Merge PR #805

# Conflicts:
#	docs/camera-preview-colour.md
#	docs/parity-checklist.md

crates/engine/src/camera_preview.rs

@@ -332,12 +332,19 @@ fn at_most(a: &CameraTone, b: &CameraTone) -> bool {
     })
 }
 
+/// Pixel cap for the embedded JPEG decoded for the fit. The codec checks it before decoding and
+/// decodes at full size, so a 66.5 MP full-frame ILCE-7RM6 JPEG transiently needs ~200 MB (3 B/px);
+/// 128 MP also admits 100 MP medium-format previews.
+const FIT_PREVIEW_MAX_PIXELS: u64 = 128_000_000;
 /// Same-size proxies of the sensor (white-balanced, baseline exposure, through `transform`'s
 /// matrix: the generic camera ≈ sRGB model) and of the file's embedded camera JPEG (linear Rec.2020),
 /// and per proxy pixel whether any sensor sample it covers is clipped.
 fn proxies(raw: &RawImage, bytes: &[u8], transform: &CameraTransform, size: usize) -> Option<(Rgb32f, Rgb32f, Vec<bool>)> {
     let edge = (2 * size).max(384) as u32;
-    let decoded = crate::files::decode_raw_preview(bytes, lightcraft_codecs::DecodeOptions { max_size: Some((edge, edge)), max_pixels: 64_000_000 })?;
+    let decoded = crate::files::decode_raw_preview(
+        bytes,
+        lightcraft_codecs::DecodeOptions { max_size: Some((edge, edge)), max_pixels: FIT_PREVIEW_MAX_PIXELS },
+    )?;
     let mut reference = decoded.to_working();
     let (a, crop) = (raw.active_area, raw.crop.clipped(raw.active_area.width, raw.active_area.height));
     if crop.width == 0 || crop.height == 0 || reference.width == 0 || reference.height == 0 {
@@ -3174,4 +3181,20 @@ mod tests {
         reference.data.iter_mut().for_each(|p| *p = [0.4; 3]);
         assert!(fit_tone_only(&sorted, &reference, None).is_none(), "a flat JPEG");
     }
+
+    /// ILCE-7RM6 full-frame: the 66.5 MP camera JPEG exceeded the old 64 MP cap, so no look was fitted.
+    #[test]
+    fn corpus_arw6_full_frame_gets_a_camera_look() {
+        const { assert!(FIT_PREVIEW_MAX_PIXELS >= 100_000_000) };
+        let path = std::env::var_os("LIGHTCRAFT_CORPUS")
+            .map(std::path::PathBuf::from)
+            .unwrap_or_else(|| std::path::Path::new(env!("CARGO_MANIFEST_DIR")).join("../../corpus"))
+            .join("raw/arw6-sony-ilce7rm6-ff-land-c.arw");
+        let Ok(bytes) = std::fs::read(&path) else {
+            eprintln!("skip: {} absent", path.display());
+            return;
+        };
+        let (_, info) = crate::files::load_bytes(&bytes, 400).unwrap();
+        assert!(info.camera_tone.is_some(), "no camera look fitted");
+    }
 }

docs/camera-preview-colour.md

@@ -5,6 +5,7 @@ ARW, NEF, RW2, RAF, CR3, CR2, PEF, SRW and ORF decoding supply sensor data and c
 ## Colour and tone are separate
 
