Blender Dyntopo: Constant vs Relative Detail Size

In Blender's Sculpt Mode, Dynamic Topology (Dyntopo) allows you to sculpt complex shapes without worrying about the underlying mesh topology by automatically adding and removing geometry as you brush. The two primary detail sizing methods—Relative Detail and Constant Detail—dictate how Blender calculates the density of the generated polygons. Relative Detail scales mesh resolution based on your viewport zoom level, while Constant Detail applies a fixed geometric resolution based on the 3D scene units regardless of your camera view.

Relative Detail

Relative Detail is the default sizing method in Blender. It ties polygon density directly to your screen space and camera zoom level.

  • How It Works: When you zoom in close to your model, your brush strokes create finer, smaller polygons. When you zoom out, the same brush size produces larger, coarser polygons.
  • Detail Size Unit: Measured in pixels (px). A smaller pixel value produces a higher density of vertices.
  • Best Use Cases: Ideal for the early conceptual and block-out phases of sculpting. It allows you to quickly sketch primary forms from a distance, then zoom in to add localized details like wrinkles, pores, or sharp edges without manually adjusting brush resolution settings.

Constant Detail

Constant Detail decouples the mesh resolution from your viewport camera and binds it to the 3D world space.

  • How It Works: The polygon size remains identical whether you are zoomed far away or extremely close to the model. The resolution only changes if you manually adjust the detail value.
  • Detail Size Unit: Measured in percentage or detail units relative to the scene. A higher resolution value generates finer geometry.
  • Key Features: Includes a "Detail Flood Fill" button to apply uniform resolution to the entire model at once, as well as an eyedropper tool to sample detail sizes from existing parts of the mesh.
  • Best Use Cases: Ideal for mid-to-late stage sculpting where consistent surface density across the entire object is required. It is especially useful for hard-surface sculpting, adding uniform textures, or preparing a base mesh for multiresolution displacement baking.

Summary of Differences

  • Relative Detail: View-dependent (pixels). Zooming in increases detail density; zooming out decreases it. Best for organic, freeform concepting.
  • Constant Detail: World-dependent (scene units). Zooming has no effect on detail density. Best for uniform geometry distribution and controlled detail consistency.