Cube 3D Node

Cube 3D is one of the simplest primitives in Fusion. It generates a parametric cube with six faces that can be positioned and textured independently. It is rarely a final on-screen element: it is a reference object widely used in 3D compositing as a placeholder when rebuilding a scene, as a source of cast shadows or as a support for an environment map.

For other primitives (sphere, cylinder, plane, torus, …), look at Shape 3D instead, which groups several shapes into a single node.

• Effects Library: Tools > 3D > Cube 3D
• Quick add: right-click in the Node Editor → Add Tool > 3D > Cube 3D

• SceneInput (orange) - optional. Connects another 3D scene or object; the extra geometry is added to the cube's scene.
• MaterialInput ×6 - one per face. Accepts either a 2D image (bitmap) or a 3D material (Blinn, Phong, Cook Torrance…). Materials set here do not propagate to objects connected through SceneInput.

The six material inputs match the cube faces +X, −X, +Y, −Y, +Z, −Z. In the Inspector, they appear as MaterialInput1 to MaterialInput6. You can leave some faces empty - they then get the default material.

Cube 3D is usually connected to a Merge 3D alongside a Camera 3D and a Renderer 3D to be displayed. View the Merge 3D in a viewer to work on the framing, then the Renderer 3D to get the final render.

To animate the cube (position, rotation, scale), use the node's Transform tab or feed it into a separate Transform 3D to isolate the animation.

Lock Width/Height/Depth (default: checked)

Locks the three dimensions. When on, a single Size slider drives width, height and depth at once. Uncheck it to expose three independent Width / Height / Depth sliders.

Size / Width / Height / Depth (default: 1.0 unit)

Cube dimensions in Fusion units. Animation set on Size is automatically transferred to Width when you uncheck the lock, and the other way around.

Subdivision Level (default: 1)

Number of subdivisions per face. By default the cube is made of 12 triangles (2 per face); raising the subdivision adds intermediate vertices. Essential for:

• Smooth vertex lighting - a barely subdivided cube gets shading per face (a "flat" look); more subdivisions = finer gradients.
• Deformation - before a Displace 3D, a Bender 3D or a Custom Vertex 3D, go up to 8-16 at least.

Cube Mapping (default: off)

Wraps the first texture over the six faces using standard cube mapping. The source image must be in cross format. See the dedicated section.

Wireframe (default: off)

Wireframe render (edges only). Works with a Renderer 3D in OpenGL mode, not with the software renderer.

When the Cube Mapping option is on, Fusion expects an image with a cross layout - a Latin cross where each arm matches one face of the cube:

• Center column (top to bottom): top, front, bottom, back
• Left arm: left
• Right arm: right

Typical resolutions: 4096×3072 or 2048×1536 (4:3 ratio). Fusion is not strict about dimensions, but if the layout does not follow the cross, the faces will look offset or pinched.

If you start from an equirectangular image (360° HDRI), first convert it to a cross cubemap with an external tool (HDRI-Gen, SpaceScape) or a dedicated conversion node.

Without Cube Mapping, each face gets its own texture through its material input - more flexible for product packshots or "six views" presentations.

• 3D tracking placeholder - during a camera track with the Camera Tracker, a cube placed on the ground lets you check that the camera solve is consistent (the cube must stay "stuck" to the ground plane).
• Invisible shadow caster - a cube rendered as Shadow Only in the Renderer 3D settings casts shadows on a plane without being visible itself. Handy to add the shadow of a composited object after the fact.
• Environment mapping - a giant cube enclosing the scene and textured with Cube Mapping from the inside serves as a backdrop for reflections or as a skybox.
• Multi-face product packshot - six material inputs = six different photos, for example to show the sides of a package.
• Simple 3D logo bumper - a rotating cube with a logo on each face makes quick branding without any external modeling.

• Keep Subdivision Level = 1 as long as you don't need fine lighting or deformation - it saves vertices and speeds up rendering.
• For a "physical" cube (crate, block), enable Lock Width/Height/Depth then animate Size - one keyframe to manage instead of three.
• A wireframe cube on top of a tracking solve is a very clear visual marker to validate the camera alignment.
• Avoid Cube Mapping for product packaging: each face has its own bleed and orientation, so separate inputs are easier to adjust.
• If you apply a Bender 3D or a Displace 3D after the cube, raise the subdivision to 8-16 at least, otherwise the deformation will look "stepped".

The Materials, Transform and Settings tabs share the options common to all 3D nodes: visibility, lighting, matte, blend mode, normals / tangents, object ID, and transform controls (position, rotation, scale, pivot).