Reload Footage empties the RAM preview, but an idle pass of
persist_missing may already hold a frame from before the reload. It
took the cache's drop generation when it ran, after the drop, so the old
pixels could be written back and, with the file's size and time
unchanged, served again after a restart.
Each RAM frame now keeps the disk cache its render started with and
that cache's drop generation then. persist_missing writes a frame only
to that cache, with that generation, and skips it once one of its files
was dropped since; a frame rendered without that cache is not written
from there.
Tests use tempfile directories (the layer drop test, the frame_files
folders under target/).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- A reload while the disk cache is off is kept on the session
(pending_disk_drops) and applied when the cache is switched on, before
anything is read from it.
- Each render takes the cache's drop generation before it reads (frame
worker; layer cache at the miss). An entry listing a file dropped after
that is refused under the queue's lock, so a render that was under way
during the reload can't store pixels from before it. Such a frame is
shown with an incomplete file list, so persist_missing skips it too.
The per-file record is bounded (past it, everything older is refused).
- An entry read while its drop was pending is a miss even when the
writer ends the sweep between the read and the check (a stale hit
found under load): the pending drops and the generation are taken
before the entry is read.
- Test: a reopened project serves its frame from disk with no source file
opened or decoded.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
An empty first paragraph draws no glyph, but in a box its face sets the
first baseline that places every later line. The read list took faces only
from glyphs, so it could be complete without that file. Note each style
run's resolved face as well.
The font-reads and confine tests use tempfile::tempdir(): a fresh
directory per run, removed on panic too.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- Buffers made under `Renderer::recorded` (layer content, a Time effect's input, an adjustment
layer's footprint) evaluate their masks, shapes and effects with the context that notes into
the buffer's own read log, so a file an expression reads (`sampleImage`, also at shutter times
mask motion blur adds and the cache key doesn't see) is in the buffer's files.
- A styled layer served from the cache lists the files of the cached body and passes too, so
the buffers made of it (flattened, blending space) don't get a complete list when one of them
came with an incomplete one.
- A character put in by Character Offset or Character Value without a relayout notes the font
it is drawn with; a fallback face (or none) leaves the list incomplete.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A mask with Mask Motion Blur On blurs even when the layer's own switch is off,
as long as the render's gate is open (Render Settings motion blur, and with
Switches Affect Nested Comps the precomp layers above). The key held the
layer's own motion blur state only, so a buffer blurred with the gate open was
reused with it closed (and the other way round), also from the disk cache.
Masked layer, input and footprint keys now tell a closed gate apart.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A layer whose effect reads another layer (Gradient Wipe, Set Matte, Blend…) gets
that layer drawn by its own renderer. Through a precomp layer, Switches Affect
Nested Comps, the switches inherited from the precomp layers above and motion
blur on change those pixels, but the key held only the layers' own switches: a
buffer made with the setting on was reused (from memory or the disk cache) with
it off. Such layers now derive their key with what the renderer passes down.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Layer entries share the frame entries' file list layout, which now stores the
full i128 modification time: the layer entry format version goes to
layer-entry-files-2. The module doc names the font files text is drawn with
among an entry's source files, and what keeps a list incomplete.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A frame entry's file list stores a file's modification time as i128 nanoseconds
since 1970 (negative before), as the file stamps now keep it; the frame entry
format version goes to frame-entry-files-2 so entries in the old layout are not
read.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
EvalCtx carries the render's read log. A sampleImage render records its footage
reads, keeps them with its per-thread cached buffer and notes them in the
evaluating render's log on every use. Text notes the font files it is drawn with,
with the identity their bytes were read with; a family that isn't installed, a
fallback or missing glyph, or a font from memory leaves the list incomplete.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
LayerCache::insert returns the files of the buffer it keeps, and Renderer::keep
notes them in the render's log and returns them, so a render that uses another
worker's buffer lists that buffer's files (a file changed between the two reads
then has two identities and the list is incomplete).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Footage outside the read roots shows nothing and was noted as no file at all, so a
picture made without it had a complete, empty list and could be kept on disk for
a later session with other roots. A refused read during `recording` now marks the
list incomplete.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A file's modification time was nanoseconds since 1970 as u64: a time before 1970
read as "no time" (two such files of one size looked alike) and times after 2554
were truncated. It is now i128 nanoseconds, negative before the epoch; None only
means the file system has no time. The content-key hash writes a presence tag and
the full value.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
File > Reload Footage now deletes the frame and layer entries whose file
lists name the reloaded item's files (DiskCache::drop_files), so a file
replaced by one with the same size and modification time, which the
check on hit can't tell apart, is read again. The index knows keys, not
files: the writer thread reads the files section at the start of each
frame and layer entry (not the pixels) after the writes queued before
it, and deletes the entries that list one of the files. Until it is
done, an entry made of one of them is a miss and frames aren't linked to
new keys. The entries are gone for good, so they don't come back after a
restart. Automatic reloads don't drop anything: they follow a changed
stamp, which the check on hit catches. The browser needs nothing (its
file identities are content hashes).
Removed: disk_cache::seal, which steps 2 and 3 left without callers (the
tests use seal_format). Updated the module, seal_format and forget_media
docs, and docs/architecture.md / docs/web.md describe the final design:
what is cached, keys from memory, file lists checked on hit, nothing
scans the project, incomplete lists stay in memory, Reload Footage drops
entries, edits keep the frames they can't change.
