Commit Graph
525 Commits
Author SHA1 Message Date
vangel76 6bea5fa6ae Merge upstream/main into cache-defaults-3298 2026-11-19 10:21:55 +01:00
arnav-hiwarkar c6612da798 #3830: tell a 3-entry .3dl mesh apart by the row count alone 2026-11-18 15:46:04 +05:30
arnav-hiwarkar bf31b65271 #3830: keep four .3dl tokens and a spare mesh row at small limits 2026-11-18 15:38:08 +05:30
arnav-hiwarkar 648ae78591 #3830: test the .3dl cap through an injectable limit; drop the local .spi3d precheck (#3831 belongs in craft-cms) 2026-11-18 15:27:36 +05:30
arnav-hiwarkar 104875a897 #3830: read only the three .3dl Mesh header tokens it uses 2026-11-17 22:28:16 +05:30
arnav-hiwarkar 6232bed556 #3830: stop collecting a .3dl line's numbers past the mesh limit 2026-11-17 22:20:14 +05:30
arnav-hiwarkar a65a1f746f #3830: cap the .3dl mesh and rows and cube with checked math; #3831: refuse a .spi3d short of its rows before the lattice is reserved 2026-11-17 22:11:34 +05:30
arnav-hiwarkar d550d3d476 #3830: regression tests for hostile .3dl meshes and row-less .spi3d lattices 2026-11-17 22:09:04 +05:30
vangel76 5d0bfe7869 Caching: memory is let go only where it frees memory, Purge keeps nothing decoding meanwhile, and regions read the rows they show (#3750 review)
- Cryptomatte planes: a plane a render holds is not offered for its gone channels, and the entry
  limit holds while renders hold every plane.
- Footage frames and whole stills held elsewhere are not offered to the memory accountant.
- Purge (clear frames) makes decodes in flight stale: their frames are not kept or joined.
- Plane store: a size refusal of a merged decode is not handed to a request that fits alone;
  slimming a shared set and copies for the disk cache take memory fallibly, and planes the disk
  cache can't take are not copied.
- Disk cache: the plane encoder's scratch is taken fallibly; no memory skips the entry.
- Region of interest: rows follow the projected transform of a 2D layer under a 3D parent, and
  a frame of another size than its footage is read whole.
- The box filter's work bound is checked without a wall clock.
2026-11-17 17:00:29 +01:00
arnav-hiwarkar c07dbf32ca Merge remote-tracking branch 'upstream/main' into fix/3464-drop-shadow-softness-range 2026-11-17 20:54:02 +05:30
ash 4c38f78307 Merge pull request #3817 from arnav-hiwarkar/fix/3467-radial-blur-amount-range
Radial Blur: Amount takes After Effects' range 0..1000
2026-11-17 19:03:14 +05:30
ash 0bd848c272 Merge pull request #3153 from jayzed80/fix/parse-cube-huge-size
fix(effects): parse_cube rejects a hostile LUT_3D_SIZE before cubing it
2026-11-17 18:37:21 +05:30
arnav-hiwarkar eb14583749 Merge remote-tracking branch 'upstream/main' into fix/3467-radial-blur-amount-range
# Conflicts:
#	crates/aep/src/export/fx_defs.rs
#	crates/engine/src/tests/effect_ranges.rs
#	docs/effects-parity.md
2026-11-17 18:34:15 +05:30
arnav-hiwarkar 7f14df8355 Merge upstream main into fix/3460-gaussian-blur-range
Keeps the Gaussian Blur (#3460) and Sharpen (#3466) tests in
crates/effects/src/misc.rs; Sharpen's code is as main has it.
2026-11-17 18:08:30 +05:30
ash a188867a42 Merge pull request #3810 from arnav-hiwarkar/fix/3466-sharpen-amount-range
Sharpen: Amount takes After Effects' 0..4000 range (#3466)
2026-11-17 18:06:39 +05:30
arnav-hiwarkar ff9584fa64 Merge remote-tracking branch 'upstream/main' into fix/3467-radial-blur-amount-range
# Conflicts:
#	crates/engine/src/tests/effect_ranges.rs
#	docs/effects-parity.md
2026-11-17 17:57:33 +05:30
arnav-hiwarkar 69666e68c7 Merge upstream main into fix/3460-gaussian-blur-range
Keeps the Gaussian Blur (#3460) and Turbulent Displace (#3469) rows in
the engine range test and the effects parity log.
2026-11-17 17:57:21 +05:30
arnav-hiwarkar 1654633bcf fix(effects): test the cube row bound without a huge input 2026-11-17 17:33:35 +05:30
arnav-hiwarkar ce0c4c8a75 #3464: Drop Shadow: blur before moving when the margin is cut
With the padding cut to its budget, a Distance larger than what's left
of the margin moved the layer out of the buffer before the blur, so the
blur had nothing to spread back in: a 64 x 48 layer moved 3000 px at
Softness 8000 left no shadow at all. The cut-margin path now samples
only the fraction of the move and lets the blur move the rest:
raster::gaussian_blur_unbounded_at evaluates the transparent-edged blur
of the image at a whole-pixel shift, through a window the size of the
buffer, so content outside the window still reaches into it. Margins
that fit are unchanged.