 A fixed sensor proxy of at most 96×96 is developed with the file's WB, black/white levels and active crop. The embedded JPEG is decoded with DCT scaling and a 64-million-pixel input limit, converted through the codec's colour management, and reduced to the same size. Aspect ratios must agree within 2%. For colour fitting, near-black, clipped and nonfinite samples are excluded. Learning a colour matrix requires at least 256 paired pixels, 5% coloured samples (linear channel spread >0.05) and a nonsingular input covariance with bounded condition. An existing camera profile supplies the colour model: fitting only tone still requires 256 pairs and 5% non-monochrome samples, but permits channel spread >0.005 so dark or dull scenes can use the known colours.
+A fixed sensor proxy of at most 96×96 is developed with the file's WB, black/white levels and active crop. The embedded JPEG is decoded with DCT scaling and a 128-million-pixel input limit (61 MP bodies' JPEGs, e.g. the ILCE-7RM6 full frame at 66.5 MP, exceed 64 MP; the codec checks the cap before decoding at full size, about 200 MB transient), converted through the codec's colour management, and reduced to the same size. Aspect ratios must agree within 2%. Near-black, clipped and nonfinite samples are excluded. Learning a colour matrix requires at least 256 paired pixels, 5% coloured samples (linear channel spread >0.05) and a nonsingular input covariance with bounded condition. An existing camera profile supplies the colour model: fitting only tone still requires 256 pairs and 5% non-monochrome samples, but permits channel spread >0.005 so dark or dull scenes can use the known colours.
 
 Two thirds of the usable pixels train a ridge-regularised 3×3 **chromaticity** correction (luminance-normalised RGB). This avoids trying to encode the camera's nonlinear tone curve in a matrix, which left skin grey in the earlier matrix-only estimate. Coefficients are bounded by 8. A matrix can't follow the camera's hue-dependent rendering, though: on a photo of a lime shirt among grey rocks the best matrix rendered the shirt olive (16° hue error), because the few saturated pixels have little weight against the many dull ones, and a matrix fitted to the shirt alone breaks everything else. So a **hue/saturation table** (5 values × 72 hues × 5 saturations, the DNG `ProfileHueSatMap` layout with an sRGB-encoded value axis, applied in linear ProPhoto RGB after the baseline exposure) is fitted to the residual hue shifts and saturation ratios after the matrix: Gaussian kernels around each node (2.5° hue, 0.15 saturation, 0.12 encoded value), weighted by saturation, shrunk toward identity where the photo has few samples, saturation 0 left as identity so neutrals stay neutral. The value axis matters: on a sunny waterfall scene dark yellows had to turn ~12° toward orange while brighter yellow-greens turned ~9° toward green, and one shift per hue made both worse. The table restores each pixel's luminance (the tone curve owns luminance; without that, saturation boosts darkened colours and the table lost on held-out pixels).
 
@@ -98,6 +99,14 @@ On the public CC0 EOS R100 / RF-S 18–45 mm sample, the measured framing scale
 
 This is a per-file camera-look estimate, **not measured spectral calibration or Lightroom parity**. Camera local tone, picture styles, noise reduction and lens warps cannot all be reproduced by a global matrix/curve; same-aspect crops may evade the aspect check. Skin, mixed-light scenes and saturated highlights can still differ from a supplied JPEG. Broader Sony, Nikon and Canon model/lighting coverage, chart-based calibration, absolute WB and other manufacturers remain unverified. Rejected fits and other RAW formats retain their existing behaviour.
 
+### Sony ILCE-7RM6 (A7R VI)
+
+Full-frame A7R VI files never reached the fit: their 66.5 MP embedded JPEG exceeded the fit's former 64 MP preview cap, so they opened in the neutral fallback. The cap is now 128 MP. On the full-frame corpus file the holdout RMS went from 0.124 to 0.062, and the render's mean RGB from 85/87/77 to 104/117/79 against the camera JPEG's 108/122/82.
+
+On a local set of 54 ARW6 and lossless ARW photos (the maintainer's own shots plus the 12 CC0 RawDB ILCE-7RM6 samples), 46 were fitted and 8 were rejected by the gates; the kinds of scene behind the rejections were not analysed. On 15 withheld photos (every third usable photo in sorted path order), mean CIE76 ΔE against the camera JPEG was 5.7 (render at 1200 px, comparison at 200 px, matching orientation); in a separate sanity check at the same settings, the APS-C corpus file scored 5.1 and two portrait RawDB files 4.6 and 6.0 (orientation-matched). These numbers were measured before the ARW6 companding curve was replaced by its clean-room derivation (`docs/arw6-compression.md`), which decodes highlights about 0.3 % lower; the fit was not re-measured.
+
+A pooled profile was tried and is not bundled. Pooling the 46 photos (158,549 colour pairs) and validating on withheld photos (every third photo, N = 15, trained on the other 31) took mean ΔE from 5.71 with per-file fitting to 6.40 with the profile: 11 of 15 photos regressed, 3 improved and 1 was unchanged, the largest by +4.68, and 2 of the 8 rejected files were newly accepted. Per-file fitting therefore ships for this body. The 42 private photos are one photographer's scenes, 38 of them in the Standard creative style (the 12 RawDB files were not inventoried); revisit when a varied multi-photographer set exists.
+
 ## Camera profiles (pooled per model)
 
