material setup, I had to make adjustments to break repetition and introduce more procedural variations, ensuring a more natural and believable result.

To avoid obvious tiling on repeated objects like floor tiles, I multiplied the mask with world position noise, which I could toggle on or off. Additionally, I used a vertex paint mask to refine the effect manually where needed.

I also experimented with height displacement, which allowed me to tweak object shapes directly in-engine. For example, on the sculpted pillar, I used R&G channels to control moss variation, adjusting roughness and color depending on its placement.

Additionally, this project helped me explore "Yes/No" material logic, which improved my understanding of workflow optimization. Moving forward, I plan to develop a more advanced and efficient master material, giving me greater flexibility while keeping the texturing process clear and manageable.

Since I’m reusing the same mask across multiple materials, I need to enable "Shared: Wrap" in the Sample Source parameter. This allows me to reuse the same texture without occupying additional slots, which is crucial since Unreal Engine limits materials to 16 texture slots.

Lighting

For the lighting, since the 2D concept art didn’t have a strong lighting intention, I aimed for a neutral yet natural light to reinforce the idea that the scene belongs in an outdoor environment. I used a branch in front of the directional light to create natural-looking shadows and added subtle point lights to enhance the darkest areas while maintaining a realistic feel.

I also wanted the lighting to guide the viewer’s eye and emphasize the sculpted details of the door. I took inspiration from real-world photography of ancient temples to see how light interacts with intricate carvings and aged materials. Games like Ghost of Tsushima and Black Myth: Wukong also influenced my choices, as they use lighting to enhance atmosphere and storytelling.

My former mentor, Xavier Le Guen once told me that modeling is 10%, texturing is 30%, and lighting is 60% of the final result – so I took extra time to refine the shot. I used a custom LUT to unify the colors and add a subtle warmth to the scene, reinforcing the idea of an old, mystical place bathed in natural light.

I used lighting to enhance the storytelling, positioning the directional light at a low angle to create a late-afternoon glow, reinforcing the idea of an ancient door weathered by time. To add depth, I darkened certain areas for contrast, making the sculpted details stand out. Since the scene is outdoors, I aimed for soft, natural shadows, mimicking sunlight filtering through nearby vegetation. Leveraging Unreal Engine 5’s Lumen system, I achieved realistic light bounces, avoiding a flat look and bringing more life to the materials.

Time Management

This project has been a huge learning experience, and I’m excited to apply everything I’ve learned to future endeavors. The goal was to use as many techniques as possible to create game-ready assets and simplify and speed up the overall workflow.

It took me about two months to finish the project since I’m working on it after classes and during weekends. I had to redo the pillar sculpt at some point since it didn’t match my goal rendering. I came up with new sculpting techniques and dug up some tutorials from Damien Peinoit to sculpt rocks and ornaments.

One key takeaway from this project is the importance of planning and organization from the very beginning, whether it’s gathering references or setting up a proper material pipeline.

Concept art can sometimes lack depth or clear detail, so I had to rely on real-world references and analyze similar 3D works to make informed decisions.

Another complexity was achieving a high level of detail without overwhelming the scene – especially with sculpting and texturing, where I had to find a good balance between hand-sculpted details and procedural techniques.

I also want to thank Hugo Ducrocq, a former Rubika student and now a Vegetation Artist, for his valuable feedback and tips on texturing.

In the future, I plan to focus on improving my lighting skills, as it’s crucial for creating mood and depth in environments. I aim to relight a scene to refine my technique and better understand how lighting can transform the overall look and feel of a project.

I’m also actively looking for an internship starting in July 2025. If you're interested, feel free to contact me on my LinkedIn page or by email at u.menchihinojo@rubika-edu.com.

Ugo Menchi, 3D Artist