Tests: render unit tests (same size and time passes the check but not
the drop; other files' entries stay; an entry queued before the drop is
dropped; a layer entry too; written again afterwards it is used; the
drop survives a restart; a pending drop is a miss), and an engine test
through file.reloadFootage.
Closes#3184. Part of #1492.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
With an expression anywhere in the project, every comp's disk key hashed
every item of the project, so any edit (another comp, an unused footage
item) threw away every comp's frames on disk. The key now covers what the
comp can draw plus what the expressions among it can reach
(Project::pixel_reach):
- without expressions: the items the comp draws, without silent layers,
plus the footage its precomp layers swap in (Media Replacement);
- with expressions: every layer of the comps they reach, their markers,
the comps and footage they name with string literals (comp("Name"),
footage("Name")) together with which items have those names now, and
the footage named like a reached source (its data is read by name);
- the whole project when an expression computes a name, uses the object
model's constructors or internals, dynamic code, computed keys or
writes, or text the scanner can't read for sure (regex literals,
template strings, escapes): expression_reach errs towards reaching more.
Markers count only where an expression can read them (#1083). Memory
only, no file is looked at.
Tests: scanner and reach unit tests in the project crate; engine tests
for the key (own-comp expressions keep the key through edits elsewhere,
comp("Other") follows Other and not a third comp, a second comp taking
the name counts, a computed name counts every item, markers read from
another comp); a ui-egui test that the disk frames outside an expression
layer are carried when the comp it names changes.
Part of #3184. Part of #1492.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* Layer: Un-precompose a selected pre-comp layer back into its parent composition (#2918)
layer.unprecompose copies the nested composition's layers into the parent at the pre-comp
layer's place with timing and position carried over, and refuses what it can't keep exactly.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
* Un-precompose: refuse a pre-comp layer with Solo or Guide on
A soloed pre-comp layer hides the comp's other layers and a guide layer is
left out of renders; the expanded copies carry neither switch, so the picture
would change. Listed with the other blockers, in the command description and
the agents.md Cookbook entry; the refusal test covers both switches.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* Un-precompose: warn instead of refusing, only a locked layer stops it
A locked pre-comp layer is still refused, with or without `force`. Every
other former refusal reason is a warning: without `force`,
`layer.unprecompose` returns an error that lists the reasons and changes
nothing. `dryRun` returns `{dryRun, warnings}`. With `force` the layer is
unpacked in one undo step and the result carries `warnings`. Whatever the
pre-comp layer itself carried is dropped, and other layers' parent, track
matte and layer-property links to it are removed.
Layer > Un-precompose in the UI lists the reasons in a prompt with
Unpack Anyway / Cancel (`dialog.unprecompose.force` /
`dialog.unprecompose.cancel`). A layer without reasons unpacks at once.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* Un-precompose: warn about expressions that refer to the pre-comp layer
An expression on another layer of the comp that names the pre-comp layer
would no longer find it once the layer is unpacked. The expression can
name it in quotes or call `layer(<index>)` with its index, bare or on
thisComp; another comp's comp("…").layer(…) is not counted. Each such
property is now listed once as a warning. Like the other warnings,
`force` unpacks and `dryRun` lists it. The expressions are left as
written.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* Un-precompose: warn about lost frame sampling and Position expressions, keep markers in order
Review follow-up (CodeRabbit on #3081):
- A nested comp that preserves a frame rate other than the parent's
samples its layers on its own frames, and the expanded copies would
lose that. Motion-blurred nested layers under a different shutter
would also blur with the parent's shutter. Both are now warnings.
- When the pre-comp layer's Position differs from its Anchor Point, the
move is added to the stored Position of the unparented nested layers.
An unparented layer whose Position, or a separated axis, has an
expression may ignore that move, so it is now a warning.
- Copied comp markers are sorted into time order with the comp's own.
- New tests: a 10 fps nested comp with Preserve Frame Rate in a 30 fps
comp, a different shutter, a Position expression with and without a
move, and nested matte and layer-value remapping with shifted markers
merged in order.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* Un-precompose: warn when an expression's layer numbers shift; the prompt keeps its target
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* Un-precompose: a shorter renumbering warning
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Sonnet 5.5 <noreply@anthropic.com>
Co-authored-by: FredApps <153173854+FredApps@users.noreply.github.com>
* Image sequences: a gap in the numbering holds the previous frame, as After Effects
Split out of #1876 (originally 7850cbf4), because it reverses the colour bars
documented in #297 and needs its own discussion.
After Effects 2026 shows the last existing file across a gap in a sequence's
numbering (files 1, 2, 4, 5 show frame 2 at frame 3; read back from After
Effects). EffectCraft drew colour bars there. The media pool now holds the
previous file (Footage::sequence_file_held); colour bars remain only where no
file precedes the frame. Missing-frame reports on import are unchanged.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* Image sequences: hold the first file of the preceding number across a gap; fix the import warning
A gap now holds the file sequence_file picks for the preceding number, so
files sharing a number do not switch at the gap. The Report Missing Frames
warning says the previous frame is held instead of "shown as placeholders".