The blur's intermediate rows are freed before the output is allocated,
so it holds at most three buffers at once. On an M-series Mac a
4096 x 4096 layer at Softness 30000 takes 0.51 s, and 1080p 0.13 s.
2026-11-17 17:30:44 +05:30
arnav-hiwarkar 06c41a1ee4 fix(effects): merge current upstream main into cube safety fix 2026-11-17 17:27:35 +05:30
arnav-hiwarkar 1eb4f39808 fix(effects): bound cube data row collection 2026-11-17 17:27:22 +05:30
arnav-hiwarkar 95685d8d41 Merge remote-tracking branch 'upstream/main' into fix/3469-turbulent-displace-amount-range
# Conflicts:
#	docs/effects-parity.md
2026-11-17 17:26:50 +05:30
arnav-hiwarkar ae2d595d9d fix(effects): exercise the wrapped combined cube row count 2026-11-17 17:25:47 +05:30
arnav-hiwarkar 83285272e3 Merge upstream main into fix/3460-gaussian-blur-range
Keeps the Gaussian Blur (#3460), Fractal Noise (#3486) and CC Power Pin
(#3492) rows in the effects parity log.
2026-11-17 17:25:22 +05:30
arnav-hiwarkar ffdcdb507d Merge remote-tracking branch 'upstream/main' into fix/3464-drop-shadow-softness-range 2026-11-17 17:25:13 +05:30
arnav-hiwarkar cebe224d85 Merge remote-tracking branch 'upstream/main' into fix/3467-radial-blur-amount-range
# Conflicts:
#	docs/effects-parity.md
2026-11-17 17:24:34 +05:30
arnav-hiwarkar 5adc96bfbf Merge remote-tracking branch 'upstream/main' into fix/3466-sharpen-amount-range
# Conflicts:
#	crates/engine/src/tests/effect_ranges.rs
2026-11-17 17:24:31 +05:30
arnav-hiwarkar 9bb0db0207 Merge upstream main into fix/3460-gaussian-blur-range
Keeps both the Gaussian Blur (#3460) and CC Power Pin (#3492) rows in
the engine range test and the effects parity log.
2026-11-17 16:55:18 +05:30
arnav-hiwarkar 969286ae76 #3460: review: name the open follow-up for SLIDERS_0_24, use the shared pixel budget in the test
The table holds only Gaussian Blur's row; the ranges changed since 0.24.1
by other work are tracked in #3815. The 30000 render test checks the
buffer against crate::pixel_budget instead of a copied constant.
2026-11-17 16:52:08 +05:30
arnav-hiwarkar 8fe42b29c6 Merge remote-tracking branch 'upstream/main' into fix/3464-drop-shadow-softness-range 2026-11-17 16:49:21 +05:30
arnav-hiwarkar 79d4be499d Merge remote-tracking branch 'upstream/main' into fix/3466-sharpen-amount-range
# Conflicts:
#	crates/engine/src/tests/effect_ranges.rs
#	docs/effects-parity.md
2026-11-17 16:48:30 +05:30
arnav-hiwarkar 63ec0033d1 #3486: merge main (CC Power Pin range rows in the shared tables) 2026-11-17 16:48:19 +05:30
arnav-hiwarkar a3ab20e9f7 Merge remote-tracking branch 'upstream/main' into fix/3469-turbulent-displace-amount-range
# Conflicts:
#	crates/engine/src/tests/effect_ranges.rs
#	docs/effects-parity.md
2026-11-17 16:46:35 +05:30
arnav-hiwarkar 8420bb9f4d Merge remote-tracking branch 'upstream/main' into fix/3467-radial-blur-amount-range
# Conflicts:
#	crates/engine/src/tests/effect_ranges.rs
#	docs/effects-parity.md
2026-11-17 16:46:29 +05:30
arnav-hiwarkar cec04744f3 #3464: Drop Shadow: test a huge Softness; say what the GPU test checks
A Softness far past the range (an expression can set one) overflowed
the padding's u32 before drop_shadow_pad saturated it; the new test
panicked there in a debug build and renders a finite shadow now. The
GPU test's comment and the parity row now say that a cut margin renders
on the CPU, and the CPU test quotes the measured 0.11 at Softness 8000.
2026-11-17 16:24:22 +05:30
arnav-hiwarkar f149b65c57 #3464: Drop Shadow: Softness takes After Effects' range
Softness took 0..1000, so prop.set and the value editors clamped values
After Effects accepts: its recorded definition for ADBE Drop Shadow
allows 0..30000 in the same pixels (the .aep mapping is plain). The hard
range is now 0..30000; the default (0) and the slider (0..250) are
unchanged. Projects saved with the old range take the new one on open.

