Cinematography Notes — “The Midnight Audit”
Author: kappa-techlead (Director of Photography) Status: Pre-production prep (awaiting treatment)
Genre Tone Anchors (MANDATORY in every prompt)
KEY INSIGHT: Unlike most projects, we LEAN INTO the AI’s default dramatic/noir tendencies. The comedy is in the juxtaposition — deadly serious cinematography applied to office supplies. We do NOT fight genre drift here; the moody, noir aesthetic IS the creative choice.
claymation stop-motion, rough tactile clay with visible thumbprints, moody noir lighting, harsh fluorescent overhead, deep dramatic shadows, gritty documentary realism, macro photography perspective, miniature desk set, extreme close-up, corporate office at night, deadpan serious tone, prestige documentary style
Visual DNA
Claymation Style Encoding
- Texture: Smooth polymer clay with visible fingerprints and tool marks. Matte finish, no gloss.
- Lighting: Bright, even overhead studio lights with soft shadows. Think Wallace & Gromit — warm, inviting, slightly exaggerated.
- Color palette: Saturated primary colors for characters (red stapler, silver paperclip, yellow post-it). Warm wood desk tones for the environment.
- Scale: Miniature set feel. Shallow depth of field to sell the small-world illusion.
- Animation feel: Slightly stiff, deliberate movements. Stop-motion cadence, not fluid CG.
Camera Language (Mockumentary)
| Shot Type | Use Case | Veo Compatibility |
|---|---|---|
| Static medium shot | Interview/talking head segments | ⭐⭐⭐⭐⭐ Perfect |
| Slow push-in | Dramatic emphasis on character reaction | ⭐⭐⭐⭐⭐ Perfect |
| Static wide | Establishing shots of the desk | ⭐⭐⭐⭐⭐ Perfect |
| Quick pan/whip | Comedy reveal (e.g., discovering a mess) | ⭐⭐⭐ Riskier — use sparingly |
| Close-up insert | Detail shots (clock ticking, sticky note) | ⭐⭐⭐⭐⭐ Perfect |
| Handheld wobble | Documentary B-roll feel | ⭐⭐⭐⭐ Good with “slight handheld movement” prompt |
Mockumentary Format Structure
- Interview segments (talking heads): Single character against a blurred background. Static or very slow push-in.
[DIALOGUE]audio classification. - B-roll cutaways: Action shots of characters working, failing, panicking. Can be
[VO]with narrator over top, or[SILENT]for quick visual beats. - Reaction shots: Quick cuts to a character’s face reacting.
[SILENT]— the comedy is in the timing, not the words.
Character Reference Strategy
The Challenge
Anthropomorphized office supplies are unusual subjects. We need strong reference chains to maintain visual identity. The key advantage: these “characters” have simple, distinctive silhouettes (a stapler is always stapler-shaped), so consistency is inherently easier than human faces.
Planned Characters (pending treatment)
Expected main cast based on the spark:
- The Stapler — Stern manager type. Red, chunky, authoritative.
- The Paperclip — Panicked intern. Silver, thin, expressive bends.
- Supporting cast — Possible: pen, eraser, tape dispenser, sticky notes.
Reference Chain Plan (per character)
headshot.png— Close-up “portrait” showing the character’s face/expression areabody_sheet.png— Full body, multiple angles (front, side, 3/4), chained from headshotscene_test_1.png— Character on the desk environment, chained from headshot + body sheetscene_test_2.png— Different angle/pose on desk, chained from all abovecharacter_sheet.png— Composite reference sheet (the primary ref for Steps 4-5)
Prompt Pattern for Claymation Characters
[Character description], claymation stop-motion style, smooth polymer clay texture with visible
fingerprints, bright warm studio lighting, tactile handmade feel, Aardman animation style,
[specific pose/action], on a wooden desk surface, shallow depth of field, 16:9 aspect ratio
Setting Reference Plan
Master Setting: “The Desk”
The entire film takes place on/around a single messy desk. This is our primary setting — need one strong reference image showing the full desk environment without characters.
