Shadow maps use the full scene-derived size (point-light faces capped
at 2048) under a 26M-texel scene budget halved largest-first; the PCF
filter measures its reach in output pixels with a two-texel minimum,
samples texel centres on a jittered adaptive 5..10 grid, and compares
each texel against the receiver plane's own depth so wide kernels never
self-shadow lit casters. The WGSL mirror is parenthesised (Tint rejects
the old form) and guarded by a shader-wide grammar test; the CPU
reference and shader remain step-for-step parity-tested.
On an adjustment layer Motion Tile kept the comp-sized buffer whatever the
Output Width / Height, and the adjustment draw dropped any stack result whose
size differed from the canvas. After Effects grows (or crops) the comp
rectangle there as on any layer, so a Transform after it brings the extra
tiles into view, as in the frames attached to #1727 and #1409.
- motion_tile_frame: one output frame for the CPU and GPU effect; on an
adjustment it grows about the comp rectangle on whole pixels, keeps the
buffer at 100 % and past 16 384 pixels a side.
- adjustment_crop: the CPU and GPU adjustment draws put a stack result whose
frame changed back onto the canvas instead of dropping it (Enc::crop on
the GPU).
Tests: the #1727 After Effects frames (910 x 340 % and 90 x 340 % then
Scale 12 %), tiles past the frame and cropping on CPU, GPU vs CPU for
adjustment and full-frame layers.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Review: styled() kept the buffers it rendered instead of the ones LayerCache::insert returned, so
without a blending-space conversion concurrent misses still returned separate body and pass
buffers (and the GPU uploaded each).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Preview workers that start together miss the layer cache together, and each kept its own copy
of every layer; on the GPU each copy was uploaded and packed into its own Classic 3D atlas, which
ran the device out of memory on a cold start. LayerCache::insert now keeps the first buffer of a
key and returns it, and the renderer goes on with the returned buffer, so the workers share one.
Renders don't wait for each other: on the shared frame pool a blocking wait could end up beneath
the work it waits for (tried, it stalled playback for the wait's timeout).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Review: mask motion blur draws an animated mask at the shutter's other times, but the layer key
hashed its value at this time only. Two frames with the same mask value (a hold before a move)
shared a cached layer, with the wrong blur. This applied to solids, text and shapes already, and
to stills since this PR.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Footage layers had no layer cache key, so a still made a new layer buffer every frame on every
preview worker, and the GPU compositor uploaded it again each time (its upload cache is keyed
by buffer). Stills (without a proxy) are now keyed like text and shape layers, without the
source time; movies and sequences are unchanged.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
After Effects 2026 in a 32 bpc project, Levels (Individual Controls) with RGB
(0.1, 0.9, gamma 1.5, 0.05, 0.95), Red Gamma 0.8, Green Input White 0.8 and
Blue Output Black 0.2 on grey steps 0..255 (case 904 E07): every step matches
"each channel's controls, then the RGB controls" with nothing clipped
(values below Input Black end below Output Black, above Input White above
Output White). EffectCraft ran RGB first and clipped "Off for 32 bpc Color"
(it could not see the bit depth): up to 27/255 off on 95 % of the grid; now
0 grid points off, max 1/255, CPU and GPU.
EffectEnv gains float_depth (the renderer sets it for 32 bpc projects), so
"Off for 32 bpc Color" clips in 8 / 16 bpc projects as before. Unclipped
values below 0 take the gamma mirrored through 0. Plain Levels keeps its
order (its channel controls were not measured); its clipping follows the same
rule. EffectCraft's own projects default to 8 bpc and are unchanged.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
After Effects 2026 (sRGB working space, not linearised) maps red, green,
blue and 50% grey solids at Tint's defaults to 76, 150, 29 and 128/255:
Rec. 601 weights on the working-space values, not Rec. 709. CPU and GPU
kernel changed; tests pin the four greys within 1/255.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
With Units set to Index the selector already counts characters (words,
lines), but the timeline kept showing Start, End and Offset as
percentages ("3.0%"), so switching looked like it did nothing. They are
now shown without the % sign while Units is Index
(render::text::in_index_units).
Synthetic bold grows a glyph's outline in its own space, so a
non-uniform character Scale stretches it more along one axis. The patch
and layer bounds reserved it through mean_scale and could clip the
stretched side; they now reserve how far the bold reaches along x and y.
A run of solid fills grows its glyphs together in layer space by one
radius, which only matches when the bold stays round there: a bold glyph
whose matrix stretches or skews is no longer gathered into a run.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Replacing a virtual file bumps a counter the UI lookup reads. The content hash runs on that miss, from a callback, so the frame path does not stat every footage item or hash on every store write.
The layout marks a glyph synth_bold when Faux Bold is on, or when a bold
style is asked of a family without a bold face, but the renderer never
read it, so both drew as regular.
