✅ Glossary
A quick guide to framework terminology.
This glossary lists the core terminology and notation used throughout Auteur Script. Each card provides a concise definition and its practical usage, linking directly to the guides where it is explained in depth.
NoteInteractive Reading Across the Documentation
Every term defined in this glossary is active across all guides and specifications on this website. When you encounter an underlined term in any article, hover or tap it to view its definition in the margin without losing your reading place.
1. Staging Architecture
Foundational pillars, pre-production setup, and the structural macro-block containers that define world invariants and initial framing.
The pre-production lookbook container defining the sensory world of the scene.
Locks color palette, lighting ratios, wardrobe, location dressing, and acoustic environment before recording begins. Contains no chronological time progression.
The dynamic timeline container directing sequential beats across rolling camera screen time.
Sequences camera moves, physical performances, spoken lines, and diegetic sound across time (\(S_1 \mapsto \dots \mapsto S_n\)) without repeating static staging lookbook details.
The pre-production container answering ‘Why does this shot exist?’
Defines the core dramatic objective, format constraints, and emotional subtext. Never describes technical camera moves or play-by-play actions.
# [LOGIC] Block (The Invariant Anchor)
The pre-production container locking down persistent physical rules and spatial permanence.
Primes the model’s learned priors with invariant constraints (such as room geometry or object permanence) so the AI does not hallucinate impossible physical mutations during brief execution beats.
# [OPENING] Block (\(S_0\))
The composed starting still canvas (\(t = 0\)) establishing initial camera framing, character placement, and prop grips before motion begins.
Eliminates first-frame ambiguity so the model knows where every subject and camera starts. Contains zero kinetic state transitions (\(A \to B\)).
A structured prompt framework that organizes cinematic intent into static pre-production staging and dynamic sequential timeline beats.
Replaces unstructured paragraph prompting with an explicit state ledger, allowing directors to mentally simulate scenes and systematically evaluate AI output divergence.
# Phase 1: STAGING (\(S_0\))
The pre-production setup where static world rules, visual lookbooks, and the opening frame are locked before motion begins.
Eliminates prompt bloat by defining the room, physics invariants, and wardrobe once so kinetic execution lines stay brief and punchy. Contains zero temporal progression.
# Phase 2: EXECUTION (\(S_1 \to \dots \to S_n\))
The active shooting phase where camera movement, actor choreography, and sound evolve beat by beat across screen time.
Governs temporal pacing and kinetic mutations. Never redundantly restates static staging details already locked in Phase 1.
The workflow separation between Phase 1: STAGING (setup before rolling) and Phase 2: EXECUTION (active rolling camera).
Mirrors physical filmmaking by dressing the set, propping actors, and dialing in lighting before directing camera movement and actor performance.
The foundational division of an Auteur Script into Intent (purpose), Aesthetic (sensory world), and Execution (timeline).
Separates conceptual stakes, world lookbooks, and kinetic screen progression into distinct cognitive layers without enforcing a rigid authoring order.
Treating the generative video model like a skilled film crew rather than a deterministic code interpreter.
The director establishes structural coordinates and physical boundaries in the script, while the AI’s learned visual priors calculate natural momentum, weight, and gap-filling.
2. State Mathematics & Timeline
The state vector formulation, timeline decomposition, coordinate tags, and the recursive staging equation governing screen progression.
# Chaining Operator (
->)
A dual-role syntax operator commanding temporal progression across lines and multi-modal layering within a line.
Across lines, denotes sequential temporal beats (\(S_0 \to S_1\)). Within a line, layers camera, performance, dialogue, and sound into a single synchronized beat interval (\(\Delta t\)).
# Core Staging Equation (\(S_n = f(S_{n-1} \mid \text{STAGING})\))
The recursive state formula modeling each beat (\(S_n\)) as a transition from the previous state (\(S_{n-1}\)), conditioned on pre-production staging.
Encapsulates that every new moment on screen carries forward previous momentum, bounded by the invariant lookbook, rules, and initial anchor locked in Staging.
# Macro-State (\(S_n\))
A single complete line inside [EXECUTION] representing one continuous stretch of screen time.
Acts as the primary human pre-visualization unit, allowing the director to mentally simulate pacing, framing, action, and sound beat by beat before generating.
The cognitive role of a Macro-State line in pacing out human mental simulation.
Breaking a scene into one-line beats allows creators to scan, evaluate, and verify physical continuity without getting lost in dense walls of prose.
# Sub-State (\(s_i\))
An atomic coordinate segment within a Macro-State line that modifies strictly one physical, optical, or acoustic dimension.
Isolates distinct cinematic variables (such as camera optics or actor movement) to prevent instruction conflation and vector collisions.
# Sub-State Tag (\(\tau \in \mathcal{T}\))
An inline bracketed category label (such as [CAM], [ACT], or [AUDIO]) indicating which coordinate dimension is mutating.
A cognitive tool for the director to separate thinking into cinematographer, choreographer, and sound designer roles. Can be omitted when character limits are tight.
The three baseline coordinate dimensions ([CAM], [ACT], [AUDIO]) directing camera optics, actor performance, and diegetic sound.
Covers optics, physical choreography, and diegetic contact in standard cinematic lines, expandable to custom dimensions ([LIGHT], [PROP]) as needed.
Using mathematical state equations (\(x \mapsto f(x)\)) as a compact cognitive shorthand for human mental staging.
Serves as the creator’s visual state ledger to organize creative intent, explicitly avoiding claims that AI neural networks execute deterministic physics code.
3. Directorial Craft & Continuity
The relationship between human instructions and latent priors, state inheritance, physical continuity, and post-generation diagnostics.
The directorial philosophy of setting structural transition fences (\(S_{n-1} \mapsto S_n\)) without over-constraining the model with artificial second-by-second timestamps.
Gives generative video engines the creative freedom to calculate natural momentum and micro-acting within well-defined directorial guardrails.
# Camera Persistence [CAM 01]
Holding the existing camera framing and angle across multiple lines unless explicitly changed.
Allows the director to pace out complex, multi-beat acting performances across multiple lines of text without triggering accidental camera cuts.
Structural boundary coordinates set in the script to eliminate latent ambiguity.
By locking starting positions, mutating parameters, and target destinations, creators establish clear fences while allowing the AI’s learned priors to calculate natural momentum and micro-acting.
The post-generation practice of using explicit script coordinates to systematically isolate why an output diverged.
Evaluates rendered video against Staging and Execution coordinates to distinguish human prompt omission from stochastic AI variance, training directorial intuition over time.
The deep physical, anatomical, optical, and temporal patterns absorbed by generative video models from massive datasets.
The underlying generative engine that calculates natural gravity, weight, muscle tension, and motion momentum when prompted with clean structural boundaries.
The accumulated physical reality (held props, strain, wounds, postures) carried forward from beat to beat.
Reminds creators that characters remain physically bound by their previous actions, anchoring spatial and prop continuity across shots.
# State Inheritance (\(S_n = f(S_{n-1})\))
The continuity rule that every new beat automatically inherits the physical and optical reality of the moment immediately preceding it.
Ensures that held objects, arm fatigue, wounds, wardrobe conditions, and camera perspectives do not reset or vanish across cuts unless explicitly altered.
The practical rule of explicitly restating continuity-critical details across consecutive beats.
Compensates for token attention drift in AI models by reinforcing which hand holds a prop or whether a character is seated, serving as a pre-generation checklist.
TipExploring the Foundations
To see how these concepts integrate into an end-to-end directorial workflow, start with The Auteur Script Blueprint or explore verified production scenes in Reference Scripts.