Extruded text keeps per-character 3D: each glyph's outline is extruded in its own space and
placed by its animated matrix; shape layers extrude their fill outlines (gradients use their
mid colour). Shadow-map bias scales with the texel footprint at the receiver and the PCF
reach (no acne on curved casters), identically on CPU and GPU. Docs: Advanced 3D in
architecture.md, parity table updated (3D, Text).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Aux-channel model (effectcraft_raster::channels3d: depth, object/material
ID, normals, UV, coverage, Cryptomatte ranks + manifests, MurmurHash3);
EffectHost::aux; multi-layer OpenEXR channels via the exr crate
(FootageSource::aux); the compositor's depth / layer-ID / Cryptomatte pass
(Renderer::comp_aux) so the effects work on precomps of 3D comps.
3D Channel Extract, Cryptomatte, Depth Matte, Depth of Field, EXtractoR,
Fog 3D, ID Matte, IDentifier.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Advanced 3D comps render their 3D runs as one scene of world-space triangles: model and
primitive layers (meshes, skinning, animation clips by layer time), extruded text/shape
layers, and textured cards for every other layer with their Material Options. Shading is
physically based in linear light: Cook-Torrance GGX + Lambert for parallel/spot/point lights
(AE falloff and cone), ambient lights, image-based light from an Environment light's layer
(equirectangular; cosine irradiance map + roughness mip chain + split-sum BRDF fit), and
PCF shadow maps (spot/parallel/point cube faces). 2x2 SSAA, depth of field as a gather blur
by circle of confusion, back to the working encoding.
The software rasteriser (depth buffer, top-left rule, perspective-correct, opaque visibility
then sorted transparent blending) is the reference and CI path; effectcraft-gpu implements
Accelerator::raster_3d with a wgpu render pipeline (advanced3d.wgsl) and matches it within
tolerance. Environment layers are hidden from both renderers.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
8/16 bpc no longer re-quantises the whole comp after every layer: 2D layers
report the canvas region they can change (transformed buffer bounds plus the
resampling margin; stencil modes, motion blur, perspective, styles,
adjustment and 3D runs fall back to the full canvas), and only that region
is quantised. A quantised buffer placed pixel-exactly (integer translation,
Normal, 100%) over a blank canvas needs nothing; invisible layers need
nothing; collapsed precomps quantise per nested layer. The redundant final
comp pass is skipped unless the comp is re-encoded. Solids are filled with a
pre-quantised colour and skip the source pass unless masked. Row-parallel
vectorisable kernel. Debug builds verify every pixel outside the region is
already quantised (whole workspace test suite run with 8 bpc forced: no
violations).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Collapsed precomps draw their layers straight into the parent with
concatenated transforms and multiplied opacity (nested blend modes and
adjustment layers act on the parent); a non-Normal mode, track matte or
Preserve Transparency composites the collapsed result as a whole; masks,
effects or styles force a flattened render. Nested 3D layers use the
parent's camera and lights and join the parent's 3D run when the precomp
layer is 3D. Continuously Rasterize re-rasterises text and shape layers at
their on-screen scale. Draft quality samples nearest-neighbour, Wireframe
draws the layer bounds. layer.slip / slipBack (Alt+PageUp) / slipForward
(Alt+PageDown) and dragging the timeline source bar slip the source under
fixed in/out points, clamped to the source duration; undoable.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
8/16 bpc projects clamp and quantise layer pixels after source/masks and
every effect, and the comp after every layer (16 bpc = 0..32768); 32 bpc
keeps over-range values. Blend modes without HDR support clamp their inputs
in 32 bpc. Working space (sRGB, Rec. 709, Rec. 2020, Display P3; matrices
derived from the published primaries), footage colour profiles from stream
metadata or Interpret Footage, Linearize Working Space and Blend Colors Using
1.0 Gamma. Layer cache keys include the colour settings. Project panel depth
button: click cycles, Alt-click opens Project Settings.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Range selector complete (units, based on, modes, shapes, smoothness, ease
high/low, randomize), Wiggly and Expression selectors combined by mode;
all AE animator properties incl. line anchor, character offset/value,
skew axis, tracking type, fill/stroke HSB+opacity, blur; More Options
(anchor grouping, fill & stroke order, inter-character blending); Path
Options (text on a mask path); per-character 3D rendered as planes in
the 3D compositor; text animation commands and original presets.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Resolve conflicts with perf (layer cache / Arc<Buf>), export (RenderOpts.view: None
for output), timeline (separate dimensions + one-node POI visibility), mask/pen
viewer gestures (alongside camera tools and view-aware layer matrices), key
dialogs (alongside Camera/Light Settings), render queue + MCP command modules.
3D path tangent honours separated position; DOF blurs a per-frame copy so cached
layer buffers stay shared. New test: cached renders match uncached after camera,
focus, light and 3D view changes. Fix a clone-on-copy lint from main.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- layer.autoOrient {mode: off|alongPath|towardsCamera|towardsPointOfInterest}
(Layer > Transform > Auto-Orient…, Cmd+Alt+O); render support for Orient
Along Path (2D and 3D) and Orient Towards Camera (with recursion guard).
- Layers parented to cameras/lights follow their full rotation.
- Draft 3D turns off lights and shadows (and DOF).
- Turning the 3D switch off discards Z values, Orientation and X/Y Rotation.
- Tests for all of the above.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Brandon Thomas