The .aep range guard loses its allow-list row, and the engine test sets
both ends through prop.set, undo / redo, save / reopen and an old
project. Keyframe text keeps no unit for it (only 0..100 sliders read
as percent).
2026-11-17 16:24:22 +05:30
arnav-hiwarkar 8903090007 #3464: Drop Shadow: keep the whole blur when the padding budget cuts the margin
Drop Shadow pads its buffer by Distance + 1.5 x Softness, but the pixel
budget (#2261) cuts that margin for a large Softness: from about 870 on
a 1080p layer. The box passes then dropped what they spread past the
edge, so the shadow came out too faint: at the centre of a 64 x 48
layer, 0.11 of the Gaussian at Softness 8000. With a cut margin the CPU now blurs as
if the buffer went on with transparency (the Gaussian is unchanged,
sigma = Softness / 2). The GPU hands that case to the CPU and keeps
every margin that fits, so those results are unchanged.

Measured on an M-series Mac: 1080p at Softness 30000 takes 0.10 s, and
a 4096 x 4096 layer (an 8192 x 8192 buffer) takes 0.99 s on the CPU.

The pad is computed once for both paths (drop_shadow_pad) and saturates
instead of overflowing on a huge Softness.
2026-11-17 16:24:22 +05:30
arnav-hiwarkar e60d6ac706 #3486: review: the octave test checks the bound; clearer bound comment
The octave test asserts both values stay within OCTAVE_SCALE_MAX (it
passed with the bound removed). The comment names Perspective Offset as
what multiplies the offset by the depth, and says why an octave at the
bound has no detail.
2026-11-17 16:22:07 +05:30
arnav-hiwarkar a62564b567 #3469: locked pinning clamps without f64::clamp
The sub-pixel bounds lx + ix and lx + lw - ix round separately, so at third
resolution they can still cross by an ulp and f64::clamp panics; a NaN
layer size panicked too. max/min equals clamp whenever the bounds are
ordered and cannot panic. The test now covers half and third resolution at
fractional offsets and a NaN layer size. The parity note says Amount, not
the displacement, is bounded to 2^24 px.
2026-11-17 16:19:44 +05:30
arnav-hiwarkar d5e74c880f #3467: Radial Blur docs say the wide Zoom shape is unmeasured
The spin helper's name gives its unit (degrees), the zoom helper's doc
uses |amount| as the code does, and the parity row and tests say what
is and is not checked against After Effects.
2026-11-17 16:15:51 +05:30
arnav-hiwarkar 4cf7741340 #3469: Turbulent Displace locked pinning on a sub-pixel layer
Locked pinning clamped samples to lx + 0.5 .. lx + lw - 0.5. On a layer
under a buffer pixel across (a 1-pixel layer at half resolution) the bounds
cross and f64::clamp panics. They now meet at the layer's centre there and
are unchanged otherwise.
2026-11-17 16:14:01 +05:30
arnav-hiwarkar 040fe61b84 #3469: Turbulent Displace stays finite at any Amount
Expressions are not clamped to the hard range. Padding computed
amount.abs().ceil() as u32 + 1, which overflowed (a panic in debug builds)
from about 4.3e9 on both the CPU and GPU paths, a NaN Amount filled the
layer with NaN, and an infinite one sent infinite coordinates to the
samplers' integer pixel positions.

turbulent_amount_px, shared by both paths, skips a zero or NaN Amount and
bounds the rest to 2^24 px. Values inside the hard range are unchanged.
Tests cover both ends of the range with every displacement kind, pinning
and Resize Layer (finite, padding within budget), Amounts past it, and
CPU/GPU agreement at -10000, 10000 and infinity.
2026-11-17 16:13:10 +05:30
arnav-hiwarkar 2e17ed1d83 #3469: Turbulent Displace Amount takes After Effects' valid range
The hard range was -1000..1000, so prop.set and the Effect Controls and
Timeline editors clamped Amount values After Effects accepts. The recorded
ADBE Turbulent Displace definition allows -10000..10000 (plain units, no
conversion). The default and the -200..200 slider are unchanged.

