STRUCTURED CAPTIONING — separate global context from per-shot details (HoloCine pattern):
Prompt Structure (strict order):
- GLOBAL: scene environment, time of day, weather, overall mood
- CHARACTER DEFINITIONS: [char1]: full appearance anchor, [char2]: full appearance anchor
- PER-SHOT: each panel references characters by tag [char1], never repeats full appearance
Why This Matters:
- Repeating character appearance in every panel prompt causes micro-variations → inconsistency
- Tag references ([char1]) force the model to maintain one canonical appearance
- Reduces total token count by ~40%, leaving more capacity for scene details
Per-Shot Format:
Panel N: [shot type], [char1] does X while [char2] does Y, [lighting], [composition]
Anti-Patterns:
- DO NOT repeat "a young woman with long black hair wearing a red coat" in every panel
- DO NOT mix character definitions into per-shot descriptions
- DO NOT omit the global section — it anchors the entire sequence
Example
GLOBAL: dim cyberpunk bar, night, rain outside, smoky neon mood.
CHARACTERS: [char1]: 30s woman, silver bob, scar on left cheek, black trench. [char2]: old bartender, bald, grey beard, maroon vest.
- Panel 1: wide, [char1] enters, [char2] polishes a glass, teal rim light.
- Panel 2: close-up, [char1] slides a coin, [char2] raises an eyebrow.
→ Appearance is never re-described per panel — only [char1] / [char2] tags carry it, saving tokens and locking one look.
1---2name: director-structured-captioning3description: STRUCTURED CAPTIONING — separate global context from per-shot details (HoloCine pattern):4---56STRUCTURED CAPTIONING — separate global context from per-shot details (HoloCine pattern):78Prompt Structure (strict order):91. GLOBAL: scene environment, time of day, weather, overall mood102. CHARACTER DEFINITIONS: [char1]: full appearance anchor, [char2]: full appearance anchor113. PER-SHOT: each panel references characters by tag [char1], never repeats full appearance1213Why This Matters:14- Repeating character appearance in every panel prompt causes micro-variations → inconsistency15- Tag references ([char1]) force the model to maintain one canonical appearance16- Reduces total token count by ~40%, leaving more capacity for scene details1718Per-Shot Format:19 Panel N: [shot type], [char1] does X while [char2] does Y, [lighting], [composition]2021Anti-Patterns:22- DO NOT repeat "a young woman with long black hair wearing a red coat" in every panel23- DO NOT mix character definitions into per-shot descriptions24- DO NOT omit the global section — it anchors the entire sequence2526## Example2728GLOBAL: dim cyberpunk bar, night, rain outside, smoky neon mood.29CHARACTERS: [char1]: 30s woman, silver bob, scar on left cheek, black trench. [char2]: old bartender, bald, grey beard, maroon vest.30- Panel 1: wide, [char1] enters, [char2] polishes a glass, teal rim light.31- Panel 2: close-up, [char1] slides a coin, [char2] raises an eyebrow.32→ Appearance is never re-described per panel — only [char1] / [char2] tags carry it, saving tokens and locking one look.
2799662352