# Fallow kitchen reference ledger

Source: [POV: Head Chef at a Top London Restaurant — Fallow](https://www.youtube.com/watch?v=Ipe9xJCfuTM), 29:09. Supplemental source: [What's in a Professional Kitchen? — Fallow](https://www.youtube.com/watch?v=y3gVno6UX4Q), 9:58. The service kitchen in the supplemental tour visibly matches the source footage. Its basement preparation rooms are outside this reconstruction.

The geometry is newly authored. Most texture maps are generated by the checked-in scripts. **One stone base-color map is derived from a small perspective-corrected crop of the unoccupied countertop at POV 04:19.** The checked-in `source/stone-reference-patch.png` records that crop; `extract_stone_reference.py` records the exact crop polygon and processing. It contains no people or printed text. The full source footage and reference frames remain outside the repository. The featured walkthrough includes an eight-second, muted excerpt at 08:06–08:14, with a visible Fallow/YouTube source link; clip metadata is in `review/featured/media.json`. There are no photographic scene backplates. This is a reference-based reconstruction, not a measured digital twin or an affiliation with Fallow.

## Observations used

| Source / time | Visible evidence | Model decisions |
| --- | --- | --- |
| POV 00:30, 04:19, 07:57 | Pale ivory/green stone, dense dark veining, slim ticket rail, narrow printed orders, white broad-rim plates and oatmeal ceramic dishes | Rectified stone surface crop with generated roughness/normal maps; individual fictional ticket planes, dishes and stainless slide rail |
| POV 01:25, 01:51 | Four small saucepans in a large rectangular heated pan; black ribbed trivet; utensil crock and round dark service trays | Open vessel geometry, handles and rolled rims; separate trivet ribs and tray edges |
| POV 03:58 | Very low pass sightline, heat lamps underneath shelf; extraction baffles on opposite wall | Separate timber shelf, stainless underside and visible strip lights; full depth kitchen |
| POV 04:55 | Central bank with compressor grille, blue display, three drawer columns; another separate prep block; broad rear cooking line | Distinct islands, ventilated refrigeration housing, individual fronts and handles |
| POV 08:19, 10:00 | Narrow working aisle, plate/bowl stacks, side tray at prep block and undercounter plate storage | Retain real-looking aisle scale and accessible plate shelves |
| POV 08:55 | GN pans with herbs, butter, sauces; yellow boards; mixing bowls; differently colored squeeze bottles | Individual containers, rigid movable bottles/utensils/bowls and visual ingredient dressing |
| POV 12:34, 23:48 | Figured dark timber fascia, black circular mounting plates and brass fasteners, overhead plate stacks | Characteristic suspended pass assembly with supports, end hardware and storage |
| POV 13:54, 14:24, 22:21 | Rounded glazed burger buns and plated service food | Separate modeled bun halves and patty, textured crust/crumb; exact filling and recipe proportions are inferred |
| POV 15:54 | Wall-mounted salamander with two knobs, red timer, glowing coils and grill bars; fire pull and conduit; induction top | Detailed cooker, coils, controls and suppression hardware; optional generic thermal/controller model, inferred parameters |
| POV 17:53 | Small end station with countertop oven, lower storage and short stone return | Modeled end preparation station |
| POV 26:40 | Charcoal floor, central prep-block end, small inset utensil holder | Dark fine aggregate floor, end tray, crocks |
| Tour 05:55 | Twin fryer basins with wire baskets and long black handles | Open basins, basket wires, handles, static oil appearance |
| Tour 06:03, 06:31 | Patinated wall panels, steel shelf/brackets, gas burner and solid top, stockpots, racked trays | Rear equipment hierarchy, baffles, vent slots, shelf supports, pan handles |
| Tour 07:23 | Broad perforated extraction fascia, suppression pipes/nozzles, pilot switches | Individual grille slats, pipes and wall controls |
| Tour 07:45 | Open dining room beyond pass, black framing and large windows | Shallow restaurant context for camera depth and lighting |
| POV 01:12, 12:34, 18:24, 24:50 | Concentric rectangular black trivet channels, pot lids and ladles, cylindrical chip cups and folded white linen in a plastic tub | Replaced parallel trivet ribs and metallic linen tub; added fitted lids, service utensils, modeled folds and individual chips |
| POV 04:54, 10:15 | Rolled GN rims, nested deep tubs, worn yellow boards, long knife, utensil cylinder with red spatula | Added flange geometry, varying container depth, scored board maps and separate spatulas; retained open physics targets |
| POV 05:21 | Salamander timer with four digits, three small buttons and metal controls; dense pans and caddies | Added buttons and fasteners; no manufacturer or electrical circuit identified |
| POV 08:03, 10:15 | Blue-lidded translucent shaker with top loop; inset utensil storage | Replaced the plain blue bottle with a shaped body, separate lid, spout, loop and volume marks |
| POV 13:56 | Dense mise en place, multiple sauces and utensils, plate storage below pass | Additional container lids and nesting, variable props and wear; ingredients still partly inferred |

