One of the most common questions I get from artists transitioning into game-ready asset creation is when to use displacement maps versus normal maps. Having spent years working in Modo as part of my BFA in Computer Graphics at NYIT and through professional 3D production work, I have developed a clear mental model for this decision. Both techniques simulate surface detail, but they solve fundamentally different problems with very different performance characteristics.
How They Differ Under the Hood
A normal map is an RGB texture where each pixel encodes a surface direction rather than a color. The red, green, and blue channels represent the X, Y, and Z components of the surface normal at that point. When the renderer calculates lighting, it uses these modified normals instead of the actual geometry normals, creating the illusion of bumps, grooves, and surface detail. The critical word is “illusion.” The underlying geometry remains flat. Normal maps use no additional polygons, which makes them fast to render and ideal for real-time applications.
Displacement maps take a different approach entirely. They physically move the vertices of the mesh according to the grayscale values in the map. White pushes vertices outward, black pulls them inward, and gray leaves them in place. This produces actual geometric detail. When you view a displaced surface up close or at a grazing angle, the silhouette shows real depth. The tradeoff is significant: displacement requires a highly subdivided mesh to capture the detail, which means more polygons and substantially longer render times.
When to Use Each
In Modo’s rendering pipeline, the choice comes down to the final delivery medium. For game assets, architectural visualization, and any real-time context, normal maps are almost always the right choice. The rendering budget in games is measured in milliseconds per frame. You cannot afford to render millions of displacement-generated polygons when a normal map on a low-poly mesh achieves a visually comparable result at 60 frames per second.
Displacement maps earn their keep in offline rendering scenarios: cinematic sequences, product visualization, and print-resolution stills. When a camera pushes in close to a surface, the flat silhouette of a normal-mapped mesh becomes obvious. Cracks that should cast shadows do not. Bumps at the edge of a model look painted on rather than sculpted. Displacement maps hold up under this scrutiny because the detail is physically there.
The hybrid approach is what I use most often in production. I model the primary and secondary forms in the base mesh, use displacement for tertiary detail in hero shots that need to hold up at high resolution, and bake everything down to normal maps for the real-time version. Modo 901’s baking workflow makes this straightforward: you can bake from a high-poly displaced mesh to a low-poly cage, capturing the displacement detail as a normal map along with ambient occlusion in the same pass.
Practical Baking Workflow in Modo
The baking process starts with setting up a render output for the normal map in the Shader Tree. In Modo, you create a bake item, assign your high-poly source and low-poly target, set the cage distance to encompass the displacement detail, and render. The key settings that trip up most artists are the tangent space convention, which must match your target engine, and the cage offset, which needs to be generous enough to capture all displaced geometry without introducing artifacts from self-intersection.
For game pipelines, I export tangent-space normal maps in OpenGL format for Unity and DirectX format for Unreal. Getting this wrong produces inverted green channel lighting that makes every surface look like it is lit from below. It is a small detail that causes enormous confusion if you are not aware of the convention difference. The investment in understanding these fundamentals pays dividends across every asset you produce, regardless of which DCC tool or engine you end up working in.