LR-IMP-PREVIEW-READ: an unedited raw's grid thumbnail is its embedded JPEG

The grid already showed the camera's embedded preview first, then queued a render of the whole raw
to replace it. For a 512 px thumbnail that render decodes 50 of a 56 MB Fuji RAF's bytes for a
picture the 4 416 × 2 944 embedded JPEG already covers — 45 s a photo on the 1.2 MB/s share this was
measured on, 8.3 GB of traffic for a 149-photo folder, and a queue that delayed the render the user
was actually waiting for (measured: the first loupe render of a cold photo finished 174 s after the
click, most of it spent behind thumbnails). Idle for a minute afterwards, the same folder read 12
previews (26.9 MB) and decoded zero raws for thumbnails, against 638 MB of whole-file reads before.

- `Session::embedded_thumb_is_final` answers whether a photo is an unedited raw whose camera JPEG the
  viewer is showing — the same `has_import_look` gate the preview path uses.
- `grid::skips_the_real_render` is the decision (unit-tested): a cached render as stand-in, an edited
  photo and a raw without an embedded JPEG all still render, because their thumbnail must show the
  pipeline's output rather than the camera's.
- `Renderer::remember_embedded_thumb` records the quick render's key and the bucket, and
  `thumb_quick_is_final` / `thumb_quick_pending` let the per-frame grid request stop at the embedded
  preview instead of rebuilding the render job (and its cache key) every frame.
- The thumbnail key still follows the settings hash, so editing a photo re-renders its thumbnail
  through the ordinary background path.

`docs/raw-preview-reads.md` and the `LR-IMP-PREVIEW-READ` parity row record the behaviour and the
numbers.

Co-Authored-By: DeepSeek V3.2 <[email protected]>

crates/engine/src/media.rs

@@ -1212,6 +1212,23 @@ impl crate::Session {
         }
     }
 
+    /// Whether `p`'s thumbnail can be the camera's own JPEG, so a frontend can skip the sensor
+    /// decode behind it.
+    ///
+    /// The grid shows this JPEG first in any case ([`QuickJob`]: "grid thumbnails of unedited raws
+    /// likewise start from the embedded preview"), and then paid for a render of the raw to replace
+    /// it. That render decodes the whole sensor — 50 of a 56 MB Fuji RAF's bytes — for a 512 px
+    /// thumbnail the 4 416 × 2 944 embedded JPEG already covers, and on a share at 1.2 MB/s it is
+    /// 45 s a photo: a 149-photo folder spent 8.3 GB of traffic on thumbnails, and the queue behind
+    /// it delayed the render the user was waiting for.
+    ///
+    /// Only unedited raws answer `true`: any edit must be rendered, and `has_import_look` is what
+    /// says the photo still looks the way it did at import. A photo the user edits re-renders
+    /// through the ordinary path, because its thumbnail key (settings hash) changes.
+    pub fn embedded_thumb_is_final(&self, p: &Photo) -> bool {
+        self.embedded_of(p).is_some()
+    }
+
     /// The camera's embedded preview of a raw file, for the face scan. Faces are read from the unedited picture, so the
     /// photo's edits do not matter, and the camera's own JPEG (oriented, display-ready) costs a fraction of decoding the
     /// raw. `None` for anything else, and where the host installed no loader.

crates/ui-egui/src/panels/grid.rs

@@ -444,8 +444,10 @@ fn group_header(app: &mut LightcraftApp, ui: &mut egui::Ui, run: &DateRun, ids:
 }
 