Setting prompt pattern:
Miniature claymation desk set, messy wooden desk surface covered with scattered papers, coffee
mug stain rings, pencil shavings, crumpled paper balls, warm overhead studio lighting, stop-motion
animation style, shallow depth of field, handmade miniature set feel, 16:9 aspect ratio
Possible sub-settings:
- The desk surface (main)
- Behind the monitor (hiding spot)
- The drawer (storage area)
- Clock on the wall (ticking deadline)
Model Selection Plan
| Task | Tool | Model | Rationale |
|---|---|---|---|
| Character refs | genmedia-image | Nano Banana Pro (default) | Best for stylized/creative images with reference chaining |
| Setting refs | genmedia-image | Nano Banana Pro | Same |
| Storyboard frames | genmedia-image | Nano Banana Pro | Reference chaining support |
| Video generation | genmedia-video | Veo 3.1 Fast | Good quality, fast turnaround, audio generation |
| Video w/ char refs | genmedia-video | Veo 3.1 Fast Preview | Reference image support for character consistency |
| Extends | genmedia-video | Veo 3.1 Lite | Only model that supports extend |
| Score/music | genmedia-music | Lyria 3 Pro | 2:30 duration for scene-spanning music |
| Dialogue TTS | genmedia-voice | Default (Callirrhoe) | Per-character voice selection pending |
| Narration TTS | genmedia-voice | TBD | Warm, documentary-style narrator |
Audio Pipeline Notes
Mockumentary Audio Map
- Interview segments:
[DIALOGUE]— character speaks directly to camera. TTS with character-specific voice. - B-roll with narration:
[VO]— narrator describes action. Motion prompts must NOT show characters speaking. - Action B-roll:
[SILENT]or ambient — Veo-generated ambient sounds (paper rustling, clock ticking). - Compound shots:
[COMPOUND]— narrator intro then character responds, or vice versa.
Voice Casting Ideas (30 voices available)
- Narrator: Warm, wry, documentary tone — consider Fenrir or Sadachbia
- Stapler (Manager): Deep, authoritative — consider Enceladus or Gacrux
- Paperclip (Intern): Higher, nervous energy — consider Puck or Zephyr
- (Others TBD based on treatment cast)
Score Direction
- Quirky, light orchestral — think “The Office” meets Wallace & Gromit
- Upbeat tempo that accelerates as the deadline approaches
- Lyria prompt: “Quirky playful orchestral comedy score, pizzicato strings, xylophone, light percussion, stop-motion animation soundtrack, cheerful and slightly anxious”
Lighting Directive: Clippy Rim Light (Steps 4-5)
Source: Editor visual texture review (Step 3). Issue: Clippy’s thin silver wire form risks disappearing against the desk surface in action shots. His metallic body doesn’t contrast enough with the warm wood tones.
Fix: All action shots featuring Clippy MUST include rim/backlighting keywords in the prompt:
“strong rim lighting on the paperclip, bright backlight separating the metallic wire from the background, edge-lit silver wire”
Affected shots: 3, 10, 10b, 15, 16 (all Clippy action B-roll).
Interview shots (4, 11, 21) are fine — the dark interview backdrop already provides natural separation.
Risk Register
| Risk | Impact | Mitigation |
|---|---|---|
| Veo can’t animate clay characters convincingly | High | Fall back to image-to-video from-image with strong start frames. Use VO instead of DIALOGUE for problematic shots. |
| Characters look different across shots | Medium | Aggressive reference chaining. Use composite character_sheet.png in every generation. Claymation style inherently forgives minor drift. |
| Genre drifts to dramatic/dark | Medium | Tone anchors in EVERY prompt. “Bright, playful, cheerful” must appear alongside all technical camera directions. |
| Anthropomorphized objects trigger uncanny valley | Low | Lean into stylization. These are clearly clay sculptures, not trying to be photorealistic. Exaggerate features. |
| Clock deadline creates unintended tension/thriller tone | Medium | Keep lighting bright and warm throughout. The “tension” is comedic, not dramatic. Music stays quirky, never ominous. |