 One photo can show too little of a colour for its own fit to learn it: in a second shot of the lime shirt only a few dozen proxy pixels show it, next to a hillside of foliage at the same hue that the camera renders differently, and the shirt stayed yellow (hue 68° against the JPEG's 82°; a 192-pixel proxy only reached 74°). `lightcraft-cli calibrate [--max N] [--out DIR] FOLDERS…` therefore pools the colour pairs of many ARWs, NEFs or RAFs per camera model — each against its own embedded JPEG, at a 192-pixel proxy, at most 4000 pairs per photo, files spread evenly over the folders — and fits one matrix (from white-balanced camera RGB, so no single photo's white point is baked in) and one hue/saturation/value table. Models need at least 5 usable photos.

docs/parity-checklist.md

@@ -245,6 +245,7 @@ differences and missing bindings are in [ui-parity.md](ui-parity.md#shortcuts).
 | LR-PROF-ADAPTIVE | Adaptive profiles | P2 | ⬜ | | |
 | LR-PROF-CAMERA | Camera-matching looks | P2 | ⬜ | | |
 | LR-PROF-CAMERACOLOR | Camera colour calibration (own) | P0 | 🟡 | `crates/raw/src/color.rs`, `crates/raw/src/profile.rs`, `crates/raw/src/gaintable.rs`, `crates/engine/src/camera_preview.rs` | DNG files use the colour matrices and the camera-profile look they carry (`ProfileHueSatMap`, `ProfileLookTable`, `ProfileToneCurve`, read from the file at run time per DNG spec ch. 6 — Lightroom-converted DNGs rendered flat and muted without them, issue #138). A DNG's `ProfileGainTableMap` / `ProfileGainTableMap2` (Apple ProRAW's local tone map; both versions, DNG 1.7 precedence, the file's byte order with a big-endian fallback) is read and written back by DNG export and not rendered by default, as Lightroom Classic renders ProRAW without it (measured on one CC0 iPhone 12 Pro ProRAW); the Profile option "Camera local tone mapping" (`profile.camera_local_tone`, per photo, off by default, `crates/pipeline/src/local_tone.rs`) renders it as the camera does, per output pixel in the scene-linear stage (CPU, also under the GPU path); Sony ARW, Nikon NEF, Panasonic RW2, Fujifilm RAF, Canon CR3 and CR2, Pentax PEF, Samsung SRW and Olympus ORF get a guarded file-local fit to their own embedded JPEG — chromaticity matrix + hue/saturation/value table + tone and chroma curves, relative WB — or, with a bundled or local per-model profile pooled from many photos (`lightcraft-cli calibrate`, ARW, NEF and RAF so far, `crates/engine/src/camera_profiles.rs`), the profile's colour and only their own tone/chroma curves (on 29 held-out ILCE-7M4 photos mean ΔE vs the camera JPEG 3.54 → 3.10; bundled ILCE-7CR from 125 photos: 26 held-out 4.07 → 3.29, and every one of the 131 photos accepted, 8 of which fell back to neutral without it), kept with smart previews (see docs/camera-preview-colour.md); the tone fit also uses dark JPEG samples excluded from colour training, so deep shadows follow the camera toe (#1307); Sony ARWs shot with DRO on (maker note `0xb025`, on by default) keep that colour but their tone curve is lowered to Sony's curve without DRO, whose lifted shadows are in the camera JPEG only (issue #244: 8 ILCE-7CR photos +6.4 → +0.8 L* from Lightroom's default, median ΔE 7.5 → 3.3); 52 models without a profile whose JPEG fit fails (or that have no usable JPEG) start from DNG-style matrices fitted to their