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* Image sequences: Interpret Footage says a gap holds the previous frame
The Interpret Footage note for a numbered sequence with gaps still said the
missing frames are shown as colour bars; the media pool holds the previous
frame there now. The note, its comment and two test doc comments say so,
and a forms test checks the note's wording.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: FredApps <153173854+FredApps@users.noreply.github.com>
* Echo: swap Composite In Back and Composite In Front (#3124)
After Effects 2026 renders of the importer sample (opaque test card with a
moving square, one echo at -1/30 s) show Composite In Front keeping the
layer on top (the output equals the current frame) and Composite In Back
putting the echo over the layer. EffectCraft had the two the other way
round, which made ~0.5 % of the pixels differ at the moving edge. The CPU
and GPU (fxn_echo) arms are swapped together.
A regression test rebuilds row 130 of the test card and checks Add,
Maximum, Minimum, Screen and both Composite operators against the measured
After Effects row. Blend is left unchanged: the sample (one echo, intensity
1, decay 1) cannot tell its normalisation apart.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* Echo: Blend sums the weighted frames like Add (#3124)
New After Effects 2026 renders (our own 1110-extra sample, 16-bit PNGs,
opaque test card with a moving square) show that Blend does not normalise:
- 3 echoes at intensity 1 and decay 1 give four times the layer.
- 2 echoes at intensity 0.5 and decay 0.5 give alpha 0.875
(0.5 + 0.25 + 0.125).
EffectCraft divided by the number of frames. The division is gone on the
CPU and in fxn_echo_fin on the GPU.
The same sample pins Composite In Back with three echoes: each older echo
goes over what is there with its weight as opacity, so the oldest is on
top, as the branch already does. Composite In Front with three echoes
equals the current frame.
A regression test checks four measured 16-bit rows within 4 codes (24 on
square-edge pixels): Blend 3/1/1, Blend 2/0.5/0.5 with alpha, and both
composite modes at 3 echoes, -0.08 s, decay 0.5.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* Echo: doc comments on the CPU and GPU entry points
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: FredApps <153173854+FredApps@users.noreply.github.com>
Part of #3184 (step 3 of 4).
- A frame-blended footage frame's layer-cache key now hashes the
sequence's file list: with no project-wide salt, two projects whose
items share an id and a first frame must not share its disk entry.
- Test: a frame worker syncing, patching and rendering a project with an
unused 5000-frame sequence asks its file hook only about the file its
buffer reads (the salt asked about every file at every sync).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Part of #3184 (step 3 of 4).
Layer buffers on disk (and in the browser's layer store) now carry the
files they were made of, like frame entries since step 2, and are checked
against exactly those files when used. The project-wide footage salt is
gone: nothing stamps every footage file any more.
- Layer entry format 4: the buffer's EntryFiles, then the pixels. A hit
checks only the listed files (metadata, or the VirtualFileId hook for
files from memory; Missing must still be missing); a change, a damaged
entry or the old format 0 is a miss and the entry is deleted, so the
re-rendered buffer replaces it. A hit reports its file list, so frames
made of buffers from disk are complete and get stored again.
- Buffers whose files can't all be named are not written.
- Layer disk keys: layer_disk_key = hash of the layer entry version, the
renderer version and the layer cache key (which names the footage items
with their paths and settings, from memory).
- PrefetchStore checks on use; an entry made of changed files is dropped
and forgotten, so the new buffer is queued over it. FrameServer passes
its virtual-file hook to its layer cache.
- Removed: footage_salt(_with), file_stamp, footage_stamp,
set_disk_salt, set_store_salt, Session::refresh_disk_salt,
poll_disk_salt, footage_fingerprint, disk_salt, disk_salt_background,
disk_salt_job, DiskSaltJob, the UI's per-frame poll, the salt in
FrameServer and in the availability worker (with its cache_enabled
plumbing). The disk cache is attached as soon as it is configured.
- Docs: architecture.md, web.md.
Tests: layer entries check their files (size, time, missing/back,
folder, damaged, old format, incomplete); virtual files; prefetch store;
media tests with a real pool across restarts (only the changed file is
read, projects with the same item ids never share an entry, missing,
unreadable and damaged never serve stale buffers); engine tests that the
cache is attached at once and edits do no file work (#1926, #3106).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* 3D: Orientation and X/Y/Z Rotation turn Z first, keyed Orientation slerps, as After Effects
After Effects composes a 3D layer's (and a camera's) Orientation and its X/Y/Z
Rotation as Rx * Ry * Rz each, Orientation before the axis rotations; EffectCraft
used Rz * Ry * Rx. Measured on synthetic AE 2026 cases: Orientation [10,20,30]
(488) was 5.2 % of the frame off and now matches to the anti-aliased edge (0.03 %);
X/Y/Z Rotation 10/20/30 (800), Orientation + X Rotation (801) and one-node
cameras (804, 805) likewise.
