* 32-bit: legacy Brightness/Contrast, Channel Mixer, Photo Filter and B&W tint unclipped
In 32-bit documents Photoshop keeps these adjustments' results above 1
and below 0. On the photoshop corpus's rgb32/gray32 files:
- Legacy Brightness/Contrast pivots its contrast at 127 of 255 (8/16-bit:
128) and is exact on every pixel.
- Channel Mixer cuts each source percentage to 1/1024 toward zero, counts
255/512 of the constant and is exact on pure red, green, blue, black
and white.
- Photo Filter works on the linear samples, negative results included.
- The Black & White tint is the Color blend mode without the clip at 1,
as the blend mode is there; white tints to (1.05, 1.00, 0.87).
We clipped all four to 0..1. The GPU kernels follow suit under F_HDR
(the legacy pivot and the mixer's steps come from the planner), and the
HDR parity test covers them.
Photoshop oracles: rgb32 and gray32 brightness-contrast-legacy, rgb32
channel-mixer and black-white-tint pass (145 -> 149 of 258); floor
145 -> 149. rgb32 photo-filter improves (0.19 -> 0.035 max); the rest is
its colour (#3790).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* Masks: feathers blur with exactly the feather's variance
A mask feather is a Gaussian with sigma = feather, which we approximate
with three box passes per axis. Whole-pixel boxes rounded the variance:
a 6 px feather blurred like 6.48 px and a 3 px one like 3.46 px, so
feathered edges were too soft by up to 4 levels against Photoshop
(psd-tools mask-parameters-no-real-channel.psd: a 6 px vector and 3 px
pixel feather).
Each pass is now an "extended box" (Gwosdek et al. 2011): weight 1 on
the whole radius and a fractional weight on the two samples beyond it,
so three passes have variance sigma^2 exactly. It stays O(n), reusing
the window's leaving sample for the low end weight (50 MP at a 1000 px
feather: 180 ms parallel, was 172).
psd-tools: mask-parameters-no-real-channel 0.0168 -> 0.0060 (PASS),
layer_mask_data.psd 26.9 % -> 3.9 % of pixels off (psb 25.9 % -> 1.8 %).
Corpus test 239 -> 240; floor 236 -> 240.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* Threshold: split on Photoshop's luminosity of whole samples
Photoshop's Threshold splits on the luminosity 0.30 R + 0.59 G + 0.11 B
(the blend modes' lum, as Gradient Map uses), not the Rec. 601 grey,
taken over the document's whole sample values:
- 8 bits: 30 R + 59 G + 11 B > 100 level - 50, i.e. the luminosity
rounded to a level with halves down reaches the level;
- 16 bits: 30 R + 59 G + 11 B >= 100 x the level in 16 bits
(round(level x 32768 / 255): mid grey 16448 is level 128);
- 32 bits: the luminosity reaches the level.
Exact on every pixel of psd-tools threshold_rgb (36432 px; Rec. 601 split
28 the other way), threshold_16bits_rgb (100) and threshold_16bits_
grayscale, and the photoshop corpus's rgb16 threshold-128. The sums are
exact in f32, on the CPU and in the GPU kernel, so ties (grey exactly on
the level) split the same way on both.
psd-tools: threshold_rgb and threshold_16bits_rgb pass; corpus test
239 -> 241, floor 236 -> 241.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* Posterize: 16-bit documents bin the 16-bit sample, as Photoshop does
At 16 bits Photoshop doesn't round to 8-bit levels before binning. It
bins the 16-bit sample, floor(v16 * n / 32768), and lands on 8-bit levels
through a 16-bit step: floor(b * 32768 / (n - 1)), converted with
floor((o * 255 + 16384) / 32768). So the middle of 3 levels is 128 where
8 bits give 127, while 42.5 and 25.5 stay 42 and 25.
Exact on every sample of psd-tools posterize_16bits_rgb (levels 3, 7,
13, 21) and posterize_16bits_grayscale (2, 6, 9, 11); the 8-bit rule
missed up to half the range at bin edges. The same rule applies to
CMYK and Lab channels at 16 bits: on the stored inks (#4945's native
path) psd-tools posterize_16bits_cmyk goes from 4.26 % to 0.01 % of
pixels off. 8-bit Posterize is unchanged.
The GPU kernel runs the same rule; 16-bit layers reach it as half
floats, so a sample within their precision of a bin edge can land in
the neighbouring bin there (the CPU, which exports, is exact).
psd-tools 239 -> 241; floor 236 -> 241.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* Groups: a pass-through clipping base with an adjustment renders isolated
A pass-through group composites its children straight into the
backdrop, so an adjustment layer inside reaches the layers beneath.