measured spectral sensitivities instead of the neutral fallback (`crates/raw/src/spectral.rs`, rawtoaces-data v1.1.0, Apache-2.0, aliases included; not yet measured against the camera JPEG); ARW: on 10 public raw.pixls.us samples from 8 bodies 7 fits were accepted, mean ΔE vs the camera JPEG 17–26 → 3–10; NEF: 13 of 13 decodable samples (D750, D780, D850, D7500, Z 50 and local D7500 shots) accepted, ΔE 13–46 → 3–8 (issue #150); RW2 / RWL: 140 of 174 public samples accepted, from 93 of the 114 bodies with a preview, median ΔE 5.0 with matching lightness (13 of the 14 corpus files with a preview: ΔE 9.6–27 → 2.8–8.9; most rejections are compacts and kit zooms whose camera JPEG is distortion-corrected); a fit that fails the gates on all pixels is retried away from edges, where a lens-corrected camera JPEG still lines up (issue #232: the six distortion-corrected public ILCE-7RM2 samples ΔE 16.6–16.9 → 3.7–3.8, the corpus GX80 19.1 → 7.9); RAF: bundled X-H2S/X-T4 profiles from 201/545 photos; on 150 withheld files mean ΔE improves slightly over per-file fitting (X-H2S 1.70 → 1.62, X-T4 2.43 → 2.34); 23 of 24 references from 13 bodies accepted, including the dark X-H2S samples (8.74 → 2.25); ORF: the 3 decodable corpus files (E-1, E-400, XZ-2) accepted (`docs/camera-preview-colour.md`); a raw whose camera JPEG is black and white (a monochrome picture style) takes only the JPEG's tone curve, on the camera profile's, spectral or neutral colour (issue #577); this is not measured camera calibration; other raws and rejected fits use a neutral fallback (camera RGB ≈ linear sRGB, `matrix_is_fallback`) with as-shot white balance, so colours are muted and not accurate. Needs our own per-camera calibration: matrices the files carry themselves (e.g. Olympus `ColorMatrix`), fitting each camera to its own embedded JPEG, then chart shots. Adobe matrices / DCPs are never used. Biggest image-quality gap today |
+| LR-PROF-CAMERACOLOR | Camera colour calibration (own) | P0 | 🟡 | `crates/raw/src/color.rs`, `crates/raw/src/profile.rs`, `crates/raw/src/gaintable.rs`, `crates/engine/src/camera_preview.rs` | DNG files use the colour matrices and the camera-profile look they carry (`ProfileHueSatMap`, `ProfileLookTable`, `ProfileToneCurve`, read from the file at run time per DNG spec ch. 6 — Lightroom-converted DNGs rendered flat and muted without them, issue #138). A DNG's `ProfileGainTableMap` / `ProfileGainTableMap2` (Apple ProRAW's local tone map; both versions, DNG 1.7 precedence, the file's byte order with a big-endian fallback) is read and written back by DNG export and not rendered by default, as Lightroom Classic renders ProRAW without it (measured on one CC0 iPhone 12 Pro ProRAW); the Profile option "Camera local tone mapping" (`profile.camera_local_tone`, per photo, off by default, `crates/pipeline/src/local_tone.rs`) renders it as the camera does, per output pixel in the scene-linear stage (CPU, also under the GPU path); Sony ARW, Nikon NEF, Panasonic RW2, Fujifilm RAF, Canon CR3 and CR2, Pentax PEF, Samsung SRW and Olympus ORF get a guarded file-local fit to their own embedded JPEG — chromaticity matrix + hue/saturation/value table + tone and chroma curves, relative WB — or, with a bundled or local per-model profile pooled from many photos (`lightcraft-cli calibrate`, ARW, NEF and