Keyed Orientation moves along the shortest arc between the keys (quaternion
slerp, eased in time, angles written in [0, 360)): keys [0,0,0] -> [10,20,30]
read back [6.2643, 9.3531, 15.2190] half-way where EffectCraft gave [5,10,15];
[0,0,0] -> [60,80,40] linear and eased match AE's read-backs within 0.002 deg.
lookAt() returns AE's angles (z axis on the target, x axis level; two AE
read-backs within 1e-3 deg). The camera tracker, VR editor, Look At and orbit
commands use the new decomposition.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* Keyframes: a linear side next to an eased side has 16.67 % influence, as After Effects
In a segment that is Bezier on one side and linear on the other, the linear
side keeps the segment's average speed, but with the 16.67 % influence After
Effects shows for linear keys in the Keyframe Velocity dialog (and reports to
scripts), not 33.3 %. Measured on an After Effects 2026 time-remap render (a
square at x = 50 + 455 * remap): out of a key at 2 s (3.5, speed 0, influence
50 %) into a linear key at 4 s (0.5), AE plays 3.290 at 2.5 s and 2.657 at
3 s; EffectCraft played 3.308 and 2.629 (8 and 13 px off), now the same
square positions to the pixel.
LINEAR_SIDE_INFLUENCE is used by evaluation (scalar, per-dimension, paths and
spatial), side_ease (Graph Editor, velocity dialog, keyIn/OutTemporalEase) and
the Lottie export, so exported easing plays as EffectCraft does. Segments
linear on both sides stay straight lines.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* Timewarp: retime the layer's source without the effects above it, as After Effects
After Effects' Timewarp reads the layer's source (with its masks) at the
retimed times and ignores the effects applied before it: a red solid with a
Gradient Ramp and then Timewarp (Speed 25) renders as the red solid in AE
2026 (EXPORTER e2 read-back). EffectCraft retimed the layer with its
effects (mean difference 127.5/255 over the comp), now identical (0, CPU and
GPU). Echo, Posterize Time and the other time effects keep reading the
effects before them, as measured before.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* Image sequences: a gap in the numbering holds the previous frame, as After Effects
After Effects 2026 shows the last existing file across a gap in a sequence's
numbering (case 860: files 1, 2, 4, 5 show frame 2 at frame 3; VERIFY
read-back). EffectCraft drew colour bars there. The media pool now holds the
previous file (Footage::sequence_file_held); colour bars remain only where no
file precedes the frame. Missing-frame reports on import are unchanged.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* Review: keyed Orientation uses LINEAR_SIDE_INFLUENCE for linear sides; real NaN assert
evaluate_orientation still gave a linear side (and the unreachable missing-key fallback)
1/3 influence while segment_value, side_ease and the Lottie export use
LINEAR_SIDE_INFLUENCE (1/6), so a keyed Orientation next to an eased key moved on a
different curve than every other property (CodeRabbit on #1876). New test:
orientation_linear_side_next_to_an_eased_side_matches_a_scalar (both directions).
The always-true NaN assert in keyed_orientation_slerps_as_after_effects now checks that
a non-finite key comes back as given; Property::value_at says why the match_id
"orientation" routing is safe.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* Format keyframe orientation test (cargo fmt)
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
* Move the image-sequence gap change out of this branch
The gap change (hold the previous file instead of colour bars) reverses
#297's documented behaviour and goes to its own PR, as requested in review.
pool.rs, sequence.rs, the footage test and docs/footage.md / docs/agents.md
are back to main.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: Codex <noreply@openai.com>
Co-authored-by: FredApps <153173854+FredApps@users.noreply.github.com>
* SAM 2 mask decoder: transposed convolutions as matrix product plus pixel shuffle (#3031)
Both transposed convolutions of the mask decoder have a 2x2 kernel, stride 2
and no padding, so each input pixel owns its own 2x2 output block. That is one
matrix product plus a pixel shuffle, which is about twice as fast on the CPU
as candle's conv_transpose2d. The first convolution also takes the decoder's
key tokens directly, without the transpose and copy to an image.
The candle path stays in the code for the tests. They compare the two to
1e-6 (relative to the largest output value) on the decoder's real sizes,
odd and degenerate sizes, large values, zeros and a single hot pixel, check
that other kernels are refused, and compare a whole decoder forward pass
(single and multi mask) on random weights.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
* SAM 2 decoder test: bound the weight discovery loop
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Sonnet 5.5 <noreply@anthropic.com>
Co-authored-by: FredApps <153173854+FredApps@users.noreply.github.com>
* SAM 2 tracking: prompt frames in frame order, size-check frame reused (#3030)
With several objects prompted on three or more frames, the prompts went to
the tracker object by object. The tracker keeps features for two frames, so
every prompt after the first object fetched and encoded its frame again.
Each prompt frame was also fetched once only to check its size and then
fetched again for the encoder.
Prompts now go to the tracker in frame order (then object order), and the
frame fetched for the size check is handed to the tracker the first time it
asks for that frame (up to 256 MB of pixel data kept; beyond that the frame
is fetched again as before). Objects do not influence each other, so the
results are unchanged.
Tests cover the prompt order, one host fetch per prompt frame, the order of
several prompts of one object on one frame, and the memory cap.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
* SAM 2 tracking test: the held-frame cap is cumulative
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Sonnet 5.5 <noreply@anthropic.com>
Co-authored-by: FredApps <153173854+FredApps@users.noreply.github.com>
Part of #3184 (step 2). Fixes#3175: opening a comp no longer stamps every
footage file of the project before its first frame can show.
- The comp's content key (`disk_cache::comp_content_key`) is taken from the
project in memory alone: no footage or proxy file is stamped, no
`VirtualFileId` is asked. `comp_content_key_with` is gone; `frame_key`
hashes a new entry version, so entries of the earlier format are never
looked up again (LRU evicts them).