When the group is the base of a clipping group and holds an adjustment
layer, Photoshop renders it isolated instead: in psd-tools
passthrough_clipping_mask_adjustment an Exposure in such a group leaves
the layers beneath alone (we darkened them: half the image 170 levels
off). The same group with a Linear Dodge child and no adjustment keeps
reaching the backdrop (passthrough_clipping_mask_blendmode), and so do
clipping-mask3/4/5, so the rule is the adjustment, not the clipping
alone (isolating every pass-through clipping base broke those two).
CPU compositor and GPU plan alike. psd-tools 239 -> 240; floor 240.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: Brandon Thomas <echelon@gmail.com>
* Threshold: split on Photoshop's luminosity of whole samples
Photoshop's Threshold splits on the luminosity 0.30 R + 0.59 G + 0.11 B
(the blend modes' lum, as Gradient Map uses), not the Rec. 601 grey,
taken over the document's whole sample values:
- 8 bits: 30 R + 59 G + 11 B > 100 level - 50, i.e. the luminosity
rounded to a level with halves down reaches the level;
- 16 bits: 30 R + 59 G + 11 B >= 100 x the level in 16 bits
(round(level x 32768 / 255): mid grey 16448 is level 128);
- 32 bits: the luminosity reaches the level.
Exact on every pixel of psd-tools threshold_rgb (36432 px; Rec. 601 split
28 the other way), threshold_16bits_rgb (100) and threshold_16bits_
grayscale, and the photoshop corpus's rgb16 threshold-128. The sums are
exact in f32, on the CPU and in the GPU kernel, so ties (grey exactly on
the level) split the same way on both.
psd-tools: threshold_rgb and threshold_16bits_rgb pass; corpus test
239 -> 241, floor 236 -> 241.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* Posterize: 16-bit documents bin the 16-bit sample, as Photoshop does
At 16 bits Photoshop doesn't round to 8-bit levels before binning. It
bins the 16-bit sample, floor(v16 * n / 32768), and lands on 8-bit levels
through a 16-bit step: floor(b * 32768 / (n - 1)), converted with
floor((o * 255 + 16384) / 32768). So the middle of 3 levels is 128 where
8 bits give 127, while 42.5 and 25.5 stay 42 and 25.
Exact on every sample of psd-tools posterize_16bits_rgb (levels 3, 7,
13, 21) and posterize_16bits_grayscale (2, 6, 9, 11); the 8-bit rule
missed up to half the range at bin edges. The same rule applies to
CMYK and Lab channels at 16 bits: on the stored inks (#4945's native
path) psd-tools posterize_16bits_cmyk goes from 4.26 % to 0.01 % of
pixels off. 8-bit Posterize is unchanged.
The GPU kernel runs the same rule; 16-bit layers reach it as half
floats, so a sample within their precision of a bin edge can land in
the neighbouring bin there (the CPU, which exports, is exact).
psd-tools 239 -> 241; floor 236 -> 241.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* Groups: a pass-through clipping base with an adjustment renders isolated
A pass-through group composites its children straight into the
backdrop, so an adjustment layer inside reaches the layers beneath.
When the group is the base of a clipping group and holds an adjustment
layer, Photoshop renders it isolated instead: in psd-tools
passthrough_clipping_mask_adjustment an Exposure in such a group leaves
the layers beneath alone (we darkened them: half the image 170 levels
off). The same group with a Linear Dodge child and no adjustment keeps
reaching the backdrop (passthrough_clipping_mask_blendmode), and so do
clipping-mask3/4/5, so the rule is the adjustment, not the clipping
alone (isolating every pass-through clipping base broke those two).
CPU compositor and GPU plan alike. psd-tools 239 -> 240; floor 240.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: Brandon Thomas <echelon@gmail.com>
* Posterize: 16-bit documents bin the 16-bit sample, as Photoshop does
At 16 bits Photoshop doesn't round to 8-bit levels before binning. It
bins the 16-bit sample, floor(v16 * n / 32768), and lands on 8-bit levels
through a 16-bit step: floor(b * 32768 / (n - 1)), converted with
floor((o * 255 + 16384) / 32768). So the middle of 3 levels is 128 where
8 bits give 127, while 42.5 and 25.5 stay 42 and 25.
Exact on every sample of psd-tools posterize_16bits_rgb (levels 3, 7,
13, 21) and posterize_16bits_grayscale (2, 6, 9, 11); the 8-bit rule
missed up to half the range at bin edges. The same rule applies to
CMYK and Lab channels at 16 bits: on the stored inks (#4945's native
path) psd-tools posterize_16bits_cmyk goes from 4.26 % to 0.01 % of
pixels off. 8-bit Posterize is unchanged.