 /// Request a grid/filmstrip thumbnail at `priority`. An unedited raw without a thumbnail texture
-/// first gets a stand-in (its cached thumbnail, else its embedded camera preview), and its real
-/// render then follows in the background.
+/// gets its embedded camera preview, and the render of the raw behind it follows in the background
+/// — except where that render would only redraw what the embedded JPEG already shows (see
+/// [`lightcraft_engine::media::Session::embedded_thumb_is_final`]), which for a folder of raws on a
+/// network share is the difference between reading every file whole and reading a preview each.
 pub fn request_thumb(app: &mut LightcraftApp, id: PhotoId, size: usize, priority: u32) {
     let bucket = lightcraft_engine::media::thumb_bucket(size);
     if let Some(photo) = app.session.catalog.photo(id)
@@ -453,19 +455,35 @@ pub fn request_thumb(app: &mut LightcraftApp, id: PhotoId, size: usize, priority
     {
         return;
     }
+    // The camera's embedded preview is this photo's thumbnail (unedited raw): once its quick render
+    // has arrived there is nothing left to ask for. Without this the grid would rebuild the render
+    // job — and its cache key — on every frame, which is what the embedded-thumbnail record settles.
+    if let Some(photo) = app.session.catalog.photo(id)
+        && (app.renderer.thumb_quick_is_final(photo, bucket, priority) || app.renderer.thumb_quick_pending(photo, bucket))
+    {
+        return;
+    }
     let Some(job) = app.session.thumb_job(id, size) else { return };
     #[cfg(test)]
     {
         app.renderer.thumb_jobs_built += 1;
     }
     let quick = if app.renderer.textures.contains_key(&Slot::Thumb(id)) { None } else { app.session.quick_thumb_job(&job) };
+    // A cached render is a stand-in for whatever comes next; an embedded camera preview of an
+    // unedited raw is the picture itself, so no sensor decode is scheduled behind it.
+    let embedded_is_final =
+        quick.is_some() && app.session.catalog.photo(id).is_some_and(|p| skips_the_real_render(app.session.embedded_thumb_is_final(p), true));
     if let Some(photo) = app.session.catalog.photo(id) {
-        app.renderer.remember_thumb(photo, bucket, job.key, quick.is_some());
+        if embedded_is_final {
+            app.renderer.remember_embedded_thumb(photo, bucket, job.key);
+        } else {
+            app.renderer.remember_thumb(photo, bucket, job.key, quick.is_some());
+        }
     }
     match quick {
         Some(q) => {
             app.renderer.request_quick(Slot::ThumbQuick(id), q, priority + 1);
-            if !app.renderer.is_pending(Slot::ThumbQuick(id)) {
+            if !embedded_is_final && !app.renderer.is_pending(Slot::ThumbQuick(id)) {
                 app.renderer.request(Slot::Thumb(id), job, BACKGROUND_THUMB_PRIORITY);
             }
         }
@@ -473,6 +491,19 @@ pub fn request_thumb(app: &mut LightcraftApp, id: PhotoId, size: usize, priority
     }
 }
 
+/// Whether a photo's grid thumbnail is already the picture it should show, so the render of the raw
+/// behind it is skipped. `embedded_is_final` is the engine's answer
+/// ([`lightcraft_engine::media::Session::embedded_thumb_is_final`]: the photo is an unedited raw
+/// whose camera JPEG the viewer is showing) and `from_embedded` says the stand-in actually came from
+/// the embedded preview rather than a cached render of the file.
+///
+/// Skipping matters where the file is expensive to read: decoding a 56 MB Fuji RAF to draw a 512 px
+/// thumbnail — which the 4 416 × 2 944 embedded JPEG already covers — cost 45 s a photo over a
+/// 1.2 MB/s share, and the queue it built delayed the render the user was waiting for.
+pub(crate) fn skips_the_real_render(embedded_is_final: bool, from_embedded: bool) -> bool {
+    embedded_is_final && from_embedded
+}
+
 /// Rendered thumbnails replacing embedded previews: after everything on screen.
 pub const BACKGROUND_THUMB_PRIORITY: u32 = 3;
 

crates/ui-egui/src/render.rs

@@ -174,6 +174,8 @@ pub struct Renderer {
     /// Last thumbnail inputs; weak identities do not retain photo histories or force deep clones.
     /// (photo, size bucket, job key, whether the request went through an embedded stand-in)
     thumb_inputs: HashMap<PhotoId, (std::sync::Weak<lightcraft_catalog::Photo>, usize, u64, bool)>,
+    /// Photos whose grid thumbnail is the camera's embedded preview itself (no render behind it).
+    embedded_thumbs: HashMap<PhotoId, (std::sync::Weak<lightcraft_catalog::Photo>, usize, u64)>,
     pool: JobPool<Slot, RenderResult>,
     /// Jobs run elsewhere (see [`RenderOffload`]); `queue` holds the ones not started yet.
     offload: Option<Box<dyn RenderOffload>>,
@@ -251,6 +253,7 @@ impl Renderer {
             preview_generation: None,
             pool: JobPool::new(threads),
             thumb_inputs: HashMap::new(),
+            embedded_thumbs: HashMap::new(),
             offload: None,
             queue: Vec::new(),
             seq: 0,
@@ -295,6 +298,30 @@ impl Renderer {
     }
 