- Frame entries (formats 2 and 3) store the files the frame read
(`EntryFiles`: size and modification time as read; size and content hash
through `VirtualFileId` for files from memory; missing files), then the
pixels. Formats 0 and 1 read as misses and are deleted.
- Check on hit: `get_frame`/`get_frame_f16` look only at the listed files
(one metadata call each) on the frame worker. A changed, back or unreadable
file, or a damaged entry, is a miss: the entry is deleted and the frame
rendered and stored again.
- A frame whose file list is incomplete is shown but not stored. The RAM
preview keeps each frame's files, so `persist_missing` writes only frames
with a complete list.
- Remote frame worker: `FrameServer` records each frame's reads
(`take_files`); the web worker stores them with the entry and the page
checks them through the file table before using it. The browser's media
epoch existed only for the old key and is removed. Export does not use
frame entries.
- The cache indicator still only asks whether an entry exists.
Tests: entry format round trip and damage, the check per file, and in the
viewer: a changed file re-renders exactly its frames (and a reopen reads the
rest from disk); a frame checks only its own files in a project with a large
unused sequence and the key looks at no file; missing, unreadable files and
damaged entries never serve a stale frame.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* av1enc: motion search with a visited set and early-exit SAD, fixed-size SATD, region-only BlockDecoded snapshot
- motion_search remembered nothing: the diamond search evaluated the previous
centre again and every SAD ran to the last row even when it had already
lost. Full-sample and sub-sample positions are now tracked in a visited
list (costs are deterministic, so a position that was evaluated cannot beat
the best cost again), and the SAD stops after a row once sad + mv cost can no
longer be below the best cost. The stop test is the same f64 expression as
the strict "<" the callers apply, and it is monotone in the SAD, so the
chosen vector is the same.
- satd runs with a fixed block size (satd_k::<4> / satd_k::<8>): the
difference block is a [[i32; K]; K], the vertical butterflies work on whole
rows at once, and the old per-column gather is gone. The sum is an integer
sum of the same values.
- A snapshot copies only the BlockDecoded cells of the block area (at most
16x16 per plane), the only flags coding inside the block can change, instead
of the whole 3x35x35 array.
The encoded stream is unchanged. A unit test compares the new SATD with the
previous implementation for 4..32 sample blocks at 8 and 10 bit.
Fixes#2387
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
* av1enc: tests for the region-only BlockDecoded snapshot and the bounded SADs
- `snapshot_restore_brings_back_every_blockdecoded_flag`: for every block
size of a 64x64 superblock (4x4 .. 64x64) at every aligned position in two
superblocks, from an all-clear and a random flag state, marking every
sub-block of every size inside the block and restoring gives back the whole
3x35x35 array, so the 16x16 region covers every flag coding can set and
leaves the neighbours alone.
- `bounded_sad_stops_only_once_the_cost_cannot_win`: `sad_ref` and
`sad_bounded` return the full SAD whenever `sad + lb < bound`, and
otherwise a partial sum that already loses the strict `<` of the callers.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Sonnet 5.5 <noreply@anthropic.com>
Co-authored-by: Codex <noreply@openai.com>
Co-authored-by: FredApps <153173854+FredApps@users.noreply.github.com>
* opusenc: memoise SILK rate-loop trials, ring buffer for noise feedback
Rate loop: each probe clones the encoder and the range coder and codes the
whole frame, about 7 trials per frame. The hard-limit check recomputed the
trial of the chosen offset and the final step coded the frame a third time.
Keep every trial (offset, bits, encoder state, range coder): the hard-limit
check now hits the memo and the final step adopts the memoised state and
coder. If the chosen offset is somehow not in the memo it is coded as before.
Quantiser: the noise-feedback history was shifted with copy_within for every
sample (the quantiser runs about 7 times per frame). Keep it in a
double-length ring so the 16 most recent errors are one contiguous slice;
the dot product reads the same pairs in the same order, and the history is
written back to the channel at the end.
The packet streams are byte-identical to before (cmp on 24 configurations
of 90 s: speech-like and music-like input, mono and stereo, SILK NB/MB/WB
and hybrid SWB/FB).
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
* opusenc: noise-feedback ring as a helper type, equivalence tests for the SILK rate loop
`quantize` was at cognitive complexity 31 on current main (limit 30). The
double-length ring for the noise-feedback history moves into `NfRing`
(push / window / to_array), and the end-of-frame bookkeeping of `encode`
into `end_frame`.
Tests: the ring matches a shifted history over 100 pushes, and an encoder
that adopts the memoised trial produces the same bytes, noise-feedback
history and LPC state, frame after frame, as one that codes every frame
again at the chosen offset (the step before the memo), for 8/12/16 kHz,
mono and stereo, with and without a tight hard limit.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Sonnet 5.5 <noreply@anthropic.com>
Co-authored-by: FredApps <153173854+FredApps@users.noreply.github.com>
Part of #3184: ConfinedFootage forwards FootageSource::recording to the
source it wraps, so renders of an automation session list their files
instead of coming back incomplete. The read tests also render through the
viewer's CancellableFootage wrapper.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Part of #3184 (step 1): renders now report which footage files they read,
each with its identity (size and modification time) as read. Nothing uses
the list yet: images, cache keys and cache behaviour are unchanged.