The GPU kernel runs the same rule; 16-bit layers reach it as half
floats, so a sample within their precision of a bin edge can land in
the neighbouring bin there (the CPU, which exports, is exact).
psd-tools 239 -> 241; floor 236 -> 241.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* Groups: a pass-through clipping base with an adjustment renders isolated
A pass-through group composites its children straight into the
backdrop, so an adjustment layer inside reaches the layers beneath.
When the group is the base of a clipping group and holds an adjustment
layer, Photoshop renders it isolated instead: in psd-tools
passthrough_clipping_mask_adjustment an Exposure in such a group leaves
the layers beneath alone (we darkened them: half the image 170 levels
off). The same group with a Linear Dodge child and no adjustment keeps
reaching the backdrop (passthrough_clipping_mask_blendmode), and so do
clipping-mask3/4/5, so the rule is the adjustment, not the clipping
alone (isolating every pass-through clipping base broke those two).
CPU compositor and GPU plan alike. psd-tools 239 -> 240; floor 240.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: Brandon Thomas <echelon@gmail.com>
Repeating a duplicate or paste of a 12 MP layer cost as much as the first time because the
GPU compositor's lower-stack cache stayed at the checkpoint of the first edit after it was
filled: an edit higher up only checked that the layers below that checkpoint were unchanged,
so every refresh re-ran all passes above it (in the #209 layered document, 87 of 110 passes
over the 12 MP damage rect; the duplicate itself takes 0.3 ms, its tiles are already shared).
Each cached chunk now remembers its own checkpoint. A render captures at the highest
checkpoint below the edit, also where it resumed from a lower cached one, so the next
refresh of that edit (its undo, redo or repetition) resumes just below it. Chunks a refresh
doesn't reach keep their lower prefix; a chunk moved by a partial capture holds only that
rectangle afterwards. Validity is unchanged: a chunk survives only while the layers below its
checkpoint are untouched.
Release perf_scenarios, layered group, Windows 11, RTX 5070 Ti, 7 interleaved runs:
P9 p50 22-29 -> 12-17 ms, P11 p50 37-41 -> 11-18 ms; P8 and P10 fall too; P1-P7, P12 unchanged.
Co-authored-by: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
The compositor kept every pattern texture it ever uploaded (RGBA32F,
16 bytes per pixel) until the device was lost: `forget_doc` and
`retain_docs` skipped them and no budget applied, so a long session
that opened and closed documents with pattern fills or pattern effects
only ever grew.
Each cached pattern now records the document that last painted with it
and the frame it last painted in:
- closing a document (`forget_doc`, `retain_docs`) frees its patterns;
- a pattern no frame has painted with for 240 frames is dropped;
- over PATTERN_BUDGET (256 MiB) the least recently painted go first,
never one the current frame paints with.
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
* fix(gpu): contain errors in unwatched compositor work
* test(gpu): gate recovery fixtures on format capabilities
---------
Co-authored-by: haranishi <223786623+haranishi@users.noreply.github.com>
GpuCanvas::retain called forget_doc only for documents still in its
texture map. A document whose entry had been dropped earlier (texture
creation failed, a fresh render failed after uploading, or the canvas
outgrew its budget after Canvas Size) kept its resident pages, effect
maps and lower-stack composite after it was closed, until another
document pushed the budgets over; with no other GPU document, for good.
Compositor::retain_docs forgets every document the caller doesn't list
as open, whatever the texture map says, and retain calls it each time.
The open documents' ids, their preview ids and the adjustment proxies
are already in that list, so nothing in use is dropped.
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
A layer with Blending Options › Blend If made the planner refuse the
whole document, so every refresh of it ran the CPU compositor: on the
#209 layered document an opacity edit took 140 ms instead of 6 ms.
Blend If is compose::apply_blend_if: the composite with the layer mixed
back towards the backdrop by a per-pixel weight from the layer's own
content with its masks ("This Layer"; an adjustment's result) and the
backdrop ("Underlying Layer"). The planner now keeps the backdrop,
composites the layer as before (channel restrictions included, as on
the CPU), renders its own content and emits one Lerp pass with the new
F_BLEND_IF flag and the ranges packed exactly in its parameters. Base
and clipped layers both go through it; Advanced Blending still falls
back to the CPU.
Hard points: compositing can leave an 8-bit level a few millionths
below itself (a level-20 backdrop at partial alpha reads 19.999998),
which put it on the wrong side of a hard point at that level on the
CPU, and the GPU's arithmetic differs there. Both now treat a value
within 1e-4 of a level as that level; 8/16-bit values off a level are
at least 1/257 away. On 16/32-bit documents the GPU reads layers as
RGBA16F, so a value within about a tenth of a level of a hard point can
still land on the other side, as with the GPU Threshold adjustment.
