Cameras and lighting
These tools emit scene contributions. Combine them with models using Scene Group, then preprocess and render the result; see 3D scene pipeline. They are distinct from the built-in camera controls of Model display and from QML cameras in Qt Quick 3D.
Camera
Emits a Scene containing a perspective camera and its transform. Eye is the camera position, Target is the look-at point, FOV is the field of view in degrees, Near/Far set clipping distances, and Roll rotates around the viewing direction. The initial Eye is (1, 1, 1) looking toward the origin.
Automate Eye and Target for a camera move. Keep the clipping range appropriate for the asset scale; a camera inside geometry or an overly large near plane can hide the model.
Camera Array
Emits six cameras around Origin, facing the cubemap directions +X, -X, +Y, -Y, +Z, -Z. Near and Far set their clipping range. The views are square cubemap faces, not six arbitrary camera positions.
Use this with a compatible multiview render pipeline for cubemap captures, reflection probes or point-light shadow views. A non-multiview consumer sees the active +X camera rather than an automatically stitched panoramic image. Backend multiview support and the renderer’s target configuration still matter.
Camera Switch
Accepts Camera 0–3 scene inputs. In Select mode, Index chooses one input. In Blend mode, Weights supplies four weights: negative values are clamped to zero, missing inputs contribute nothing, and usable weights are normalized.
Blend interpolates position, clipping distances and field of view, and blends orientation quaternions. Use camera-only inputs for predictable transitions; it is not a general scene morph. With no effective input weight the output is empty.
Light
Emits a Scene with a light. Mode selects Directional, Point, Spot, Rect, Disk, Sphere or Dome. Color and Intensity apply to all modes. Positional lights use Range (0 for unbounded falloff), with Falloff None, Linear, Quadratic (physical) or Cubic.
- Spot lights use Inner cone ° and Outer cone °.
- Area modes use Width, Height or Radius according to their shape.
- Position and Rotation place the light; rotation determines directional/spot orientation.
- Cast shadow, Shadow bias and Shadow normal bias author shadow settings.
A light component describes the light; the downstream shader determines which light shapes and shading models it implements. Creating a light does not make an unlit shader evaluate it.
Shadow Cascade Setup
Takes Scene In and produces Scene Out with directional-shadow cascade data. Set Cascade count (1–8), Shadow distance and Split lambda (linear/logarithmic split mix). Keep Camera near and Camera far consistent with the view camera. Light direction can override the scene’s directional-light direction.
This is setup data, not a complete shadow renderer. The patch still needs shadow-map rendering, a matching shadow texture resource and a shader that samples it. See Scene Resource Route in Environments and cubemaps and Render Pipeline for the resource/rendering side.