- render: `reads` module with `FileStamp`, `ReadFile`, `ReadFiles` (sorted,
deduplicated, bounded; marked incomplete when it can't vouch for every
file) and the thread-safe `ReadLog`. `Renderer::reads` collects a frame's
files; `FootageSource::recording` lets a source note what it reads.
`file_stamp` uses the same identity.
- Layer cache entries keep the files their buffer was made of and report
them on every hit; derived entries (styles, blend space, flattened) keep
their base's. Buffers from the disk cache or the browser store don't know
their files yet and mark the list incomplete.
- Media pool: every file read notes its identity, taken from the file it
opens. Decoded frames, opened movies, GIFs, models and EXR channels keep
it with the cache entry, so a frame served from memory reports its file
without touching the disk.
- Viewer frame worker collects each frame's list and carries it to where
the disk entry is written (step 2 stores it).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* Subspace Warp: QL eigen solver and banded mesh solve (#2319)
plan() spent 43-96 % of its time in the cyclic Jacobi sym_eigen, whose
absolute stop test never triggers on Gram matrices of magnitude ~1e8, and
then ran a dense Cholesky on a mesh normal matrix of bandwidth ~30.
Use Householder tridiagonalisation plus implicit QL inside subspace.rs
(only the leading subspace is used, so basis and sign do not matter), and
accumulate and factor the mesh normal equations in lower-band storage,
without a Vec per data row. camtrack/linalg.rs is untouched.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
* Subspace Warp: replace the QL eigen solver with a clean-room symmetric QR
The previous sym_eigen_ql followed a well-known published tred2/tqli listing
too closely for the clean-room rules. It is replaced by sym_eigen_qr, written
from the textbook method (Golub & Van Loan, Matrix Computations, ch. 8) in
its own structure:
- householder_tridiagonal: explicit reflectors with the symmetric rank-2
update B -= v w^T + w v^T (p = beta B v, w = p - (beta p^T v / 2) v), Q
accumulated as rows of Q^T.
- wilkinson_qr_step: one implicit QR step with the Wilkinson shift, Givens
rotations chasing the bulge from the top of the unreduced block down.
- sym_eigen_qr: deflation of negligible off-diagonal entries and the step
driver, split so each function stays under the complexity limit.
It returns None for non-finite input, a size mismatch or no convergence
within 30 n steps; smooth_trajectories then falls back to the Jacobi
sym_eigen, so it can never loop for ever or be worse than before.
Tests: random symmetric matrices (n = 1..61) against Jacobi eigenvalues,
residual |A v - lambda v| and V^T V = I; identity, repeated 3x3 blocks, zero,
1x1, 0x0, rows scaled from 1e-8 to 1e8 and a 1-2-1 tridiagonal; NaN, infinity
and size mismatch return None; the leading-subspace projector of Gram-like
low-rank matrices matches Jacobi. An ignored benchmark times plan() on a
300-frame synthetic parallax shot.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Sonnet 5.5 <noreply@anthropic.com>
Co-authored-by: FredApps <153173854+FredApps@users.noreply.github.com>
* Roto Brush: propagate on the band bounding box
propagate() ran the signed distance, gauss, graph cut, model fit and
clean over the whole frame. Work on the crop of the band bounding box
(|sd| <= r plus the core, padded) and paste the labels back; fall back
to the full frame when the crop is not smaller. warp_prior stays
full-frame.
Fixes#2323
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
* Roto Brush: band crop padded by the prior blur, models fitted on the full frame
Review follow-up for the band bounding-box crop in propagate:
- The crop pad is now the reach of the prior blur (three box passes of
matting::gauss, exposed as matting::gauss_reach) instead of 1 px, so every
band pixel sees the same blur window as on the full frame.
- When the previous frame has no colour models they are fitted on the full
frame before cropping, as the full-frame path does.
- The free pixels and the clean seeds are cut out of the full-frame signed
distance instead of being recomputed on the crop; the seeds equal
core_of(prior.soft), which saves one signed distance per frame.
- The full-frame path and the crop share recut_band/cut_band; full frame is
just the whole rectangle. A frame with no band and no core (no
foreground) skips the cut, which was a no-op.
- The w - 1 / h - 1 underflow on zero-sized frames is gone (band_rect
returns None); regression test for 4x0, 0x4 and 0x0 (panicked before in
debug builds).
Tests: crop vs full frame labels on a disk, a disk at the frame corner, a
donut, r = 2 and r = 3 with prior edges wider than the band, and no
foreground, each with fitted, given and equal colour models; the padded crop
blurs the prior like the full frame at every band pixel while a 1 px pad
does not.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Sonnet 5.5 <noreply@anthropic.com>
Co-authored-by: FredApps <153173854+FredApps@users.noreply.github.com>
* Roto Brush: reuse scratch vectors in the distance transform
matting::distance allocated four vectors per column plus a staging copy
and three vectors per row. Run the columns and rows through four reused
scratch vectors instead; the dt1 arithmetic is unchanged.
Fixes#2324
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
* Roto Brush: keep the distance transform parallel, with per-split scratch
The previous commit made the column and row passes serial, which was measured
on one thread only. Both passes are parallel again: the column pass runs on a
transposed buffer so every column is a contiguous chunk, and rayon's
for_each_init gives each work split one set of scratch vectors instead of
four allocations per line. The square root is folded into the row pass.
