EXECUTION — Timeline Engine

Block Header: [EXECUTION]
Parent Pillar: EXECUTION (The Kinetic Flow)
Production Phase: Phase 2: EXECUTION (Kinetic Production / Rolling Camera)
Placement: Block 5 (The Temporal State-Transition Engine)

1. Purpose & Functional Role

[EXECUTION] delivers the continuity-aware timeline of visual and auditory moments. It translates the static directorial staging established in Phase 1 into an unbroken sequence of Macro-States (\(S_1 \mapsto S_2 \dots \mapsto S_n\)).

Under the core philosophy of “Boundaries, Not Handcuffs,” [EXECUTION] defines the structural boundary conditions for each beat, allowing the human director to mentally track pacing while giving the visual reasoning layer the creative freedom to render fluid motion, physical weight, and natural performance arcs between states without unprompted drift or hallucination.

2. Formatting & Line Architecture

  • Header: [EXECUTION] on its own line.
  • Line Structure: Exactly one Macro-State per line.
  • Human Pre-Visualization Unit: Line breaks represent cognitive reading units for human mental simulation. Breaking actions across lines prevents dense walls of text and allows creators to mentally simulate the pacing of the scene line-by-line before generation.
  • Chaining Operator (->): Micro-events occurring chronologically within a single Macro-State are chained from left to right using ->.
  • Tag Discipline:
    • Sub-States begin with a single-dimension coordinate tag (e.g., [CAM], [BLOCK], [ACT], [AUDIO]).
    • Tags are labels only; descriptive text follows outside the bracket with a single space.
    • Square-bracket tags are mandatory by default and may only be omitted if the user explicitly commands "tagless" or "strip tags".

3. The First Macro-State Anchor Rule (\(S_0 \to S_1\))

The very first line of [EXECUTION] (\(S_1\)) is strictly bound to the ground-truth canvas established in [OPENING] (\(S_0\)):

\[\text{[OPENING]} \equiv S_0 \implies S_1 = f(S_0 \mid \text{STAGING})\]

[!IMPORTANT] The Zero-Jump Law: 1. Initial Tag Declaration: Because [OPENING] is written in tagless prose, Line 1 of [EXECUTION] must open with a bracketed [CAM] tag (or named anchor [CAM 01]) declaring the exact shot size, lens angle, and perspective established in [OPENING]. 2. Zero Spatial or Optical Jump: Line 1 cannot jump-cut to a new angle, change camera distance, or swap character frame coordinates relative to \(S_0\). 3. Prop Grip Continuity: Any held object established in \(S_0\) (e.g., “holding key in right hand”) must serve as the physical origin point for kinetic actions on Line 1. 4. Posture & Eyeline Lock: Initial body postures (seated, standing, slouched) and gaze directions on Line 1 must inherit directly from \(S_0\).

4. Timeline Mechanics & Forward State Flow

\[\begin{aligned} S_0 \ (\text{[OPENING]}) &\implies S_1 \ (\text{Execution Line 1}) \\ S_1 &\implies S_2 \ (\text{Execution Line 2}) \\ &\ \ \vdots \\ S_{n-1} &\implies S_n \ (\text{Execution Line } n) \end{aligned}\]

  1. Sequential State Inheritance: Each subsequent Macro-State line inherits the terminal visible condition, held objects, and camera position of the preceding line unless an explicit transformation is commanded.
  2. Same-Camera Multi-Line Pacing: Multiple consecutive Macro-State lines frequently share the same camera setup (e.g., repeating [CAM 01] across 2–4 lines) to pace out micro-actions, subtle eye shifts, or dialogue exchanges cleanly.
  3. Camera Transitions & Cuts: Declare a new camera coordinate (e.g., [CAM 02]) only when changing lens framing, switching coverage, or executing a motivated cut.

5. Detail Inference Levels

Select the appropriate line density based on scene complexity:

  • Default Format (Core Triad): For standard single-subject or simple kinetic beats:
    [CAM] ... -> [ACT] ... -> [AUDIO] ...
  • Hybrid Format (Spatial & Axis Anchoring): For multi-character dialogue and strict screen direction tracking:
    [CAM] ... [AXIS] ... -> [BLOCK] ... -> [ACT] ... -> [AUDIO] ...
  • Expanded Format (State-Locked): For critical prop transfers, continuous physical transformations, or high-stakes continuity pivots:
    [CAM] ... [AXIS] ... [STATE IN] ... -> [BLOCK] ... -> [ACT] ... -> [STATE OUT] ... -> [AUDIO] ...

6. Dedicated Execution Subsystem & Companion Specifications

[EXECUTION] serves as the Macro-Block container for Phase 2. The internal craft, visual blocking, taxonomy, and atomic coordinate rules are modularly specified across dedicated companion documents:

Directorial Dimension Dedicated Specification Document Directorial Scope
Macro-Block Rules & Grammar block_rules.md 3 Core Pillars, Staging vs. Execution phases, and the Block Extension Principle.
Directorial Theory & Staging execution_director.md Screen geography (Frame-L/C/R, FG/MG/BG), 180° axis, Literal Over Literary performance translation, and the Repetition Mandate for continuity.
Tag Taxonomy & Hierarchy tag_taxonomy.md 3-tier tag hierarchy overview, Macro-Block vs. Sub-State disambiguation, and tag routing.
Atomic Sub-State Grammar substate_grammar.md Single-dimension law, chaining syntax, and atomic tag standards.
Individual Tag Specifications grammar/substates/ 1-to-1 dedicated coordinate specs (tag_cam.md, tag_act.md, tag_dialogue.md, tag_audio.md, tag_block.md, tag_prop.md, etc.).

7. Prohibited Anti-Patterns & What to Avoid

  • ❌ Unanchored First Line: Never open Line 1 with a camera angle, character position, or prop allocation that contradicts [OPENING].
  • ❌ Meta-Prompt Bleed: Never include instructions about video length, script formatting, or pacing advice inside execution lines (e.g., avoid “Shot 1 lasts 4 seconds” or “Make this scene dramatic”).
  • ❌ Dense Wall-of-Text Paragraphs: Never combine multiple distinct camera setups or unrelated action beats into a single unformatted block of text.
  • ❌ Prop Ghosting & Hand Amnesia: Never assume the model remembers which hand holds an object across cuts; explicitly re-state critical held props across shots until they are set down.
  • ❌ Abstract Literary Feelings: Avoid directing non-observable emotions (e.g., “Samira feels betrayed”); direct observable physical anatomy, micro-expressions, posture, and breath rate.

8. System Mechanism & Upstream Relations

\[S_n = f(S_{n-1} \mid \text{STAGING})\]

\[\text{STAGING} = \langle \text{INTENT}, \text{LOGIC}, \text{AESTHETIC}, S_0 \rangle\]

  • State System Role: Acts as the Dynamic State-Transition Trajectory (\(S_1 \mapsto \dots \mapsto S_n\)), driving the temporal evolution of the scene frame-by-frame.
  • Upstream Input:
    • From [INTENT]: Inherits dramatic stakes, tone, and canonical character identities.
    • From [LOGIC]: Bounded by invariant spatial rules (180° axis, prop permanence, physical degradation).
    • From [AESTHETIC]: Grounded in the static lookbook (palette, baseline lighting, room tone, wardrobe tokens).
    • From [OPENING]: Initialized directly from the frozen ground-truth frame \(S_0\).
  • Output Role: Produces the final, continuity-locked instruction sequence verified by the human director and executed by the visual reasoning engine.
Back to top