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: Brandon Thomas <echelon@gmail.com>
A duplicate shares its source's tiles (Arc), but its resident texture was new, so the
compositor converted and uploaded every tile again: duplicating a 20 MP photo layer uploaded
330 tiles. When a new resident's surface shares tiles with another resident of the same role,
region, texel kind, format and default pixel, its texture now starts as a copy of that one
(Stats.tiles_shared counts the tiles taken) and the usual per-tile pass fixes up whatever
differs. Layer textures gain COPY_SRC.
The copy is recorded in the frame's encoder, after whatever brought the donor up to date this
frame. Queued writes land before the encoder runs, so both textures are marked `ordered` and
their writes go through the encoder from then on; otherwise a write to either in the same
refresh would be overwritten by the copy or leak into it. A resident whose region only moved
still keeps its own texels (keep_overlap) first.
Duplicating a 20 MP photo layer in perf_scenarios: 25.1 -> 6.3 ms, 330 tile uploads -> 0
(RTX 3090, Vulkan). Paste is unchanged: the pasted layer has new tiles.
Test: duplicates_copy_their_source_texture covers a plain duplicate (no uploads), a
duplicate and its source each painted with one tile in the same refresh, and a duplicate
missing a tile, all against the CPU composite. Without the `ordered` routing the second case
is off by 167.6/255.
The compositor submitted only between 2048x2048 page cells, so one cell of a deep stack
recorded every pass into a single command buffer: 700 full-size layers make 1403 render
passes, and wgpu-core opens two hal command buffers per pass. Metal allows 4096 command
buffers per queue, so a few thousand passes in one cell can lose the device.
run_cell now counts copies and render passes and, when a refresh may submit in between,
submits the batch after FLUSH_OPS (1000) of them, waiting for the previous batch as the
page-cell flush does. Chunk textures, uniforms and bind groups outlive the submit and the
queue keeps the order, so a chunk continues in the next command buffer. A device lost in
such a submit ends the refresh with an error instead of recording more work.
Not measurably slower on Vulkan (RTX 3090): perf_scenarios rows are within noise.
Test: many_layers_submit_inside_one_chunk (700 noise layers on a 64x48 document) needs
flushes > 0 inside one chunk and matches the CPU composite; main records all 1403 passes
without a flush.
A group's effect region and the empty-tile check were computed from
layer_bounds, which for a group is the union of its children's pixels.
A child's outside stroke reaching past that region was cut off on the
canvas (GPU) and in tiled CPU renders, and the group's drop shadow
followed the cut shape.
New photocraft_compose::effect_source_bounds: for a group, the children
grown by their own effect reach. Used by compose::effect_region,
compose::empty_in and gpu::bounds::effect_region.
generated with Claude Code
* Curves: pencil mode, Smooth and Edit points (#3088)
The Curves editor (Image > Adjustments dialog and the adjustment layer's
Properties panel) gains Photoshop's "Edit points to modify the curve" and
"Draw to modify the curve" buttons plus Smooth, with tooltips. A drawn
curve is a 256-level map per channel, evaluated piecewise-linearly.
- doc: Adjustment::Curves gains `pencil` (serde default), with
curve_map / curve_map_outputs helpers.
- compose: curve_map_lut / curve_lut_mode for drawn curves.
- engine: curves params take `pencil` and `smooth`; channels accept
256-level maps; switching modes draws the spline into maps or fits up to
16 points to the drawing. Eyedroppers fit points first.
- io: the PSD `curv` map flag (first byte, psd-tools `is_map`) was read
as padding, so pencil curves from Photoshop were misparsed; they are now
read and written as 256-byte maps, and smart-filter Curves use `Mpng`.
- scorecard UI-217-15 note updated.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* Merge upstream/main: Curves in new compose/gpu HDR tests gets the pencil field (#3088)
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: Brandon Thomas <echelon@gmail.com>
* Preserve HDR range in float adjustments and stack statistics
* Cover HDR exposure round trips and Levels record order
* Report the backend used by GPU parity tests
* Compare float PSD oracles at native depth and preserve signed Exposure
* Retain merged alpha metadata when decoding float references
---------
Co-authored-by: Brandon Thomas <echelon@gmail.com>
Every effect with a contour (Drop/Inner Shadow, Outer/Inner Glow, Satin,
Bevel & Emboss Gloss Contour and Contour element) now shows a contour
thumbnail with a preset picker and an Anti-aliased checkbox. Clicking the
thumbnail opens the Contour Editor: add/drag/delete points (the shared
point-curve editor), Corner per point, Input/Output, Preset, and Load…/Save…
of .shc contour files, read and written from Adobe's public File Formats
Specification (Contours). Edits go to the dialog's pending params, so the
preview follows and OK applies them in the one layer.layerStyle.replace.
