使用 Substance 3D Designer 节点制作逼真可平铺森林地面材质

Environment Artist

Interviewed by

19 October 2026

Zeqi Mu explains how he crafted the intricate, true-to-life forest floor texture in Substance 3D Designer, comments on how he achieved a believable root growth pattern, and shows his lighting and rendering setup for the Unreal Engine scene.

Introduction

Hello, my name is Zeqi Mu. I began my career as a 3D environment artist in 2018 and moved into procedural material creation in 2019. At the university, I studied computer science before transitioning into art.

Since then, I've worked at several major studios, including Perfect World, Papergames, Happy Elements, and Tencent. I've contributed to large-scale open-world projects such as World of Jade Dynasty, Infinity Nikki, and Silver Palace, and served as the Lead Material Artist.

In this interview, I'd like to share my process for creating realistic forest ground material.

Getting Started with Realistic Forest Ground

I began the project by gathering real-life references. My goal was to see whether I could create a forest ground material with scan-level quality using only procedural nodes in Substance 3D Designer.

After analyzing reference images of tree roots, I initially used the Cells 2 node in Adobe 3D Designer to generate their main shapes.

However, I quickly realized that this approach couldn't reproduce the natural hierarchy found in real root systems. I therefore tried a different method, using spline generators to 'grow' the roots in a more organic and controllable way.

Texturing with Substance 3D Designer

By using Scatter Splines on Splines, I was able to recreate the natural growth patterns of real tree roots. 

This method gave me precise control over both root density and thickness. The density of the main roots is intentionally quite low. This was important because the material is tileable, and having fewer dominant shapes helps reduce visible repetition.

To make the roots feel even less repetitive, I multiplied several Gaussian noise patterns into the height map, creating more variation in their elevation and surface shape.

Next, I created several large, visually prominent stones and scattered them between the roots. 

For the smaller stones, I used the SDF rock feature in the new Shape splatter v2 node. Within its function graph, I introduced random values for properties like roundness and facet count. This gave each stone a slightly different shape and made their distribution feel more natural.

I then used Shape splatter v2 to add moss to the damp areas. I generated the moss normals directly from Shape splatter v2, which produced better results than deriving them from the combined height map.

To capture the complexity of a real forest floor, I built the material from many layers. From bottom to top, the hierarchy consists of tree roots, large stones, small stones, tiny stones, moss, dead leaves, old twigs, micro-stones, clover, lichen, broken branches, grass, fallen leaves, fallen twigs, additional clover, and ferns.

Lighting & Rendering

In Marmoset Toolbag, I used a studio HDRI to present the material clearly under controlled lighting. I also created a simple forest-style scene using a single directional light and a high-quality HDRI to place the material in a more natural context.

For the character-running demonstration in Unreal Engine 5, I used UE5’s default procedural sky setup and adjusted it to create a sunset atmosphere. I enabled Nanite displacement for the material and added a subsurface scattering mask to the clover leaves.

Conclusion

The material took roughly one week to complete. I spent the first three days developing the spline-generation system for the tree roots.

The biggest challenge was refining the composition of the roots so that the material looked natural while avoiding obvious repetition across the tiled surface.

My main advice is to study plenty of real-life reference images before deciding how to approach the material procedurally. Pay close attention not only to individual shapes, but also to how those shapes grow, overlap, and form a hierarchy.

In this case, choosing a spline-based system instead of relying on cell patterns made a significant difference. It allowed me to create a much more convincing hierarchy of primary and secondary roots while retaining enough control to make the material tileable.