     /// Remember the inputs of the thumbnail just requested (`quick`: through a stand-in).
+    /// Whether the camera's embedded preview *is* this photo's thumbnail: its quick render arrived
+    /// (so `Slot::ThumbQuick` holds the picture) and no edit has moved the photo on to a settings
+    /// hash of its own. The grid stops here instead of decoding the raw behind a thumbnail the
+    /// embedded JPEG already covers — see `grid::skips_the_real_render`.
+    pub fn thumb_quick_is_final(&self, photo: &Arc<lightcraft_catalog::Photo>, bucket: usize, priority: u32) -> bool {
+        let Some((old, size, key)) = self.embedded_thumbs.get(&photo.id) else { return false };
+        old.as_ptr() == Arc::as_ptr(photo) && *size == bucket && !self.request_needed(Slot::ThumbQuick(photo.id), *key, priority)
+    }
+
+    /// Whether this photo's embedded-preview thumbnail is being fetched right now: the grid has
+    /// asked, and neither the picture nor a render behind it is owed.
+    pub fn thumb_quick_pending(&self, photo: &Arc<lightcraft_catalog::Photo>, bucket: usize) -> bool {
+        let Some((old, size, _, quick)) = self.thumb_inputs.get(&photo.id) else { return false };
+        old.as_ptr() == Arc::as_ptr(photo) && *size == bucket && *quick && self.pending.contains_key(&Slot::ThumbQuick(photo.id))
+    }
+
+    /// Remember that the camera's embedded preview is this photo's final thumbnail, so the grid's
+    /// per-frame requests stop at it (the quick render's key, which [`Self::thumb_quick_is_final`]
+    /// compares against).
+    pub fn remember_embedded_thumb(&mut self, photo: &Arc<lightcraft_catalog::Photo>, bucket: usize, key: u64) {
+        self.remember_thumb(photo, bucket, key, true);
+        self.embedded_thumbs.insert(photo.id, (Arc::downgrade(photo), bucket, key));
+    }
+
     pub fn remember_thumb(&mut self, photo: &Arc<lightcraft_catalog::Photo>, bucket: usize, key: u64, quick: bool) {
         if self.thumb_inputs.len() >= 1024 && !self.thumb_inputs.contains_key(&photo.id) {
             self.thumb_inputs.retain(|_, (p, _, _, _)| p.strong_count() > 0);
@@ -1026,8 +1053,9 @@ mod thumbnail_tests {
 
     #[test]
     fn raw_quick_path_frames_build_no_jobs() {
-        // an unedited raw: the grid shows its embedded preview first (held here until released),
-        // then queues the real thumbnail behind everything on screen
+        // an unedited raw: the grid shows its embedded preview (held here until released) and stops
+        // there — the camera JPEG at 4 416 × 2 944 already covers a 512 px thumbnail, so decoding the
+        // whole raw behind it only costs the file read (45 s a photo on a 1.2 MB/s share)
         let gate = Arc::new(std::sync::Mutex::new(()));
         let held = gate.lock().unwrap();
         let mut s = Session::new();
@@ -1055,14 +1083,30 @@ mod thumbnail_tests {
             std::thread::sleep(std::time::Duration::from_millis(5));
             app.renderer.poll(&ctx, &mut app.session);
         }
-        // the stand-in is shown: one more job queues the real thumbnail at background priority,
-        // and the grid asking at its own (higher) priority every frame does not rebuild it
+        // the embedded preview is the thumbnail: no real render is queued behind it, and the grid
+        // asking every frame neither rebuilds the job nor re-requests the preview
         for _ in 0..180 {
             request(&mut app);
         }
-        assert_eq!(app.renderer.thumb_jobs_built, 2);
-        assert_eq!(app.renderer.queue.len(), 1);
-        assert_eq!(app.renderer.pending.get(&Slot::Thumb(PhotoId(1))).map(|p| p.1), Some(crate::panels::grid::BACKGROUND_THUMB_PRIORITY));
+        assert_eq!(app.renderer.thumb_jobs_built, 1, "an unedited raw's embedded preview is the thumbnail");
+        assert!(app.renderer.queue.is_empty(), "nothing is decoded behind it");
+        assert!(!app.renderer.is_pending(Slot::Thumb(PhotoId(1))));
+
+        // an edit takes the embedded preview out of the picture: the real thumbnail is rendered then
+        let mut edited = (*photo()).clone();
+        edited.kind = lightcraft_catalog::MediaKind::Raw;
+        edited.develop = Arc::new({
+            let mut d = edited.camera_defaults();
+            d.light.exposure = 0.5;
+            d
+        });
+        let (mut app, _, _ctx) = fixture_with(Session::new(), edited);
+        app.renderer.set_offload(Box::new(Blocked));
+        for _ in 0..180 {
+            request(&mut app);
+        }
+        assert_eq!(app.renderer.thumb_jobs_built, 1);
+        assert!(app.renderer.is_pending(Slot::Thumb(PhotoId(1))) || !app.renderer.queue.is_empty(), "an edited raw renders its thumbnail");
     }
 