PlacedGlyph now carries the bold radius (0.03 em, which brings Inter
Regular's stems to Inter Bold's). A bold fill is the outline's coverage
merged with a round stroke twice that wide, and a bold glyph's stroke is
the band between the outline grown by radius ± half the stroke width, so
it follows the bolder shape. Runs of solid fills only merge glyphs of the
same bold radius, and the patch and layer bounds grow by it.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
The Advanced 3D renderer folded Accepts Shadows (Off/On/Only) into a
bool, so a shadow catcher set to Only drew as an ordinary lit surface.
The material now carries shadows_only. Such a fragment draws black at
the strongest shadow a light casts on it, as Classic 3D does, and it
goes in the blended pass so what is behind it shows through. The CPU
shader and advanced3d.wgsl get the same branch.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
draw_bounds padded every stroke by width × miter limit / 2 on each side, joins or
not. A straight 1080 px wide wipe used as a track matte became a 4326×6366 layer
buffer (~440 MB as RGBA f32), made anew every frame while its trim animates, then
cached on the CPU and uploaded to the GPU. With 12 frames exported in parallel a
16 GB Mac ran out of memory and froze.
Now the miter limit applies only to miter-joined paths with a join (two or more
segments, or closed); square caps reach √2 half widths, other strokes one; Taper
and Wave keep the widest padding. Rendered pixels are unchanged.
A full-res frame of the affected ad: 1.7 GB → 0.7 GB peak.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A layer selected by another layer's effect (Displacement Map, Blend...) had
its structure, values, matrix and time hashed into the owner's key, but not
the comp camera/lights its own scene-reading effects (Caustics, CC Particle
World...) see. The owner's key included the scene only when the owner itself
ran such an effect, and a cached owner never runs the selected layer's
effects, so moving the camera or a light reused stale owner pixels.
Dependencies::layer now hashes comp_scene for every selected (or nested)
layer that reads it, with the same policy as owner keys.
Fixes#1252
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01N9kwMcw2pYQ5fSjKa3ZPnh
Guide layers render only in their own comp's viewer: a nested renderer (precomp, collapsed
precomp in 2D, Classic 3D and Advanced 3D, comps read by effects, aux passes, the GPU walk's
resident precomps) now runs with guides off (Renderer::nested_opts), so a precomp's guide
layers no longer show in the parent comp's viewer or in a "Current Settings" render.
The Timeline layer and Composition viewer right-click menus (shared layer_menu) gain a
Guide Layer item (a plain toggle, no check mark) acting on the selection, and guide layers show a guide mark
before their name in the Timeline.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
In After Effects solo is not time-dependent: while any AV layer is soloed,
non-soloed layers are excluded at every time. The renderer, nested/collapsed
3D paths, engine viewer helpers and viewer hit tests only counted solo layers
active at the current time, so outside the soloed layer's in-out range the
whole comp reappeared. Drop the time condition (audio already ignored it).
Update the #66 viewer regression test to the AE behaviour and add a render
test for it.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Expansion was a fill plus a stroke 2·|expansion| wide, blended per pixel. Once the
expansion passed the shape's curvature radius the stroke's inner edge inverted: a
20 px circle expanded by 60 left a ring with the layer showing through the middle,
+600 cut only the original dot, and negative values left islands. The outline is
now offset with Offset Paths' geometry (round joins) and filled.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A Merge Paths item appended to a shape layer's Contents cleared every draw
collected so far, including fills and strokes inside groups above it, so
the layer's shapes vanished from the render. AE's rule is per group: only
paint items in the merge's own group above it are dropped. Draws now
record which contents level created them; the merge drops only its own
group's paint, and the group-bubbled draws keep rendering.
Adds a regression test (Merge Shapes and Add modes) that renders nothing
without the fix.
Stretched layers copied the nearest source sample for each output sample, which adds error: at
48 kHz the worst error on a 440 Hz tone was 0.025, and 0.15 at 8 kHz. They now blend the two
source samples either side. Layers at 100% and 50% speed keep their exact samples.
The render tests cover a plain precomp, which mixes the same audio as its layers, and the stretch
error against the exact tone.
Co-Authored-By: Claude Haiku 5.5 <noreply@anthropic.com>
Echo, Posterize Time and the other Time effects read the layer at other times with none of
its effects. After Effects renders them with the effects before them: SceneYard "Task
Preparation UI" applies Echo after CC Particle Systems II on black solids, and AE shows the
echoed fireworks over a white page where we echoed the black solids and blacked it out
(Posterize Time after Turbulent Displace, the usual stop-motion setup, depends on the same).
The host already bounds the count to the effects before the one rendering.
Mean abs error vs AE: Task Preparation UI 0.1855 -> 0.0071, Lightbulb Idea 0.0126 -> 0.0102;
the 15 other corpus projects with a Time effect after another effect are pixel-identical.
(cherry picked from commit 46870452db2656c72406f63245e8b1609ceb04d4)
(cherry picked from commit 334d102582e5a64db32e2f3e55836d55ac96bdaa)
(cherry picked from commit 734da3e35dc57c01b73823458bdd87611ac1bf99)
(cherry picked from commit 87a733e076)
aiongg