Skip to Content
3D & cameras3D layers

3D layers

Premation’s target is After Effects’ Classic 3D: flat planes positioned in real 3D space, projected through a real camera, lit and shadowed. It is a true 3D pipeline with a proper 4×4 matrix stack — not a perspective-looking hack — but it is not a 3D modelling application. Imported glTF models and a PBR shading mode exist as well; see Lights & materials.

Turning a layer 3D

The 3D switch is in the Transform section of the Properties panel. Turning it on adds three properties:

  • Position Z
  • Rotation X
  • Rotation Y

Everything else about the layer works as before, and the switch is reversible. New 3D layers answer lights and shadows without further setup.

What you get

FeatureNotes
Perspective projectionThrough the composition camera, with correct parallax and tilt
Z-sorting3D layers sort by depth rather than by stack order
Depth of fieldBlur driven by the camera’s focus distance
Per-fragment lightingLambert + Blinn-Phong. See Lights & materials
Material optionsPer-layer properties feeding the shading model
ExtrusionReal depth on 2D shapes and text
BevelsOn extruded geometry
Ground gridAt Y = 0, for orientation — drawn by the engine in the viewport only, depth-tested and with adaptive spacing
Axis widgetBottom-left, projected through the current camera
Auto-orientRotate a layer to face its direction of travel
Layer stylesDrop shadow, outer glow and the rest work on 3D and extruded layers, not only flat ones
Near-plane clippingGeometry derived from a 3D layer is clipped against the camera’s near plane, rather than folding back through it

3D primitives and 3D text

Beyond flat planes, Layer → New → 3D Primitive and the toolbar’s New Layer menu offer:

  • 3D primitives — cube, sphere, cylinder, plane, cone, torus, capsule and box. A primitive stays parametric: its Primitive section in the Properties panel holds its dimensions.
  • 3D Extruded Text — extruded type, with per-character 3D available so each glyph occupies its own position in space.
  • Imported models — File ▸ Import ▸ 3D Model… takes glTF / GLB, OBJ, FBX and USD (.usda / .usdz). See Lights & materials.

Extrusion and bevels

Give a 2D shape or text layer depth, then bevel the edges. Both are per-layer properties in the Properties panel and both are keyframeable — animating extrusion depth from zero is a clean way to have flat artwork “become” dimensional.

Draft 3D

A toggle that trades shading fidelity for interaction speed: ambient and key light only, with no shadows, ambient occlusion, depth of field or reflections. Use it while blocking out a scene, turn it off to judge the result. It affects the viewport only, never the export. See Navigating 3D.

Performance in 3D

Since 1.0 the 3D renderer caches each 3D layer’s effect chain at the layer’s own size, and reuses shadow maps, ambient occlusion and reflections between frames when the camera, the lights and the shadow casters have not moved. A layer that only moves reuses its effect chain rather than running it again. SVG layers are rasterised at the size they are displayed — including under a camera push-in — so they stay sharp. For a heavy lit scene, Composition Settings ▸ Renderer trades quality for speed — see Lights & materials.

Z-sorting vs stack order

The thing that catches people out: once layers are 3D, the layer tree no longer decides what draws in front. Depth does. A layer at the bottom of the tree with a nearer Z will occlude one at the top.

2D layers in a mixed composition still obey stack order, and sit at their position in the stack relative to the 3D group.

  1. Select a layer, turn on 3D.
  2. Duplicate it twice, and give each copy a different Position Z — say 0, −400 and −800.
  3. Layer → New → Camera…, pick a lens preset.
  4. Press C for the camera tool and drag — the first mode, Unified, orbits on the left button. The parallax between the three layers is the whole point of 2.5D.
  5. Press 1 to return to the Active Camera view.
Last updated on
Shapes & masks from text and SVGCameras