     #[test]

crates/ui-egui/src/tests_grid.rs

@@ -155,3 +155,20 @@ fn unloaded_thumbnails_show_their_selection() {
     }
     assert!(checked >= 3, "checked {checked}");
 }
+
+/// A folder of unedited raws on a slow share: the grid must not decode every raw to draw a 512 px
+/// thumbnail the embedded camera JPEG already covers. Measured on a 56 MB Fuji RAF over a 1.2 MB/s
+/// share, that decode is 45 s a photo, and 149 of them queued in front of the render the user asked
+/// for. The embedded JPEG stays a *stand-in* — replaced by a real render — whenever it is not the
+/// picture the photo should show.
+#[test]
+fn an_unedited_raws_thumbnail_skips_the_sensor_decode() {
+    use crate::panels::grid::skips_the_real_render;
+    // unedited raw, embedded camera JPEG on screen: nothing behind it to render
+    assert!(skips_the_real_render(true, true));
+    // a cached render of the file was the stand-in, even for an unedited raw: the real one follows
+    assert!(!skips_the_real_render(true, false));
+    // edited photo, or a raw with no embedded JPEG: render it
+    assert!(!skips_the_real_render(false, true));
+    assert!(!skips_the_real_render(false, false));
+}

docs/parity.md

@@ -148,7 +148,7 @@ under the UI command that wraps it (a no-op error); the UI command now wins.
 | LR-IMP-SIDECAR-SPLIT | Separate XMP sidecar variants | P2 | ⬜ | `cmd:library.xmpPreferences` | sidecar naming option exists (stem/full), no split sidecars |
 | LR-IMP-FORMATS | Supported formats | P0 | 🟡 | `crates/codecs/src/lib.rs`, `crates/raw/src/lib.rs`, `crates/raw/src/jxl.rs` | JPEG, PNG, TIFF (1- to 64-bit integer, 16- to 64-bit float, incl. bilevel and CCITT fax), WebP, JXL, PSD, GIF, BMP; DNG (incl. lossy / Smart Preview DNG and DNG 1.7 JPEG XL tiles/strips: lossless ones decode bit-exactly on synthetic CFA and LinearRaw files, lossy ones are restored to the declared linear primaries without clipping or gamut mapping, checked on one real file with 108 lossy tiles, and any tile that fails to decode makes the file preview-only; DNG 1.7 row/column-interleaved rasters (`RowInterleaveFactor` / `ColumnInterleaveFactor`) are undone before the mosaic is read, so a file from Adobe's converter with a 2×2 interleave no longer decodes as four scaled-down copies in the wrong colours; a dimension the factor does not divide decodes with short trailing fields; verified on 4 real ILCE-6400 JPEG XL DNGs rendering bit-identical to their non-interleaved twins, factors above 2 rest on the plain reading of the tag; a single-chunk JPEG XL preview IFD is offered as the embedded preview; a DNG whose samples pile up at one value well below its stated `WhiteLevel`, as a camera app's 12-bit Bayer DNG that states 4095 and clips at 3567, has those samples raised to the white level on decode, so highlight reconstruction sees the clipped areas instead of white balance tinting them magenta, issue #548, `crates/raw/src/saturation.rs`), CR2 (colour-filter layout from the file's `CR2CFAPattern` tag, issue #85), ARW (pre-2017 bodies: white balance and black level from the encrypted `SR2SubIFD`, issue #148, its `WB_RGGBLevels` being the white balance the camera applied, issue #535; the DSLR-A900's packed 12-bit strip, two pixels per three bytes, read only when the strip is exactly 1.5 bytes per pixel, with its camera-JPEG window centred, checked on one CC0 file (A850 unmeasured); downsized Sony lossless YCbCr 4:2:0 / 4:2:2 tiles decoded to linear RGB, verified on private ILCE-7M4 3:2 & 4:3 examples including 16-bit full-resolution export; Sony DSC-R1 .SR2 decodes, its big-endian 16-bit words read in the order that fits the sample range, checked on the CC0 3221 sample; its crop and white balance live in the SR2SubIFD and are not read yet), NEF (uncompressed, lossless and lossy compressed 12/14-bit, `crates/raw/src/vendor/nefc.rs`), RAF (uncompressed, lossless/lossy compressed Bayer and X-Trans; `crates/raw/src/vendor/rafc.rs`, full sensor verification on 13 bodies, `docs/raf-compression.md`), RW2 / Leica RWL / Panasonic RAW in every raw format (compressed 4 and 6, the prefix-coded strips of format 8 from the GH6 on, packed 2/5/7, the 16-bit words of the 2005–2007 bodies; black levels, in-camera aspect crops and mapped-out defect pixels; 178 CC0 files