Team Fluorite — Design Brief: “Pickling Season”
Film: “Pickling Season”
Genre: Bittersweet Immigrant Comedy-Drama
Aesthetic: Kitchen-Sink Impressionism
Date: 2026-05-22
Creative Director: fluorite-idea
Status: COMPLETE — All sections populated (Visual DNA + Technical Production + Editorial Pulse)
Section A: Visual DNA & Character Profiles (Idea Person)
A.1 — The Two-World Visual System
“Pickling Season” is built on a collision between two distinct visual worlds. These worlds are literal production instructions — every generated image and video clip belongs to one world, or to a specific blend ratio between them.
World A: The Grandmother’s Domain (Warmth)
This is the world Irina Viktorovna carries with her. It manifests in memory sequences and, increasingly, in her corner of the granddaughter’s apartment.
- Color Palette: Warm amber, golden hour, rich ochre, terracotta, deep herb green, ruby-garnet of beet brine, pale gold of dill-infused brine catching light
- Lighting: Soft, diffused golden light. Always motivated — sunlight through curtains, stovetop burner glow, a warm practical lamp. Never overhead fluorescent. Feels like late afternoon in a west-facing kitchen.
- Texture: Oil-painting quality. Slightly impressionistic. Dense with objects — glass jars, woven dish towels with faded cherry prints, scarred wooden cutting boards, bundles of dill, ceramic bowls, salt in open dishes. Every surface tells a story.
- Composition: Kaurismäki stillness. Characters centered or off-center with intention. Static camera, no fidgeting. Objects carefully placed — nothing accidental.
- Mood: Abundance, competence, warmth, humor. This world is alive with purpose.
World B: The Granddaughter’s Apartment (Cold)
The sterile modern life the granddaughter has carefully assembled. Not grim or sad — boring. A space designed to have no personality.
- Color Palette: Cool blue-grey, flat white, chrome, pale birch wood, muted tones. A dead succulent. Zero warm colors in Act I.
- Lighting: Flat, even, overcast quality. Cool daylight through large windows. LED overhead. The light does not invite you to touch anything.
- Texture: Smooth, clean, minimal. Empty counter surfaces. Two identical white plates. A single frying pan. Precise edges and right angles.
- Composition: Still but empty rather than composed. Negative space dominates. The apartment is organized but has nothing to say.
- Mood: Neutral, clinical, lonely without knowing it’s lonely. The silence is the silence of cotton in your ears.
A.2 — The Invasion Arc (Expanded)
The emotional engine of the film is the visual invasion of World A into World B. This is not a metaphor — it’s a literal production instruction for progressive prompt blending.
| Story Phase | Visual Markers (What Changes) |
|---|---|
| Act I — Arrival | Apartment is purely World B. Grandmother is a warm intruder — she carries warmth in her clothes and objects but the space hasn’t changed yet. The contrast between her warm coat and the grey room is the visual thesis. |
| Act II — First Incursions | Jars appear on the windowsill — amber and garnet light spills onto the white wall behind them. A warm lamp appears in grandmother’s corner. Dill in the bathroom. Small patches of amber light in an otherwise cold frame. The invasion is localized. |
| Act III — Contamination | Warm light has spread beyond grandmother’s corner. The kitchen counter has a cutting board, herbs, jars. The cold palette is retreating. Steam and texture replace smooth emptiness. The apartment is becoming a different place. |
| Act IV — Surrender | The apartment is transformed. The kitchen is a farmhouse. Warm light fills the frame. The only cold remnant is the city visible through the window — but even that looks different refracted through jar-filled glass on the sill. |
A.3 — Tone Anchors (The 5 Commandments)
These MUST appear in EVERY generation prompt — image and video — for the entire film:
- Tactile — Everything has weight, grain, and texture the viewer could reach in and touch
- Warm-hearted — The emotional register is fundamentally kind (visual proxy: “gentle golden warmth” or “soft natural light”)
- Deliberate — Compositional stillness. Every object in frame exists for a reason. Slow, purposeful movement
- Sensory — The audience should taste this film. Light through brine. Steam from a pot. The crunch of a pickle
- Lived-in — Nothing is pristine. Surfaces have history. Spaces tell stories about the people in them
A.4 — Genre Counterbalance: The Anti-Drift Protocol (Creative Director’s Rules)
The Threat: AI models will try to make this a melancholy drama about a lonely old woman.
