Shape3D

Shape3D is a foundational procedural-geometry generator in classic Fusion 3D. Its output is not an RGBA image but a scene3d stream: the selected primitive exists as real geometry, can be lit, cast and receive shadows, pass through 3D modifiers, and becomes an image only in Renderer3D.

The important input model has two distinct sockets. The orange SceneInput accepts an existing 3D scene stream. The green MaterialInput is dual-type: it accepts a complete 3D material, and it also explicitly accepts a normal 2D image as the diffuse texture for the built-in basic material.

Shape3D does not require external geometry to operate. Even with empty inputs it can generate Plane, Cube, Sphere, Cylinder, Cone, Torus or Ico. SceneInput is for serially adding the new primitive to an upstream scene, not for creating the primitive itself.

Shape is dynamic. The geometry controls below it change with the selected primitive. Plane/Cube expose size dimensions; round primitives use Radius; Cone adds Top Radius; Cylinder/Cone expose caps; Torus adds Section; several round shapes expose angular and/or latitude ranges.

Lock Width/Height/Depth is not the same as final uniform Transform Scale. It controls the procedural dimensions before the object transform. When unlocked on supported shapes, independent dimensions become available.

Cube Mapping belongs specifically to Shape3D's Cube mode and determines how the incoming 2D texture is mapped. Shape3D Cube is still one primitive with one MaterialInput; the dedicated Cube3D tool is a different, more specialized workflow.

Start/End Angle removes part of the circular sweep for Sphere, Cylinder, Cone and Torus. Start/End Latitude further slices Sphere/Torus. These settings change generated geometry; they are not camera crops or 2D masks.

Bottom/Top Cap determines whether Cylinder/Cone end polygons exist. That is fundamentally different from face culling, which merely decides whether already-existing polygons are rendered based on orientation.

Torus Section controls tube thickness while Radius controls the overall ring radius. Treat them as independent geometric dimensions.

Subdivisions controls mesh density, not texture resolution and not a generic quality slider. More tessellation is useful for deformation, smoother vertex response or a large curved silhouette, but unnecessary subdivisions increase scene cost.

Wireframe is a geometry diagnostic in supported 3D render/viewer paths. It helps reveal tessellation before Bender3D/Displace3D; it is not a stylized 2D outline effect.

Visible is the hard object visibility switch. Unseen by Cameras is narrower: the object can remain in the scene while being hidden from direct camera rendering. Shadow behavior can depend on the selected renderer path.

Cull Faces Front/Back drops polygons based on facing direction. If an object disappears after a camera/object turn, inspect culling and normals before compensating with material opacity.

Suppress Aux Channels for Transparent Pixels controls whether fully transparent pixels contribute Z, UV, normals and other auxiliary channels. It is a technical pass control, not a normal alpha slider.

Affected by Lights, Shadow Caster and Shadow Receiver are object-level participation flags. Receives Lighting and Receives Shadows in the Material tab are material-level flags; isolate these layers one at a time when troubleshooting.

Is Matte turns the geometry into a spatial 3D holdout. It is not another opacity control: matte geometry can remove objects behind it in Z, so it should be used deliberately.

Hardware and Software Blend Mode belong to the 3D render path and expose different lists. The current captured Inspector shows a broad artistic hardware list and a shorter software list including Replace and matte modes.

Normals/Tangents Scale plus Show Normals/Tangents are diagnostics. They help read surface orientation and UV tangent basis without changing the actual mesh.

ObjectID and Material ID serve different auxiliary passes. ObjectID tags the Shape3D object; Material ID tags the material/shader. Planning both early makes later isolation masks much easier.

When a 2D image enters MaterialInput, it acts as a diffuse texture for the built-in basic material. When a full 3D material from a Phong/Blinn/CookTorrance-style chain enters the same socket, that external material becomes the surface shader instead.

Within the built-in Material tab, Diffuse defines the base response, Specular Intensity defines highlight strength and Specular Exponent defines falloff/tightness. Keep them separate while diagnosing the surface.

Transmittance is not the same as Opacity. Opacity controls visible surface transparency, while Transmittance concerns light passing through the material and transmissive shadows; Alpha Detail and Color Detail refine how texture information participates.

Two Sided controls back-side material/lighting behavior. It does not override face culling: a back side can have a two-sided shading model yet still be culled by object settings.

Transform operates on the already generated object: Translation places it, Rotation with an explicit order orients it, Pivot changes the rotation/scale center, and Scale resizes the finished object. Do not confuse these controls with procedural Size/Radius.

Use Target enables target-driven orientation. Import Transform transfers compatible transform data, but it does not import mesh geometry or replace FBX/Alembic import.

The clean data boundary is image/material → Shape3D.MaterialInput, scene3d → Shape3D.SceneInput, Shape3D.Scene → Merge3D/Renderer3D, and Renderer3D.Image → the normal 2D Fusion pipeline.

For simple motion-graphics scenes Shape3D is often faster and cleaner than an imported mesh. Once a shot needs complex topology, bespoke UVs or sculpted form, use an appropriate mesh source rather than forcing the design out of many primitive nodes.