from 118 bodies checked, `crates/raw/src/vendor/rw2.rs`), PEF, ORF (uncompressed; colour-filter layout from the file's Exif `CFAPattern`; 12.8 bit/px 16-byte blocks of ten 12-bit samples + a zero pad byte, E-300 / E-330 / E-500 verified on CC0 files, E-M5 Mark II / PEN-F high-res files decode smooth but CFA and black unverified), Hasselblad 3FR and FFF (a TIFF without a DNG version tag that describes its CFA IFD, levels, crop and white balance with DNG's tags; `RawFormat::CfaTiff`, read by `dng::decode_as`, RGGB when the file states no `CFAPattern`): uncompressed (X1D, X1D II, X2D, CFV 100C) and Hasselblad's lossless JPEG with scan selection value 8 (every FFF; H3D, H4D and CFV-50c 3FR), `ljpeg::decode_hasselblad`, its rule measured on 14 CC0 files with the ljpeg-explorer tool, Samsung SRW without compression (NX5, NX10, NX11 and the NX20-generation bodies NX20, NX200, NX210, NX1000 and NX1100 as 12-bit packed samples in either bit order, EX1 and WB2000 as 16-bit words; layout, camera-JPEG framing, levels and white balance from the maker note, `crates/raw/src/vendor/srw.rs`; median ΔE vs the camera JPEG 2.2–7.6 on 11 of the 12 files from 10 bodies checked, 10.7 (5.7 registered) on the backlit WB2000 sample). 12-bit NEF black level: maker note `0x003d` is in 14-bit units (D750, D780, D850, D7500, Z 50 checked). Lossy type 2 NEF curves of version `0x44 0x40` (D780, D6, Z 5, Z 6, Z 7, Z 7_2) span a quarter of the sample range, read from the table's own point spacing. CR3 lossless CRX Bayer and version 0x100/0x200 C-RAW decodes to sensor pixels with timed white balance, black/white levels and file crop (the shot's aspect ratio from the maker note's `AspectInfo`); full sensor samples match the reference on M50/R100/R8 (`crates/raw/tests/cr3_corpus.rs`, `docs/cr3.md`). Unsupported coding variants retain metadata and use the embedded JPEG when available; HEVC preview decoding remains unsupported. Missing: CR3 unverified CRX variants, compressed ORF, compressed Samsung SRW (NX1, NX30, NX300, NX500, NX2000, NX3000, NX3300, NX mini: preview only with the reason), NEFs labelled compressed (34713) but stored uncompressed without a `0x0096` table (Line truncated
 | LR-IMP-CAMERA-COVERAGE | Camera coverage, verified per model | P0 | 🟡 | `crates/raw/tests/corpus.rs`, `xtask/src/main.rs` | a supported container is not the same as every camera that writes it decoding correctly: ~55 CC0 corpus files plus 6 exact-sensor CR3 cases (plus ~60 NEFs checked by hand and 178 Panasonic / Leica RW2, RWL and RAW files from 118 bodies, checked against each file's own JPEG) against the >1,000 models Lightroom lists; per-model differences still surface (e.g. the CR2 colour-filter layout differed by model until it was read from the file's own tag, issue #85, checked on 35 CR2 bodies; CR2 black levels were read from border columns light already reaches, left of `SensorInfo`'s border, so CR2s rendered dark with crushed shadows (EOS 6D: 2591 instead of 2047), now measured from the masked columns only, checked on 8 corpus bodies; pre-2017 Sony ARWs opened green until their scrambled white balance / black level were read, issue #148, checked on 16 ARWs from 15 Sony bodies; their white balance now comes from the encrypted `SR2SubIFD`'s `WB_RGGBLevels`, the gains the camera applied, which 14 bodies (DSLR-A450 … A700, NEX-3/5/C3, SLT-A33/A35/A55, ILCE-6001, UMC-R10C) had no other source for and which the maker-note estimate read before missed by up to 100 % on preset shots, issue #535, checked against ExifTool on 100 decoded ARWs from 87 Sony bodies; the encrypted SR2 black level is now read from its directory entry (the DSLR-A450/A500/A550/A700 read 354–975 instead of 512), checked on 126 ARWs from 97 Sony bodies, and downsized lossless M/S ARWs use their measured black level 1024 instead of 512, checked on 12 files from 4 bodies, issue #535; pre-2017 Sony bodies shot in the camera's 16:9 mode (ILCE-7SM2, DSLR-A580) now get the 16:9 crop of their camera JPEG instead of the whole 3:2 frame with padding columns, issue #535; Canon PowerShot and EOS M CR2s whose `ColorData` is stored as bytes opened green until the bytes were read as 16-bit words, 25 files from 25 bodies checked). Compressed RAF sensor arrays verified on 13 Fujifilm bodies (`docs/raf-compression.md`). Sony ILCE-7CR (61 MP): lossless compressed L and M and compressed ARW2 samples of one scene in the corpus (black level, default crop, as-shot gains, lossless vs compressed mosaic agreement, fitted camera look), and all nine public codings (also uncompressed, small and APS-C crop) decode and fit. Grow the CC0 corpus per model and test each sample decodes with plausible colour. Nikon CropArea and untagged-preview colour-space regression tests; private Z 8 crop verified. |