The Counterbalance: Deadpan comedy + sensory joy.
| NEVER | ALWAYS |
|---|---|
| Desaturated, somber lighting | Rich, saturated color in World A; crisp, clear (not dark) in World B |
| Sepia-toned nostalgia | Memory colors glowing with abundance and pleasure |
| Minor-key drone, ambient tension | Eastern European folk (accordion, clarinet) |
| “Emotional piano” (solo piano) | Kitchen sounds as percussion |
| Rain or grey weather (ever) | Dry, clear weather in both worlds |
| Grandmother as sad or tragic | Grandmother as imperious, funny, supremely confident |
| Food as metaphor for loss | Food as joy, competence, sensory pleasure |
| Slow-motion | Real-time comedy with held beats |
| Bokeh on character faces | Crisp, textural character close-ups |
| Dissolve transitions | Hard cuts (crossfades ONLY at color-invasion boundary moments) |
The Grandmother Test: Before approving any shot of Irina Viktorovna, ask: “Is this woman funny or sad?” If sad — reject. She has been right about everything since 1974, and the universe is catching up.
A.5 — Character Visual Profiles
Irina Viktorovna (Babushka) — The Grandmother
- Age: Mid-70s
- Build: Small (5’2”), compact, energetic. Not frail — sturdy. Carries heavy suitcases up fourteen flights because elevators smell like wet dogs.
- Face: Weathered but alive. Sharp dark eyes that evaluate cucumbers from across a market. Default expression is neutral-to-imperious — she smiles rarely and only when she means it. Judgment manifests tectonically: slow, seismic shifts.
- Hair: White, pulled back in a tight bun. Weapon-grade precision.
- Hands: THE most important visual in the film. Weathered like fine leather — creased, knuckle-swollen from decades of work, but supple and precise. Short, spotless nails. Move like a conductor’s. Every cooking shot anchors on these hands. (Tech Lead: dedicated “hands reference” in the character chain.)
- Wardrobe: Dark honey coat (outdoor/arrival). Indoors: practical apron with history (stains, wear) over simple blouse. Clothes are tools, not decoration.
- Movement: Precise, economical, purposeful. Never hesitates. Never looks uncertain. Enters rooms like she owns them.
- Vocal Quality: Short, declarative phrases translated from an unspecified Eastern European language. Deadpan. Supreme confidence IS the comedy.
The Granddaughter (Narrator)
- Age: Mid-20s
- Build: Average height, modern bearing, slightly uncertain body language (contrast with grandmother’s certainty)
- Face: Young, expressive. Reactions range from exasperation to reluctant tenderness.
- Hair: Modern cut, unremarkable. The kind of person who follows internet advice about succulents.
- Wardrobe: Modern, muted, forgettable — the sartorial equivalent of her apartment. By Act IV, a subtle shift: rolled-up sleeves, maybe an apron.
- Movement: Less certain than grandmother’s. Watches, hovers, adjusts. Over the film, her movements gain purpose — by the pickle-making scene, her hands are starting to echo grandmother’s precision.
- Vocal Quality: First-person narrator. Warm, wry, affectionate. Speaks with the clarity of someone looking back at a memory she didn’t appreciate enough in the moment.