- doc: Contour::Custom gains per-point corner flags (serde default).
- compose: contour_curve_lut splits the spline at corner points; CPU and
GPU effects use it.
- engine: the 12 standard presets with our own control points (older names
still accepted); `contour` params also take a curve object
{name, points, corners}; antiAlias params; Satin contour; the bevel
Contour element edits in place; contours_from_shc / contour_to_shc.
- io: PSD contour points read and write the Cnty (continuity) flag.
- psd: shc.rs parser/writer.
- scorecard FILE-215-15 missing -> partial.
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
A dab that painted past a layer's stored tiles changed its resident page
region, which recreated the texture and converted and uploaded every tile
of the layer in that page cell again. Copy the overlap from the old
texture on the GPU instead (in its own submission, ahead of the queued
tile writes) and keep its tile map, so only new or changed tiles upload.
32-bit float RGB/gray tiles also convert to RGBA16F straight from their
samples (same bytes as the general path) instead of decoding into a float
buffer and converting pixel by pixel.
P16 (32-bit, 20 layers) brush dab + refresh, median of 6 alternating runs:
p50 8.9 -> 3.1 ms, p95 27.2 -> 7.5 ms. P13 is unchanged (~1.6 / ~4.5 ms).
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
* Speed up large feathered mask refreshes
* Format preference label for workspace check
* gpu test: allow float noise on a one-code vector-mask parity step
The combined mask is cached as U16 now; one pixel lands exactly one 8-bit
code apart (1.0000075/255), within the <=1/255 parity rule.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: Brandon Thomas <echelon@gmail.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
* perf(gpu): reuse unchanged lower-stack composite chunks
* test(gpu): verify half-float composite checkpoint reuse
* docs: record measured GPU composite reuse gains and limits
* perf(gpu): admit whole grid chunks, serve partial damage as sub-rect copies
Cache entries were keyed by the exact damage rectangle, so brush strokes
(a new partial rect every frame) filled the budget with entries that
never hit again. Entries are now whole chunk-aligned, chunk-sized grid
cells: a capture merges into the cell it overlaps at the rectangle's
offset, a read is served as a sub-rect copy of the covered part, and
coverage is tracked per cell so a resume only happens where every pixel
of the requested rectangle is cached. After a full render, any in-bounds
partial rectangle hits.
Regression test: two series of disjoint strokes inside the same chunks
leave composite_cache_bytes unchanged (no entry-per-rectangle growth),
a full refresh still finds its chunks and a top-layer edit afterwards
reuses them. The old partial-damage test now expects the covered-chunk
hit and checks outside-the-document renders admit nothing.
* docs: record the GPU reuse measurement in target-app-parity.md
The honest assessment moved from roadmap.md to target-app-parity.md on
main; keep main's roadmap and add only the dated measurement.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: Brandon Thomas <echelon@gmail.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
A live gradient's dither adds a fraction of a level of noise, but gradient
fill layers kept it in float. Color Dodge divides by 1 - cs, so under a
bright dodge layer that sub-level noise came out as speckle several levels
high, and channels at 0 were lifted off black (coloured fringes).
Dithered gradient pixels are now rounded to the document depth's levels
(8/16-bit; float documents unchanged), as a painted gradient stores them.
Same rounding in the GPU shader. Regression tests in compose and a GPU
parity case for stacked dithered gradients under Color Dodge.
Claude-Session: https://claude.ai/code/session_01Qpa5z8GksXAFiDavFkmh4R
Co-authored-by: Leel Wasanatha <260480244+kavishkashehan726-sudo@users.noreply.github.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Colorize floored saturation at 25% on both the CPU (compose::adjust::hue_saturation) and the
GPU shader, so dragging Saturation to 0 with Colorize ticked still tinted the image. Photoshop
leaves only the lightness at 0%. The floor is gone from both paths, so 1-24% now tint
progressively too.
The floor also stood in for Photoshop's 25% starting saturation when Colorize is switched on.
The dialog's checkbox already sets that; the hueSaturation params now do the same when a call
turns Colorize on without giving a saturation, so {"colorize": true, "hue": 30} still tints.
Fixes#2326.
Co-authored-by: lloyd <lloyd@stackworx.io>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
* Blend modes: Fill below 100% works inside the blend for Photoshop's special eight
In Photoshop, Fill and Opacity are the same for most blend modes, but not
for eight: Color Burn, Linear Burn, Color Dodge, Linear Dodge (Add),
Vivid Light, Linear Light, Hard Mix and Difference. PhotoCraft treated
Fill as Opacity for all of them, so a Difference layer at 50% Fill gave
(150,90,110) where Photoshop gives (150,60,10).