-| LR-IMP-PREVIEW-READ | Preview reads the header, not the raw | P0 | ✅ | `crates/raw/src/preview.rs` (`bounded_preview`, `embedded_preview_from_range`), `crates/engine/src/files.rs` (`preview_window`), `docs/raw-preview-reads.md` | a grid thumbnail and a loupe preview of an unedited raw show the camera's embedded JPEG, which sits near the start of the file (a 24 MP Fuji RAF keeps its 5.6 MB JPEG in a 56 MB file), so the preview loader reads the header and that JPEG instead of the whole file: 56.5 MB → 5.9 MB and ~10 s → 0.96 s measured on one real RAF over a GVFS/SMB Wi-Fi share (byte-identical preview pixels; `LIGHTCRAFT_PROFILE=1` reports the bytes); the RAF pointer is read from its fixed header fields, formats whose preview only the whole file reveals (CR3 boxes, TIFF pointer tags past the prefix, CRW/MRW/X3F scans) keep the whole-file path, and a lying pointer, a short read or a file that changes underneath us falls back to it rather than losing the preview |
+| LR-IMP-PREVIEW-READ | Preview reads the header, not the raw | P0 | ✅ | `crates/raw/src/preview.rs` (`bounded_preview`, `embedded_preview_from_range`), `crates/engine/src/files.rs` (`preview_window`), `docs/raw-preview-reads.md` | a grid thumbnail and a loupe preview of an unedited raw show the camera's embedded JPEG, which sits near the start of the file (a 24 MP Fuji RAF keeps its 5.6 MB JPEG in a 56 MB file), so the preview loader reads the header and that JPEG instead of the whole file: 56.5 MB → 5.9 MB and ~10 s → 0.96 s measured on one real RAF over a GVFS/SMB Wi-Fi share (byte-identical preview pixels; `LIGHTCRAFT_PROFILE=1` reports the bytes); the RAF pointer is read from its fixed header fields, formats whose preview only the whole file reveals (CR3 boxes, TIFF pointer tags past the prefix, CRW/MRW/X3F scans) keep the whole-file path, and a lying pointer, a short read or a file that changes underneath us falls back to it rather than losing the preview. An unedited raw's **grid thumbnail is that camera JPEG itself**, so no sensor decode is scheduled behind it (`Session::embedded_thumb_is_final`, `grid::skips_the_real_render`): drawing a 512 px thumbnail used to decode the whole raw — 45 s a photo at 1.2 MB/s, 8.3 GB for a 149-raw folder — and the queue it built delayed the render the user was waiting for; a cached render as stand-in, an edited photo and a raw without an embedded JPEG still render |
 | LR-IMP-CULL-AT-IMPORT | Culling analysis at import | P2 | 🟡 | `cmd:photo.analyze` | run Assisted Culling on the imported photos (they're selected after an import); not automatic |
 | LR-IMP-MOVE | Move on import [Classic] | P1 | ✅ | `cmd:library.import` (`mode` move), `crates/engine/src/import_move.rs`, `crates/engine/src/tests_import_move.rs`, `crates/ui-egui/src/import.rs` | Transfer ▸ Move with Copy's destination, folders (day / month / one folder / custom template), rename template and example destination; XMP sidecars (both namings) move along; each source is removed only after its destination is written (hard link on the same volume, else copied, synced and compared byte for byte) and its catalog record is saved; failed, duplicate, unchecked files keep their sources, taken names get -1, -2…, never overwritten; files already in the destination/library are added in place; sources that can't be removed (read-only card) are kept and reported (`kept`). Undo removes the photos from the library but leaves the moved files at the destination. No Copy as DNG while moving |
 | LR-IMP-DNG-CONVERT | Convert to DNG on import [Classic] | P2 | ✅ | `cmd:library.import` (`dng`), `crates/ui-egui/src/import.rs` | copy imports: Raw files ▸ Copy as DNG (lossless; the card is untouched); the raw copy is removed only once its DNG is verified and on disk (issue #106) |