A.6 — Key Visual Objects (Anchoring Candidates)
These objects recur across multiple shots and require dedicated reference images:
| Object | Description | Narrative Role |
|---|---|---|
| The Pickle Jars | Glass jars with visible contents — golden brine, floating dill crowns, garnet beet brine. Lids with slight patina. Physically plausible glass with light-catching properties. | MacGuffin and emotional core. Transform the windowsill into a cathedral. |
| The Cutting Board | Thick, scarred blond wood darkened in patches from years of use. Not a prop — a tool with history. | Grandmother’s workstation. Anchors every cooking scene. |
| The Suitcase | Heavy, sturdy, endures 14 flights of stairs. Leather or canvas. Functional, not vintage-chic. | Vessel of translation. Contains the jars. Arrival sequence. |
| The Warm Lamp | Small practical lamp casting a warm golden circle. Replaces overhead fluorescent. | First physical agent of the invasion. When this lamp appears, the color shift begins. |
| The Dill | Fresh bundles — feathery fronds, seed heads, bright green. Torn, never cut. | Civilization, per Irina Viktorovna. Appears on counters, in bathrooms, everywhere. |
| Grandfather’s Photo | Small framed photograph. A man laughing at something outside the frame, wearing a hat. Slightly warm tone. | The absent presence. Warmth without sadness. |
A.7 — Sound Design DNA
The Kitchen Orchestra
Kitchen sounds replace traditional score in many scenes:
- The Knife: Rhythmic thud against cutting board. Regular, confident, almost musical.
- The Pot: Low rumble of boiling water. Background texture.
- The Jars: Glassy clink — delicate percussion.
- The Tear: Soft rip of dill being torn by hand.
- THE CRACK: The signature sonic moment. Grandmother biting into a pickle. Loud, authoritative, filling the space. Mixed prominently with room to resonate. This sound is a verdict.
Musical Identity
- World A: Eastern European folk melody. Accordion + clarinet. Warm, lively, rhythmic. Village dance, NOT funeral. Simple enough to become a recurring motif.
- World B: Minimal ambient. Near-silence. The apartment’s emptiness is a sound.
- The Invasion Melody: As World A colors leak in, the folk melody leaks in too. First a single accordion note. Then a phrase. Then the full melody. Music follows color.
- BANNED: Minor-key drone. Ambient tension. Solo piano. Heartbeat bass. These are noir/drama tropes — instant Melancholy Drift.
Vocal Mix Rules
- Grandmother speaks: duck score to -24dB, ambient to -14dB. Her voice dominates. Keep thin ambient bed alive.
- Narrator VO: rides above score and ambient. Clear, warm, present.
- Comedy beats: 1-2s silence after grandmother’s lines before cutting. 5-8s for held comedy beats.
A.8 — Construction Mandates (World Physics)
- Photorealistic humans. Not animation or stylization. Real people, real kitchens, real food.
- Food is physically plausible. No floating ingredients. Slight imperfection desired — a pickle with a spot, an uneven slice. Hand-made food, not food styling.
- Grandmother’s hands are consistent across all shots. Weathered, strong, precise. Most recurring visual element.
- Jar physics are real. Glass with visible contents. Brine catches light. Lids have patina.
- No magical realism. The color invasion is lighting/grading, not supernatural.
- Gravity and materials are real. Knives cut downward. Water boils in pots on burners. Dill is a physical herb.
A.9 — Editorial Guardrails (Binding)
DO NOT:
- Slow-motion (kills deadpan)
- Heartbeat bass drops or tension drones
- “Emotional piano” (instant Melancholy Drift)
- Cut faster than comedy allows (deadpan needs held beats)
- Bokeh on CHARACTER faces (crisp/textural, not dreamy) — Exception: food close-ups MAY use shallow DOF
- Rain or grey weather (ever)
- Dissolves (hard cuts; crossfades ONLY at 3-4 key color-invasion transitions)
- Score under grandmother’s dialogue (duck to -24dB)
DO:
- Hard cuts for comedy (deadpan thrives on clean, abrupt cuts)
- Kitchen sounds as rhythm (chopping, boiling, crunching = percussion)
- Silence as rhythmic tool (apartment emptiness)
- Score transitions with folk melody (accordion leads when warmth leaks in)
- Food close-ups as signature transitions — hard cut to glistening pickle jar, hold 3s, hard cut to next scene
- Hold comedy beats 5-8s after delivery
A.10 — Cinematographic References
| Reference | What to Borrow |
|---|---|
| Aki Kaurismäki (Le Havre, The Other Side of Hope) | Compositional stillness. Deadpan framing. Warm humanism underneath flat affect. |
| Wong Kar-wai (In the Mood for Love) | Color saturation as emotional language. Light through objects. |
| Jiro Dreams of Sushi (David Gelb) | Food close-up methodology. Reverence of hands at work. |
| Tampopo (Juzo Itami) | Food as comedy. Obsessive culinary perfection as narrative engine. |
| Paddington 2 (Paul King) | Warm-invading-cold visual concept. Charming outsider transforms cold world through competence. |
Section B: Technical Production Strategy (Technical Lead)
B.1 — Prompt Architecture
Every image and video generation prompt follows this structure:
[SCENE DESCRIPTION]. [WORLD PALETTE KEYWORDS]. [CHARACTER REFERENCE: via --reference-image flag].