Measured on Photoshop 27.11 (scripted, 24 cases at Fill 25/50/75% with
two colour pairs): below 100% Fill, Photoshop pulls the layer's colour
toward the mode's neutral colour (black for Color Dodge, Linear Dodge,
Difference; white for Color Burn, Linear Burn; 50% gray for Vivid and
Linear Light) and blends that at full strength; Hard Mix becomes
clamp((b - f(1 - t)) / (1 - f)). Opacity still mixes the result in.
That reproduces 31 of 32 values within 1 level (Hard Mix at 75%: 3).
CPU (psblend::composite_fill, the final blend of plain pixel, shape and
type layers) and GPU (Kernel::Blend takes Fill in the spare uniform slot)
agree; Dissolve keeps Fill as coverage. Layers with effects, adjustment
layers and groups keep the old behaviour for now.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* GPU: composite_fill with a single blend_rgb call, so fs_blend stays small
The first version added a third inlined copy of blend_rgb (every blend
mode) to fs_blend: composite_g's two plus blend_rgb_fill's. Windows CI
(software WARP adapter) then timed out rendering the UI in
arbitrary_rotation_preview, which passes on main; a larger shader
compiles noticeably slower there.
composite_fill now handles every case itself with one blend_rgb call
(Fill as coverage for normal modes, the special-eight blend, Lab Normal
mixing, the text gamma space) and is fs_blend's only composite path,
leaving it with fewer inlined copies than on main. Same results: the 35
GPU parity tests pass.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
- Hue, Saturation, Color and Luminosity no longer clip 32-bit documents to 0..1, so values above 1 survive in the CPU compositor (blend_rgb_hdr now uses blend_rgb_within at f32::MAX) and in the GPU shader (under the existing F_HDR flag). 8/16-bit results are unchanged apart from rounding at white.
- ClipColor returns the grey when the luminosity is outside the in-gamut range, instead of dividing by max − lum rounding noise. The old formula turned colours inside out (Luminosity of a grey of 2 over (0.5, 0.25, 0.125) gave (1.0, 2.32, 2.99)), and CPU and GPU disagreed (Hue over a grey of 2: CPU 1.0, GPU 2.82).
- Tests: 8/16-bit inputs match the ISO 32000 formula, HDR cases in compose, psblend::with_hdr nesting, and a CPU/GPU parity test across depths.
Co-authored-by: hideoneric <hideoneric@users.noreply.github.com>
Squashed onto main, with the maintainer's asks applied:
- Photoshop's blend menus (layers, brush, pencil, gradient, layer style) stay as they are;
the Layers panel and Blending Options list a Paint.NET mode only for a layer that uses one.
- The six Paint.NET mode labels are translated in every catalog, and the blend-mode i18n test
iterates layer_modes().
- pdn::pixels rejects truncated uncompressed (version 1) data before allocating its buffer.
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
An RGB pattern (every built-in one) showed in colour as a pattern fill or
Pattern Overlay in a Grayscale or CMYK document, unlike Edit > Fill, the
export, and Photoshop, which fills in grey but keeps the pattern's colours
(a Grayscale round trip brings them back).
pattern::pattern_rgba projects a pattern's pixels through the document's
mode when painted; the CPU prepared tiles and the GPU pattern textures use
it. Patterns themselves are unchanged.
Co-authored-by: Claude <noreply@anthropic.com>
Co-authored-by: Brandon Thomas <echelon@gmail.com>
The 4096-entry ramp LUT interpolated across a hard stop, showing grey where the CPU export is
white. Gradient fills, Gradient Maps and gradient effects with coincident colour or opacity stops
now return Unsupported so the canvas uses the CPU oracle. Fixes#974.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Changed tiles are grouped by shared Arc and converted once each in parallel batches; a tile
shared by several coordinates (a solid fill, a cleared area, a non-zero default) is uploaded
once to a scratch tile and copied on the GPU, and new pages take the same path. 16-bit RGB and
gray tiles convert to RGBA16F through a lookup table. Select All + Fill on a 50 MP 16-bit page
refreshes in 15-19 ms instead of about 1.4 s. Fixes#1774.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Two results exactly one 8-bit step apart can land a few f32 ulps past 1/255 after premultiplying
and subtracting (NVIDIA, Hue/Saturation in adjustment_layers); allow 4 epsilon of slack. Fixes#1848.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- Linear Dodge (Add) and Divide keep values above 1 in 32-bit float documents, clipped only at f32::MAX. Integer depths still clip at 1.