A test keeps the old per-line implementation and checks that the output is
bit-identical on a disk, its inverse, 1xN, Nx1, all-true, all-false, seeds on
the frame edge and a sparse 333x211 pattern. An ignored benchmark prints the
1080p median and the rayon thread count.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* Roto matting: size the distance-transform scratch per pass (columns h, rows w)
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Sonnet 5.5 <noreply@anthropic.com>
Co-authored-by: FredApps <153173854+FredApps@users.noreply.github.com>
On a hosted macOS runner the_shader_is_the_reference found 168 of 16384
bytes one step off, just over the 1 % a real GPU is allowed.
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
* EXR golden test: hash samples near, not bit for bit (temporary: print the new goldens)
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* EXR golden test: the new near hashes
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Native division isn't correctly rounded on real GPUs, so
posterize_matches_the_cpu_on_every_8bpc_pair failed on NVIDIA and Apple
GPUs while passing on the software renderer hosted CI uses.
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
* AEP import: map the Cryptomatte plug-in to EffectCraft's Cryptomatte
The importer reported every instance of the free Cryptomatte plug-in
(match name `Cryptomatte`) as a lost effect. It now becomes
`ec.3d.cryptomatte`:
- CryptomatteData decodes the plug-in's Settings data (its own
little-endian layout, documented in docs/aep-format.md 3.7): the
Cryptomatte layer name and the selection text, both carried verbatim.
- Output maps to Display / Matte Only (Colors -> Colors, Matted RGBA ->
Matte, Matte Only -> Matte Only). Matted Colors, Selection Mode when on,
keys or expressions and anything else stored are reported, and
unreadable data is reported as damaged.
- EffectCraft's Cryptomatte gains a Layer Name text parameter. A layer the
footage has under exactly that name wins; otherwise the Layer popup
decides as before. The CPU, GPU and pick paths share one resolver
(AuxChannels::resolve_crypto_layer_named). Without it, plug-in layer
names such as Crypto_Material would silently fall back to the first
layer.
On a real production project this takes the effect losses from 132 to 44.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* AEP import: check Cryptomatte against an After Effects 2026 sample of our own
A sample made in After Effects 2026 (plug-in 2.6.6: a fresh effect,
Selection Mode on, each Output item, with read-backs and renders)
corrected three points of the mapping:
- Selection Mode on was dropped silently: After Effects writes it as a
default-state record (tdsb 3) holding 1, which the general reader does
not take. The Cryptomatte step reads that record itself and reports it.
- No Output record is now the plug-in's default Colors, also for an
instance that borrows the project's definition (every instance after
the first), which was reported as having no reliable default and fell
back to EffectCraft's Preview. Keyed or unreadable Output values fall
back to Colors too, still reported.
- Matted Colors imports as Display Matte (the same matte over the layer's
own colours) instead of Preview, still reported as an approximation.
docs/aep-format.md 3.7 gets the sampled Output default, the tdsb 3
Selection Mode record and the real fresh Settings data (an empty text of
length 1 and its MurmurHash3, twice, not cry1 and zeros); docs/aep-import.md
the render comparison with After Effects (Matted RGBA matches; Matte Only
differs in alpha; Colors with no layer chosen shows another layer).
Tests: the exact fresh Settings bytes, tdsb 3 Selection Mode, borrowed
definitions, and an opt-in test against the sample once it is in the
corpus (aep/cryptomatte.aep, ignored until the corpus pin includes it).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* AEP import: write Cryptomatte's popup items in the test as After Effects does
The injected Cryptomatte definition wrote `pdnm` and `fnam` as `LIST`
chunks. After Effects writes them as plain chunks whose body is chunks
(`NESTED_IDS`), and the reader only looks there, so the tests never read
Output's item names: they passed through the fallback for a definition
with four items and no `pdnm`, which uses the plug-in's own list.
The harness now writes both as After Effects does. New tests check that
the injected definition reads back with its items and instance name,
that Output is mapped by the item the stored index names (items in
another order, case, an unknown item), and that the fallback without a
`pdnm` still works. With the old `LIST` form, or with the mapping
ignoring the items, the new tests fail.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* AEP import: check Cryptomatte's layer and selection against presets of our own
The After Effects 2026 sample 1400 now has two preset compositions that choose a layer and
pick one object (box_red on CryptoObject) or one material (mat_paint on CryptoMaterial).
CryptomatteData reads both as they are; the picked tokens are the EXR manifest's hashes,
which EffectCraft's selection grammar reads the same as the names.
- The u32 at 108 is MurmurHash3 of the selection text with its NUL (three texts).
- A unit test pins the 127-byte preset data; the sample test now covers all eight
compositions with their layer and selection.
- Render check: EffectCraft's matte covers exactly the pixels After Effects draws as
selected, edge coverage within 0.0002. docs/aep-format.md 3.7 and docs/aep-import.md.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* Cryptomatte import: report damaged Settings and Selection Mode records
Review follow-ups for #3109:
- CryptomatteData refuses a selection text without its NUL and checks the zero byte
at 112 for an empty text too.
- A Settings record without its data, and a Selection Mode record whose value cannot
be read, are reported as damaged data, not taken as the default.