RAF so far, `crates/engine/src/camera_profiles.rs`), the profile's colour and only their own tone/chroma curves (on 29 held-out ILCE-7M4 photos mean ΔE vs the camera JPEG 3.54 → 3.10; bundled ILCE-7CR from 125 photos: 26 held-out 4.07 → 3.29, and every one of the 131 photos accepted, 8 of which fell back to neutral without it); ILCE-7RM6: per-file fit now runs on previews above 64 MP, a pooled profile (46 photos) was tried and not bundled (withheld ΔE 5.71 → 6.40, 11 of 15 regressed), kept with smart previews (see docs/camera-preview-colour.md); Sony ARWs shot with DRO on (maker note `0xb025`, on by default) keep that colour but their tone curve is lowered to Sony's curve without DRO, whose lifted shadows are in the camera JPEG only (issue #244: 8 ILCE-7CR photos +6.4 → +0.8 L* from Lightroom's default, median ΔE 7.5 → 3.3); 52 models without a profile whose JPEG fit fails (or that have no usable JPEG) start from DNG-style matrices fitted to their measured spectral sensitivities instead of the neutral fallback (`crates/raw/src/spectral.rs`, rawtoaces-data v1.1.0, Apache-2.0, aliases included; not yet measured against the camera JPEG); ARW: on 10 public raw.pixls.us samples from 8 bodies 7 fits were accepted, mean ΔE vs the camera JPEG 17–26 → 3–10; NEF: 13 of 13 decodable samples (D750, D780, D850, D7500, Z 50 and local D7500 shots) accepted, ΔE 13–46 → 3–8 (issue #150); RW2 / RWL: 140 of 174 public samples accepted, from 93 of the 114 bodies with a preview, median ΔE 5.0 with matching lightness (13 of the 14 corpus files with a preview: ΔE 9.6–27 → 2.8–8.9; most rejections are compacts and kit zooms whose camera JPEG is distortion-corrected); a fit that fails the gates on all pixels is retried away from edges, where a lens-corrected camera JPEG still lines up (issue #232: the six distortion-corrected public ILCE-7RM2 samples ΔE 16.6–16.9 → 3.7–3.8, the corpus GX80 19.1 → 7.9); RAF: bundled X-H2S/X-T4 profiles from 201/545 photos; on 150 withheld files mean ΔE improves slightly over per-file fitting (X-H2S 1.70 → 1.62, X-T4 2.43 → 2.34); 23 of 24 references from 13 bodies accepted, including the dark X-H2S samples (8.74 → 2.25); ORF: the 3 decodable corpus files (E-1, E-400, XZ-2) accepted (`docs/camera-preview-colour.md`); a raw whose camera JPEG is black and white (a monochrome picture style) takes only the JPEG's tone curve, on the camera profile's, spectral or neutral colour (issue #577); this is not measured camera calibration; other raws and rejected fits use a neutral fallback (camera RGB ≈ linear sRGB, `matrix_is_fallback`) with as-shot white balance, so colours are muted and not accurate. Needs our own per-camera calibration: matrices the files carry themselves (e.g. Olympus `ColorMatrix`), fitting each camera to its own embedded JPEG, then chart shots. Adobe matrices / DCPs are never used. Biggest image-quality gap today |
 | LR-PROF-CREATIVE | Creative profiles (own) | P2 | ✅ | `cmd:develop.profile`, `crates/pipeline/src/profiles.rs` | 16 own looks in Film / Cinematic / Muted / B&W (tone + point-curve fades, colour grading, mixer / B&W mix); scale with `ctl:profile.amount`; sliders untouched |
 | LR-PROF-LEGACY | Legacy profiles | P2 | ⬜ | | |
 | LR-PROF-NONRAW | Profiles for non-raw files | P0 | ✅ | `cmd:develop.profile` | same looks apply to JPEG/TIFF |