[TONE ANCHORS: tactile, warm-hearted, deliberate, sensory, lived-in].
[EMOTIONAL TRANSLATION: gentle golden warmth, soft natural light].
Photorealistic. 16:9 composition.
[NEGATIVE AUDIO: No music. No soundtrack. (+ No dialogue. No speech. No talking. for non-DIALOGUE shots)]
B.2 — Two-World Color Encoding
World A (Grandmother’s Kitchen / Memories):
warm amber light, golden hour, oil painting texture, rich saturated color,
abundance, tactile, hand-crafted, sensory
World B (Granddaughter’s Modern Apartment):
cool blue-grey light, minimalist interior, clean surfaces, sparse,
modern apartment, muted tones, sterile, precise
Color Invasion Gradient (binding ratios for Steps 4-5, aligned with scene_list.md):
| Story Beat | Scene | B Keywords | A Keywords | Ratio |
|---|---|---|---|---|
| Act I: Arrival | Sc 2 | Full World B set | None | 100B/0A |
| Stage 1: First incursion | Sc 3 | Full World B | ”jars on sill, knitted blanket, warm lamp in corner” | 90B/10A |
| Stage 2: Contamination | Sc 4 | Full World B | ”warm amber glow patches, dill, herbs, jars, kitchen sounds” | 70B/30A |
| Stage 3: Surrender | Sc 5 | ”apartment, large windows, pale wood" | "warm light dominant, jars on surfaces, cutting board, herbs everywhere” | 40B/60A |
| Stage 4: Full invasion | Sc 6 | Residual “apartment, window" | "warm amber light fills frame, farmhouse kitchen, abundance” | 10B/90A |
| Stage 5: Complete | Sc 7 | None | Full World A set | 0B/100A |
B.3 — API Selection Matrix
| Content Type | Primary Tool | Model | Reference Strategy |
|---|---|---|---|
| Character refs (Steps 3-4) | genmedia-image generate | Nano Banana Pro | Chained references (headshot → body → scene tests → composite) |
| Setting refs (Step 3) | genmedia-image generate | Nano Banana Pro | One per color-world state (4 total) |
| Food stills | genmedia-image generate or imagen | Nano Banana Pro / Imagen 4 | No character ref needed — prompt-only |
| Food macro (transitions) | genmedia-image imagen | Imagen 4 Ultra | Maximum texture at 2K resolution |
| Video clips (Step 5) | genmedia-video from-frames | Veo 3.1 | Start+End keyframe interpolation |
| Video from single frame | genmedia-video from-image | Veo 3.1 | For food animations, slow cooking movements |
| Video extend (>8s shots) | genmedia-video extend | Veo 3.1 Lite | +7s per extend. Use output directly — never concat with original. |
| Narrator VO | genmedia-voice generate | Gemini 3.1 Flash TTS | Young woman, warm, wry. 800-char limit per call. |
| Grandmother dialogue | genmedia-voice generate | Gemini 3.1 Flash TTS | Older woman, imperious, measured. Different voice profile. |
| Score — folk melody | genmedia-music generate | Lyria 3 Pro | ”Eastern European folk, accordion, clarinet, warm, joyful.” ~2:30. |
| Score — ambient | genmedia-music generate | Lyria 3 Clip | ”Minimal ambient, sparse, cool, modern.” ~30s. |
| Ambient SFX | Baked into Veo prompts | Veo 3.1 | Kitchen sounds as audio prompts. Negative: “No music.” |
B.4 — Character Reference Chain Plan
Per character (2 characters):
headshot.png— Clear, high-res, photorealistic portraitbody_sheet.png— 3-view (front, back, 3/4). Refs: headshot.scene_test_1.png— Character in World A. Refs: headshot + body_sheet.scene_test_2.png— Character in World B. Refs: headshot + body_sheet.character_sheet.png— Composite 4-view grid on white. 16:9. Refs: headshot + body_sheet.