## Scale and uncertainty

No survey, calibration target or floor plan is supplied. Approximate scale comes from typical 26–29 cm dinner plates, standard GN pan footprints and approximately 90–95 cm worktops. Main service envelope is estimated at approximately 7.4 × 5.0 m; kitchen ceiling approximately 2.8 m. Pan sizes are plausible standard footprints, not measured product identification.

Equipment arrangement, surface character and principal sightlines have strong multi-view support. Exact floor dimensions, wall thickness, cabinet internals, the left doorway, restaurant spacing and exterior facade are inferred. Dining chairs/tables are simplified context. Food is represented by detailed rigid or static geometry. Ingredient composition, food density, dry friction, appliance power and thermal conductances are estimates. The optional stockpot thermal model and separate GPU liquid experiment are described in the README and have their own checks; the default scene does not simulate cooking, combustion, boiling or deformable food. No claim of robotic task success follows from scene authoring.

## Visual iterations

1. Blocked out the pass, islands, extraction and principal cooking stations; rendered three views.
2. Corrected the facing of the side cooker/oven, reduced lighting and refined steel and open vessels.
3. Added missing shelf support, corrected floor support and window transmission; replaced the inaccurate procedural stone pattern with the documented crop.
4. Added individual herb leaves and adjusted cutaway visibility and small-object placement.
5. Corrected worktop heights and overlapping props; reviewed seven higher-resolution views.
6–8. Added separate burger components, lemon interior, mushroom detail, portable food textures and gas jets; removed an overlapping plated dish and improved the food silhouette. Made the stockpot movable and added estimated material-specific friction.
9. Rendered the final review set from ten cameras at 1,800 × 1,200, 128 samples. Native physics, optional thermal and GPU particle receipts are stored separately.
10–13. Inspected 24 additional frames spanning 00:45–28:04, plus close-ups from the previous study. Refined small objects, darkened the stone map toward the observed contrast, added scored/used surface maps, and lit the exterior context. Corrected intersecting linen/crock, lids/ladles, crowded trays and chips protruding through their cups after inspecting new render angles. Rebuilt the oven with a hollow cavity, fan guard, rack, gasket and working hinged door; these internals are inferred. Expanded the final review to fourteen views and added explicit appliance/bulk-water models and a kinematic jug experiment.

The second texture pass applies contrast and a green color correction to the documented stone crop as an artistic albedo estimate; it is not a recovery of measured reflectance. Knife scratches and steel discoloration are generated, without claiming an exact match to individual marks in the footage. The pouring jug is a task-enabling interpretation of the visible small metal service vessels. Appliance powers, timers/interlocks, heat capacities, cooling behavior and hidden internals cannot be established from the footage and are explicitly modeled assumptions.

The saved iteration images are evidence of inspection, not an assertion of photoreal perfection. Remaining visual limitations include the resolution/lighting baked into the stone reference crop, simplified food microstructure, repeated props and inferred restaurant context.