The .aep range guard no longer lists the parameter as narrower, and the
engine range test covers both ends through prop.set, undo/redo, save and
reopen, and projects saved with the old range.
2026-11-17 16:12:25 +05:30
arnav-hiwarkar ec1294a1d2 #3486: Fractal Noise: Sub Scaling takes After Effects' 10..10000
After Effects records Sub Scaling's valid range as 10..10000 (slider
25..100), but the hard range stopped at 100, so prop.set and the value
editors clamped 200 to 100. The CPU renderer and the GPU pointwise path
also capped the ratio at 1, so a value above 100 rendered as 100 even
once stored.

Widen the hard range for Fractal and Turbulent Noise (they share the
parameter; slider and default unchanged) and take the ratio from one
sub_scaling_ratio (0.1..100) on the CPU and both GPU paths. Above 100
each octave is larger than the last.

Tests: the range and that 200 and 10000 render differently from 100;
finite output at both ends with Complexity 20, Perspective Offset,
Turbulence Factor 10 and a huge offset; CPU/GPU parity at 200 and 10000
on the pointwise and fxn_fractal paths; prop.set, undo/redo and
save/reopen keep 10 and 10000 (engine range table). The .aep range
guard no longer lists the parameter.
2026-11-17 16:10:16 +05:30
arnav-hiwarkar f98c3e9ba2 #3467: Radial Blur Zoom past 200 stays on its side of the centre
Zoom spreads its 32 samples over scales 1 ± Amount / 200 about the
centre. Past Amount 200, which the range now allows, that reaches below
0 and samples the layer mirrored through the centre: at 1000 a layer
white on the left and black on the right turned the right half 35 %
grey.

raster::radial_blur_zoom_scale keeps the old scales up to 200 exactly
and spreads the samples evenly from the centre to 1 + Amount / 200
beyond. The CPU and GPU paths both take their samples from it. Work per
pixel stays at 32 samples at any Amount.
2026-11-17 16:09:38 +05:30
arnav-hiwarkar 12480ab94e #3492: say when crossed Expansion sides flip the CC Power Pin source
The source rectangle flips when two opposite sides add up to less than
-100 %, not when one side alone does.
2026-11-17 16:09:32 +05:30
arnav-hiwarkar d918aa6e13 #3467: Radial Blur Amount takes After Effects' range 0..1000
After Effects records 0..1000 for Radial Blur's Amount (slider 0..100);
EffectCraft's hard range stopped at 100, so prop.set and the value
editors clamped 200 to 100. The hard maximum is now 1000; the slider
range and default are unchanged.

The engine range test covers prop.set at both ends, undo and redo, save
and reopen, and a project saved with the old range. The parameter leaves
the .aep range guard's list of known narrower ranges.
2026-11-17 16:09:29 +05:30
arnav-hiwarkar 8f7896b28a #3486: Fractal / Turbulent Noise: bound octave scales and scaled offsets
Each octave's frequency and Perspective Offset depth are Sub Scaling to
the power of the octave. Above 100 % (After Effects allows 10000) that
reaches 100^±19 at Complexity 20, and Perspective Offset or Turbulence
Factor times the offset overflowed to infinity and drew NaN.

Both stay within 1e20 and its inverse and remain each other's inverse,
so an octave's offset in lattice cells is unchanged; 10..100 never
reaches the bound. The scaled offset stays within 1e36 px, past where
the lattice index saturates. The CPU, fxn_fractal (from fractal_gpu)
and pointwise paths share the bound; one octave_origin replaces the two
copies of the origin code.
2026-11-17 16:07:18 +05:30
arnav-hiwarkar ef7dd7f3d1 #3460: Gaussian Blur: Blurriness takes After Effects' range 0..30000
After Effects records Blurriness as valid from 0 to 30000; the hard range
was 0..3000, so prop.set and the value editors clamped 4000 to 3000. The
default (0) and the 0..50 slider stay as they were. Padding stays within
the effect pixel budget at the new maximum, on the CPU and the GPU.

Tests: the engine round trip (prop.set, undo/redo, save/reopen, older
projects) gets a row, the aep guard drops its known-narrower row, and the
CPU and GPU render at 30000 is checked bounded, finite and matching.
2026-11-17 16:06:36 +05:30
arnav-hiwarkar e83be3a08f #3486: Fractal Noise: lattice neighbours wrap past the i32 range
At Sub Scaling 10 % and high Complexity an octave's lattice coordinates
pass the i32 range. The cast saturates, and adding the neighbour offset
overflowed (a panic in debug builds). Wrap it, as the GPU's i32 adds do.
2026-11-17 16:05:40 +05:30