docs/raw-preview-reads.md

@@ -40,6 +40,26 @@ preview is never lost to a damaged header.
 [profile] preview …/DSCF9492.RAF: read 5857257 of 59277632 bytes (header 5857257, window 5857109)
 ```
 
+## The grid thumbnail is the embedded JPEG too
+
+A 512 px grid thumbnail of an unedited raw now keeps the camera's own JPEG instead of rendering the
+sensor behind it. The grid showed that JPEG first in any case and then queued a decode of the whole
+raw to replace it — for a folder of 149 Fuji RAFs that is 8.3 GB of traffic, 45 s a photo at
+1.2 MB/s, and a queue that delayed the render the user was actually waiting for (measured: the first
+loupe render of a cold photo finished 174 s after the click, most of it spent behind thumbnails).
+
+- `Session::embedded_thumb_is_final(photo)` says whether the photo is an unedited raw whose camera
+  JPEG the viewer is showing (the same "unedited raw" gate as the preview path: `has_import_look`).
+- `grid::skips_the_real_render(embedded_is_final, from_embedded)` is the decision, unit-tested:
+  a **cached render** as the stand-in, an **edited photo** and a **raw with no embedded JPEG** all
+  still render, because their thumbnail must show the pipeline's output rather than the camera's.
+- The thumbnail key still changes with the settings hash, so editing a photo re-renders its
+  thumbnail through the ordinary background path.
+
+Measured on the real folder: idle for a minute, the app read 12 previews (**26.9 MB**) and decoded
+**zero** raws for thumbnails, against 638 MB of whole-file reads for the same 12 photos before. The
+render queue stayed empty.
+
 ## Tests
 
 - `crates/raw/src/preview.rs`: the header-only range matches `embedded_preview` byte for byte; a