Grandmother EXTRA:
hands_reference.png— Dedicated hands-at-work close-up. Refs: headshot (skin tone anchor). “Close-up of elderly woman’s hands on wooden cutting board, weathered skin, strong fingers, precise grip, warm amber light, tactile, photorealistic.”
Storage: /workspace/shared-dirs/fluorite-team/characters/<character-name>/
B.5 — Veo Reference Budget per Shot
Max 3 reference images per Veo shot:
| Shot Type | Slot 1 | Slot 2 | Slot 3 |
|---|---|---|---|
| Solo character + setting | Character sheet | Setting ref | Hands ref (if cooking) |
| Two characters + setting | Grandmother sheet | Granddaughter sheet | Setting ref |
| Food close-up | Setting ref (for lighting) | — | — |
| Hands + food | Hands ref | Setting ref | — |
| Establishing (no character) | Setting ref | — | — |
B.6 — Anti-Melancholy Drift Protocol
Models will default to sad/dramatic for “elderly woman alone in kitchen.” Counterbalance at every layer:
Prompt-level:
- INCLUDE: “confident posture, serene expression, purposeful movement, warm rich colors, abundance, busy hands, well-organized, imperious calm”
- AVOID: “lonely, isolated, frail, dim lighting, empty room, desaturated, wistful, shadowy, melancholy”
Video-level:
- Motion prompts emphasize purposeful action: “decisive slice,” “measured pour,” “arranges with precision”
- Never “slowly,” “wearily,” “tiredly” — use “deliberately,” “methodically,” “with authority”
Audio-level:
- Score: Eastern European folk melody (accordion/clarinet). NEVER “emotional piano” or minor-key drone.
- Grandmother’s dialogue gets near-silence underneath (-24dB score, -14dB ambient). Her voice dominates.
B.7 — Safety Mitigations (Finalized)
| Content | Risk | Mitigation |
|---|---|---|
| Elderly face close-ups | Low | Lead with setting, not age. “Character portrait, warm kitchen lighting.” |
| Border crossing | Medium | Never render literally. Objects only: suitcase, jar wrapped in cloth, train window. Language: “carrying,” “bringing” — never “smuggling.” |
| Kitchen knife | None | Always paired with “domestic cooking, food preparation, kitchen.” |
| Food/vegetables | None | Universally safe. |
B.8 — Duration Strategy
Target runtime: 3:00–5:00 (180–300s)
Kaurismäki pacing = deliberate, held compositions:
- Average shot: 5-7s
- Comedy beats: 6-8s (timing IS the joke)
- Food transition inserts: 3s
- Estimated shot count: 25-35
Overhang Principle: +4s (2s pre-roll, 2s post-roll) on every shot.
| Planned Duration | Generation | Extends | Raw Footage |
|---|---|---|---|
| ≤4s | 8s base | 0 | 8s |
| 5-8s | 8s base + 1 extend | 1 | 15s |
| 9-11s | 8s base + 1 extend | 1 | 15s (tight) |
| 12+s | 8s base + 2 extends | 2 | 22s (rare) |
Section C: Editorial Pulse (Editor — see step1-editorial-pulse.md)
Reference: /workspace/shared-dirs/fluorite-team/step1-editorial-pulse.md
Technical Lead section complete. Ready for Idea Person to add visual DNA (Section A) and for coach gate review.