- CPU: add psblend::HDR and generic::blend_rgb_hdr / blend_channel_hdr; render sets the flag per document depth.
- GPU: add the F_HDR pass flag, and compose.wgsl uses it for the same clip bound.
- Non-separable modes are unchanged.
- Tests: colour unit tests, compose depth tests, and the GPU parity test float_documents_blend_past_white.
Co-authored-by: hideoneric <hideoneric@users.noreply.github.com>
Saturation −100 greyed to the HSL lightness, so a yellow (255, 255, 0) became grey 128 where
Photoshop shows near-white: a yellow logo with Brightness/Contrast +150 and Vibrance › Saturation
−100 clipped to it stayed yellow under Lighter Color. Measured black-box on Photoshop 25.4
(16-bit runs of 7,400-colour charts and HSV sweeps; sRGB, Adobe RGB and ProPhoto documents), both
sliders work in linear light through the document tone curve, Vibrance first, then Saturation:
- Saturation moves each colour from the grey 0.288 R + 0.712 G (blue has no weight, in every
profile) by 1 + s, clipping per channel: within 0.07/255.
- Negative Vibrance is an HSV change in linear light keeping the hue, in closed form: 0.2/255 on
average, the deepest shadows (V < 0.05) drift.
- Positive Vibrance: the brightness term is exact; the saturation boost (stronger for dull and
dark colours, damped on reds and skin) is a fit, 0.4 (+10) to 5 (+100) /255 on average.
CPU (compose::adjust::vibrance_px) and GPU (compose.wgsl) share the constants; the parity test
gains a negative Vibrance and Saturation −100. photocraft-corpus adjustments/rgb8|rgb16/
vibrance.psd (Vibrance +60, Saturation −15): 25.6/255 mean, 89 max → 0.38 mean, 5 max.
Co-authored-by: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Fill 0 % + an Outside stroke is Photoshop's "outline only" look, but on pixel
layers the stroke lay beneath the whole interior at full coverage, relying on
the layer composited on top to hide it. With the fill lowered the interior
showed the stroke colour (a solid disc instead of a ring).
The share beneath an interior pixel is now (1 - a) / (1 - c), with a the
shape's coverage and c the layer's alpha after fill: 1 under an opaque
layer (100 % fill renders exactly as before, so PSD oracles don't move),
1 - a at 0 % fill. CPU compositor and the wgpu shader agree (the layer's
alpha now rides along the coverage seed for pixel layers too).
Tests: compose regression at 8/16/32-bit (0 %, 50 %, 100 % fill), and a GPU
parity case at fill 0 and 0.5 (fails without the shader change).
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
- A layer about the canvas's size keeps its whole effect region even where it overhangs the
edge, so the region keeps its size while the layer is dragged and the GPU moves its effect
maps with it instead of rebuilding the shape, distance fields and every effect each frame.
- Surface::translated moves a surface by copying encoded tile bytes tile to tile (no float
round trip, no full-size intermediate; whole-tile moves share the tiles): 2.3x faster than
to_interleaved + write_interleaved. Used by layer.translate, Move tool drags and
algo::resample::translate_surface.
- perf_scenarios: Move tool drag frame rows (styled layer, full-canvas layer with and without
a drop shadow).
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
* Photo Filter and gradient glows render like Photoshop (#442, #443)
Photo Filter multiplied encoded RGB by the filter colour and rescaled to Rec.601 luma. Photoshop
multiplies the pixel's D50 XYZ (relative to white) by the filter colour's, mixed by density, and
Preserve Luminosity restores the encoded luminosity like the Luminosity blend mode (SetLum +
ClipColor); 32-bit documents keep luminance Y. rgb16 photo-filter.psd now matches within 0.6/255
(was 125/255); rgb8 mean error 0.3/255 (was 20.8).
Outer and Inner Glow gradients were laid out across the canvas (a vertical linear gradient), so
the colour bands ran top to bottom. Photoshop colours the glow at strength v with the gradient at
1 - v (first stop at the edge) and the glow is opaque from v = range^2. outer-glow-gradient.psd:
mean 1.4/255, max 9.9 (was 252).
CPU and GPU paths updated together; photoshop oracle floor 132 -> 133.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* Merge origin/main; regenerate docs/scorecard.md
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: Brandon Thomas <echelon@gmail.com>
Rebased onto the paged GPU compositor (#187): the first-frame empty submit moves into encode_inner, and the test GPU lock covers the paged compositor too.
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
* Adjustments: full parameter sets for every kind, Hue/Sat ranges, CMYK/Lab tone channels
- doc: Hue/Saturation gains six colour ranges (HueRange), Gradient Map a dither flag,
Levels/Curves a ToneSpace (RGB, CMYK inks + black, Lab) for mode-native channels.
