A media clip whose enabled standard effects all have a GPU implementation
plans as a GPU layer carrying them (render::plan::LayerFx): parameters are
evaluated on the CPU at the frame's time (render::gpufx::FxOp::eval, shared
with the CPU path so both clamp hostile values identically), the source is
drawn into an Rgba32Float working image at the size the CPU decodes it, the
effects run as WGSL compute passes (gpu::fx, fx.wgsl) with the CPU
reference's math, and the result is placed with Motion, opacity and blend
mode as before. Layers without effects keep the single draw.
GPU effects: Brightness & Contrast, ProcAmp, Tint, Black & White, Color
Balance, Leave Color, Change to Color, Color Pass, Color Replace, Channel
Mix, ASC CDL, Gamma Correction, Levels, Extract, Invert, Posterize, Alpha
Adjust, Gaussian / Camera / Directional Blur (+ legacy aliases), Sharpen,
Unsharp Mask, Crop, Edge Feather, Transform, H/V Flip, Mirror, Offset.
Masked effects, opacity masks, non-finite parameters and minifying
Transform/Crop zooms stay on the CPU. Devices without compute shaders run
the effects on the CPU inside the compositor.
Also fixes a debug-build overflow in Image::transformed for non-finite maps.
Parity tests (exact f32 checks and composited plans), a golden test that
plans effect chains as GPU layers, and the stack3-fx playback benchmark.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
All 27 blend modes now composite on the GPU instead of sending the
whole frame to the CPU.
- render::plan: PlanLayer carries its Blend; plan_frame no longer falls
back to a CPU frame for non-Normal modes. Media, graphic clips,
multicam angles and CPU-rendered layer images (standard effects) keep
their clip's mode; inside a transition the mode is Normal, as in the
CPU reference. execute_cpu composites with each layer's mode.
- gpu: Normal keeps the single fixed-function pass. Dissolve also uses
fixed-function "over": the CPU's 64-bit pixel hash is reproduced in
WGSL on u32 pairs (pattern bit-exact). The other 25 modes copy the
accumulator region under the layer's quad into a backdrop texture and
composite in fs_blend with the CPU math (sRGB-encoded straight colour,
W3C formulas, back to linear, same premultiplied mix and clamping).
- Tests: per-mode GPU-vs-CPU parity (gpu crate: overlapping transformed
layers, partial opacity/alpha, empty/partial backdrop; golden: every
mode through plan_frame incl. an effect layer, blended clips in a
cross dissolve), exact-value edge cases (black/white, Divide by 0,
Dodge/Burn at 0 and 1, zero alpha) within 4.9e-4, Dissolve pattern
exact, off-output layers, quad bounds.
- Bench: new stack3-blend playback scenario (V2 Screen, V3 Overlay);
CPU ms/frame 151-214 -> 86-89, Full 19/173 and 138/54 -> 191/1, 192/0.
- Docs: architecture §5, gpu crate docs, testing, performance.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Rust/clippy 1.99 added `chunks_exact_to_as_chunks` and tightened
`needless_late_init` and `drain_collect`, failing `clippy -D warnings`
on main.
- `chunks_exact(N)` / `chunks_exact_mut(N)` with a constant N become
`as_chunks::<N>()` / `as_chunks_mut::<N>()` (fixed-size arrays, so
several byte reads drop their per-element indexing). relink's
fingerprint uses the `(chunks, remainder)` split; its hash output is
unchanged.
- Late-initialised `let`s in ac3 bit allocation and opus CELT
bands/rate become `let x = if/match …`.
- `dpb.drain(..).collect()` becomes `std::mem::take(&mut dpb)`.
No behaviour change; `cargo xtask ci` passes.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- gpu: `composite` returns its output view by value (a cheap handle clone) taken from the
render-target setup, so the accumulation/output views no longer `expect` their slots.
- render: lightning bolt subdivision, tracker pyramid and the Lumetri preset instance without
`expect`.
- cli: dotted parameter insertion without `expect`; the render/mcp backend arms report a usage
error instead of `unreachable!`.
- Each crate denies unwrap/expect/panic/unimplemented/todo/unreachable outside tests.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- Lumetri gains input_lut (Basic Correction) and look_lut (Creative) references: lib:<id> for
LUTs imported into the project's library (embedded .cube/.3dl text, parsed once and cached by
content), builtin:<id> for LUTs generated from code: camera log → Rec. 709 conversions from the
published curves, and the procedural Creative looks (no third-party LUTs)
- section switches basic_on/creative_on/curves_on/wheels_on/vignette_on (+ hsl_on) bypass their
sections; older projects without them render as before
- commands lut.import, lut.list, lut.remove, lut.export (cube/3dl), lumetri.setInputLut,
lumetri.setLook (lut ref or path; imports on the fly), lumetri.setSection; effects.setParam
fills in parameters missing from instances saved by older builds
- Lumetri panel: Input LUT and Look LUT menus (None, built-ins, library, Browse… file dialog),
working section switches (automation ids lumetri.input_lut, lumetri.look_lut,
lumetri.switch.<section>); combo boxes no longer widen the panel
- filmcraft-gpu: GpuLut WGSL compute pass with the same tetrahedral interpolation; parity test
vs Lut3d::apply for 2³–65³ and a non-unit domain (max |Δ| 1.8e-7)
- tests: engine (import/apply/bypass/undo/save, export→import round trips for both formats),
render (built-ins reproduce the looks, library resolution), headless UI (color_ui.rs)
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
GpuCompositor::upload converted f32 RGBA layers (effects rendered on the CPU, captions,
whole-frame fallbacks) and 16-bit YUV frames (ProRes render previews) to half floats on the UI
thread every time a frame was shown: 8.3 M values per 1080p RGBA layer, 4 M per ProRes frame.
filmcraft_gpu::prepare(plan) now does those conversions on the frame worker that built the plan
(parallel chunks; 16-bit codes through a cached per-bit-depth table) and the plan cache stores
them with the plan (frames::GpuPlan), bounded by count and bytes (1 GiB). composite_prepared()
only copies bytes into textures. The texels are bit-identical (new test compares the read-back
output of both paths for f32, 10-bit 4:2:2 and fallback-image plans).
bench, same-session A/B (GPU path, after-preview Full, 2 rounds, load avg 50–95):
UI-thread present time p50/p95: plays from previews 11–31 / 34–77 ms → 3.6–5.8 / 4–14 ms;
live effects play 17–28 / 31–56 ms → 4.1 / 12 ms (first run)
plays from previews: 184/8, 183/9, 189/3, 169/23 → 192/0, 192/0, 191/1, 192/0 shown/dropped
live effects play (CPU-bound, ~6 cores at this load): too noisy to compare (0–117 shown either way)
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
crates/golden (test-only, L3): nine procedural 320x180 scenes (transform
and opacity, blend modes, Gaussian Blur, Lumetri basic, Crop, Cross
Dissolve / Dip to Black / Wipe mid-transition, timecode/clip-name text)
rendered on the CPU and compared with committed goldens (PSNR >= 45 dB,
max <= 12, p99 <= 2); FILMCRAFT_BLESS=1 regenerates them with
attribution sidecars. testkit::golden adds PNG I/O, diff metrics and
failure images. The same scenes are composited on the GPU and compared
with the CPU.
GPU parity exposed a real bug: GpuCompositor cached uploads by
pixel-buffer address without owning the buffer, so a new frame
allocated at a freed frame's address was drawn with a stale texture
(whole wrong frames, PSNR ~12 dB). The cache now keeps the buffer
alive; regression test upload_cache_keeps_buffers_alive.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
