Building a Beautiful & Atmospheric Environment of a Fallen Titan

Introduction

Hi! My name is Shanshan He, and I'm a Senior Level Artist currently based in China, specializing in environment and level art. I studied Game Art at Gnomon School of Visual Effects in Los Angeles, located in the heart of Hollywood. Being surrounded by industry professionals from major studios had a strong impact on my early development. It was an intense period where I built a solid foundation and gained a clearer understanding of production workflows.

My path into CG was not entirely straightforward. I originally studied at a business school in Melbourne, but everything changed after I watched Avatar. The level of visual fidelity, the richness of its ecosystems, and the seamless integration of art and technology reshaped my understanding of digital worlds. That experience pushed me to transition into CG and pursue it seriously.

Since then, I've been on this journey for around 10 years. Early in my career, I joined Ubisoft Massive as a Junior Level Artist, contributing to The Division 2. Shortly after, I had the opportunity to work on Avatar: Frontiers of Pandora, which felt like a full-circle moment, since Avatar was what first inspired me to enter the industry. That experience had a strong impact on how I approach environment work. You can find more of my work on my ArtStation.

Cavern of the Fallen Titan

This project started as a personal study focused on environments shaped by gradual change over time, where organic, growth-like forms slowly take over and redefine the space. The idea was loosely inspired by environments from Monster Hunter, but rather than replicating its style, I used it as a foundation for my own interpretation.

At the same time, I wanted to explore building a cavern environment. During my time working on Avatar, I was drawn to hidden, almost secret-like spaces within the floating islands. I wanted to capture that feeling of discovering a quiet, enclosed space that feels both isolated and alive.

For references, I looked at real-world cave formations combined with more stylized and alien-inspired elements, especially in surface detail and material variation. I also studied how these growth forms spread across surfaces, using that as a guide for how the environment evolves.

Composition & Blockout

I started with composition before committing to detailed assets, with the goal of establishing a clear visual hierarchy and guiding the viewer through the space. During the blockout phase, I focused on scale, pathing, and camera composition, iterating on major forms, arches, platforms, and vertical openings, to ensure the focal point reads clearly from key viewpoints.

The primary focal point is the dragon skull, positioned off-center and framed by surrounding rock formations. It acts as both a visual anchor and a narrative element, immediately establishing a strong sense of scale. Lighting and structural forms work together to guide attention.

The main light shafts subtly point toward the skull, while platforms and the dragon's lower jaw establish a natural traversal path. Large-scale shapes help organize the space and maintain readability. Here's a quick breakdown of how the scene evolved from the initial blockout to the final result:

Assets & Optimization

Since this is a level art study, I relied primarily on marketplace assets, focusing on composition, material cohesion, and storytelling rather than creating everything from scratch. Because the assets came from different sources, I adjusted materials extensively to unify their look and developed custom shaders in Unreal Engine to better integrate them.

The dragon skull was adapted from an existing asset. I isolated the head, removed some teeth, and softened the horns to better fit the scene and support the narrative. For the rocks, I selected a small set of models with varied shapes and scales. Larger pieces define the main structure, while smaller ones refine transitions and break up the forms. To keep the workflow efficient, I limited the number of assets and relied on variation in scale, rotation, silhouette, and material adjustments to reduce repetition.

Retopologizing was not a primary focus, as the assets were not originally modeled by me. My work was mainly focused on material setup and shader development, with minor UV adjustments where needed.

Texturing & Material Blending

My texturing workflow focused on building a flexible and consistent material system, rather than relying on unique textures per asset. To keep everything seamless, I used custom shaders with vertex painting to blend materials and control how organic growth interacts with surfaces, using Roughness and Color variation rather than relying on additional geometry.

The rock assets came from various sources, with different shapes, scales, and texture styles, most of them using unique textures. This created issues with consistency, flexibility, and resolution when scaling up the scene. To address this, I kept the geometry but replaced the original textures with a unified material system. I used a small set of tiling textures across all rocks, allowing them to share the same shader setup while maintaining variation through blending.

To better illustrate how the material blending works in practice:

Moss

I integrated a moss layer, created in Substance 3D Designer, directly into the material system rather than adding it on top. Using world-aligned blending, it appears naturally on upward-facing surfaces. Sharing this layer between rocks and the landscape allows the transition to remain seamless.

Landscape

Texture repetition became very noticeable across the large landscape. To address this, I used a Macro Texture Variation approach, multiplying large, medium, and small-scale variation masks over the tiling landscape textures to break up repetition. This ensures the surface reads differently at both close and far distances, avoiding visible tiling and preserving variation across the scene.

Final Assembly & Detail Distribution

In the final stage, I focused on reinforcing composition through deliberate detail placement rather than introducing random noise. Details were concentrated around the focal area, the dragon skull, and along the main visual path, while surrounding areas were kept more restrained to maintain readability. This helps guide attention without overwhelming the scene.

Instead of evenly scattering assets, I grouped them into clusters to create variation in density and avoid uniform placement. These details also help suggest how the environment evolved, with organic growth spreading over the bones and rocks, and sediment accumulating along the lower path, reinforcing the narrative without distracting from the focal point.

Lighting & Atmosphere

For lighting, I started by defining a clear and believable light source, imagining that erosion and collapse over time had created openings in the ceiling, allowing sunlight to enter the space. The primary light comes from a skylight above, illuminating the midground and defining the silhouette of the vertical shaft and the dragon skull. A secondary opening in the background adds depth and improves readability.

Although the main skylight is not visible from the camera, I introduced a large opening in the frame, implied to be caused by the dragon, to break up the enclosed feeling while supporting the narrative. Lighting was used to guide the viewer's attention rather than evenly illuminate the scene. Most areas remain subdued, with light concentrated around the focal point and main visual path. Subtle emission in some organic elements introduces a slight variation in secondary lighting and adds a sense of life, while remaining low enough to avoid visual noise.

Working in Unreal Engine allowed for real-time iteration. Volumetric fog was used to reinforce depth and create a soft, diffused atmosphere. For post-production, I avoided relying on it early on, as it can mask issues in lighting and materials. A Post Process Volume was set up from the beginning to lock auto-exposure, ensuring consistent lighting conditions for evaluation.

Conclusion

This project was developed in my spare time, so the process was fragmented, with a long break in between, making it difficult to estimate a precise timeline. One of the main challenges was working in areas I had less experience with, particularly building a large-scale cavern and developing organic growth-driven surfaces. This required extensive experimentation and iteration.

Before building the full environment, I focused on look development using a single rock asset. The custom rock material developed through this process became the foundation for the entire scene. A key takeaway for me was the importance of iteration. Strong visual results often come from refining multiple approaches rather than settling on the first working solution.

For beginner artists, I would recommend focusing on exploration and not being afraid to discard earlier solutions. Strong results usually come from refinement, not from getting everything right on the first attempt.

Shanshan He, Level Artist

Interview conducted by Gloria Levine