- engine: new adjust_params module (from_params over a base / to_params) used by
image.adjustments.*, layer.newAdjustmentLayer.* and layer.setAdjustment. Curves,
Color Balance, Channel Mixer, Black & White, Gradient Map and Photo Filter now honour
their params; bad params are BadParams errors; setAdjustment merges (empty = reset).
- compose/gpu: hue-range tables, CMYK/Lab tone channels (CPU; GPU falls back), dithered
Gradient Map, Black & White from grey/secondary/primary parts; GPU parity extended.
- io: hue2 ranges, grdm dither, CMYK/Lab Levels/Curves records round-trip.
- destructive adjustments use the document's tone transfer.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* Adjustments UI: one editor per kind for Properties and Image › Adjustments dialogs
- adjust_editors: Brightness/Contrast (legacy), Levels and Curves (per channel in RGB, Gray,
CMYK and Lab, histogram behind), Exposure, Vibrance, Hue/Saturation (master + six ranges with
draggable range sliders, colorize, before/after spectrum), Color Balance (tones, preserve
luminosity), Black & White (six weights, tint colour), Photo Filter (presets or colour,
density), Channel Mixer (outputs, constant, monochrome, total), Posterize, Threshold
(histogram), Gradient Map (stops, presets, reverse, dither). Replaces the M7 placeholder.
- Properties previews live (GPU) while a control moves and commits one setAdjustment per gesture.
- adjust_dialog: Image › Adjustments opens the same editor modally; the canvas previews the
real command on the proxy, Preview checkbox, OK = one history step, Cancel leaves the doc.
- Curves (#43): hit-test at the press origin, Delete/Backspace removes the selected point when
the graph has focus, Cmd/Ctrl-click removes, dragging off removes (and back on restores),
arrows nudge; endpoints and the last two points stay; pressing the gradient bars does nothing.
- tone: histograms per channel space, of the image below a layer or of the target layer alone.
- kittests: every adjustment layer kind edited through the Properties panel with real input at
8/16/32-bit RGB, Gray, CMYK and Lab (#12); Curves point interactions; dialog preview/OK/Cancel.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* io tests: adjustment round trip covers the new Levels/Curves/Hue-Sat/Gradient Map fields
#60's round-trip tests predate #68's doc-model additions (tone space and black channel,
Hue/Saturation ranges, Gradient Map dither); fill them with their defaults.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Once Format stopped failing first, the Linux runner reached the GPU tests
for the first time: its software adapter (GL) can't render Rgba32Float or
R32Float. The app already picks Rgba16Float there, but the effect-map
pipelines (R32Float/R16Float) still panicked. They are now built inside an
error scope; where they can't be, documents with layer effects use the CPU
compositor and everything else stays on the GPU.
The exact-parity tests skip on adapters without 32-bit float targets (the
app uses the CPU compositor there), like they already skip without an
adapter; the fallback-path tests build with try_new_with_format.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
On D3D12 without DXC, wgpu compiles HLSL with FXC, which aborted on
fs_adjust ("internal error: compilation aborted unexpectedly"). wgpu turns
that into a panic, so every GPU test failed on the Windows runner, and the
app would crash where the same path is taken. A bisect on the Windows
runner pinned it to two cases of adjust()'s switch, Exposure and Selective
Color: both looped over dynamically indexed vector components. They are
now written per channel / per range with the same math (GPU-vs-CPU parity
is unchanged and passes on Metal and on the Windows runner).
Compositor::try_new_with_format catches validation and internal errors
while building the pipelines and returns Err; the canvas then uses the CPU
compositor (once, remembered) instead of crashing.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
CI's `cargo fmt --all -- --check` has failed on every run because the
repo had no rustfmt config and was never formatted. Use the same settings
as EffectCraft/FilmCraft (max_width 160, use_small_heuristics Max), which
also keeps the churn small. Formatting only; no code changes.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Some drivers (e.g. Intel Vulkan on Linux/Windows) can't render to Rgba32Float,
so creating the 'fs_content' pipeline panicked with 'Format Rgba32Float is not
renderable' as soon as a document/image was opened. The accumulation format is
now chosen per adapter: Compositor::preferred_acc_format() keeps Rgba32Float
where it's a valid render target and falls back to Rgba16Float otherwise. The
compositor carries the format (pipelines, accumulation texture, byte size and
the readback half-decode all follow it); the app picks it from the eframe
adapter. Capable GPUs are unchanged (parity still exact, 18/18). A new test
forces the Rgba16Float path on any GPU and confirms it renders within display
tolerance (blend + adjustment). wasm green.
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