- Layer Name is the effect's last parameter, so the earlier parameters keep their
indexes and upgraded instances have the same order as new ones.
- A Cryptomatte layer name is matched as given before it is trimmed.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* Cryptomatte settings: require the selection's NUL to end its declared length
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* Cryptomatte: match rank planes by the exact layer name
crypto_rank_planes looked planes up with a case-insensitive whole-name match, so
layers Crypto_A and crypto_a shared planes. The layer prefix now matches
case-sensitively; component letters still match in either case.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
* Format
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
* AEP: the exporter knows Extra::Cryptomatte; rustfmt
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: FredApps <153173854+FredApps@users.noreply.github.com>
* #1335: accessibility slice 2: Info, Flowchart, Script Console and the 3D Camera Tracker block named, described and keyboard-operable
The audit gets read-outs, graphics and logs as described items (a name is required in a fixed panel), dedicated maximized scenarios for the panels the shared loop squeezes, and minimum counts for the new zero panels.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
* Review: Enter/Space on a favourite star only toggles it; keyboard activation of a property label; settings page scrolls to focus; sharper UI test assertions
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
* Refresh the accessibility audit baseline after merging upstream/main
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
* a11y audit: hold the machine-dependent Media Browser scenario to its upper bound
The Media Browser opens on the home folder, so a runner with fewer files finds fewer controls and
the 'improved, lower the entry' check failed in CI. The entry stays an upper bound.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
* Merge upstream/main and updated feat-gap25-a11y-audit into feat-gap25-a11y-panels-2
Brought in the updated feat-gap25-a11y-audit (which already contains upstream/main). Conflict: crates/ui-egui/src/menus.rs, which upstream split into menus/*.rs. Took the slim module and re-applied this branch's Flowchart hunks where the code now lives: nudge() skips the Flowchart (menus/invoke.rs) and Enter is left to the Flowchart in handle_shortcuts (menus/shortcuts.rs).
Also removed a duplicate MACHINE_DEPENDENT const in tests/ui_accessibility_audit.rs that the two branches both added. The audit BASELINE still matches the measured findings.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
* Flowchart: split show() to pass the stricter clippy limits from main
Main now denies cognitive_complexity (30) and too_many_arguments; the
merged flowchart panel exceeded it. The keyboard navigation, the
key-to-node mapping, pan-into-view, node commands, the accessible node
name and the option params move into small helpers with unchanged
behaviour.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
* Camera tracker a11y test: wait for the analysis result instead of one frame
The test pressed Enter, ran one frame and expected to see the background job.
On a fast runner the job starts, ends and is written back inside that frame,
so the state was never seen. Undo the first analysis, press Enter, and wait
(bounded) for the new analysis step in the undo history.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
* Accessibility: one hours range for the accessibility pass in ui-parity.md; test that Flowchart automation labels stay English
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
* Flowchart toolbar: keep the English label for the automation registration
The translated text is shown and spoken; the automation label stays English.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Sonnet 5.5 <noreply@anthropic.com>
Co-authored-by: OpenAI Codex <noreply@openai.com>
Co-authored-by: FredApps <153173854+FredApps@users.noreply.github.com>
* Timeline: hover texts for the icon-only column headers and layer switches
The Timeline's column header is painted as one click target, so its
icon-only headers (A/V Features, Label, #, the eight switches and the
Modes column's T) had nothing to hover and nothing explained them.
HEADER_TIPS lists each one with a short name and an explanation in our
own words. Hover-only widgets on top of the header show the text, carry a
named ColumnHeader AccessKit node and register automation ids
(timeline.header.av.*, timeline.header.switches.*,
timeline.header.modes.preserveTransparency). The header's clicks and
right-click column menu still work underneath. The layer rows' A/V and
switch toggles and the Preserve Transparency box show the same texts.
The accessibility audit's Timeline baseline drops from 89/42 to 72/25
because the switches now have accessible names.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* Timeline headers: accessible description, Preserve Transparency text, English menu test
The header's node is named before its hover text is added, so screen readers get the
explanation; Preserve Transparency says the layer follows the alpha below it, partial alpha
included; the hover test sets English before checking the translated Hide This item.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* Merge main: keep both sets of timeline tests
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* Timeline tests: close the test before the merged one
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: FredApps <153173854+FredApps@users.noreply.github.com>
show() went one over the cognitive complexity limit once #3094 landed on
current main, which failed clippy on every later commit.
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
* Cryptomatte: the Layer popup lists the footage's Cryptomatte layers (#473)
The popup was the constant CryptoObject / CryptoMaterial / CryptoAsset. It now lists the
Cryptomatte layers of the layer's OpenEXR footage, and those three when there are none. The
stored value keeps its meaning (0-2 the standard layers, 3 and up the file's other layers in
name order), so older projects render the same.
Refs #473
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
* Cryptomatte Layer popup: timeline editor, one menu helper, probe and fallback tests (#473)
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
* Cryptomatte Layer popup: give further layers of a standard kind a custom slot
A.CryptoObject keeps slot 0; B.CryptoObject and later ones get 3+ instead of
being dropped, so they can be selected, picked and rendered.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
* Format; note same-kind layers in the footage docs
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
* Format the merged code with rustfmt
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Sonnet 5.5 <noreply@anthropic.com>