Skip to content

Latest commit

 

History

History

Folders and files

NameName
Last commit message
Last commit date

parent directory

..
 
 

README.md

Core Fields and Typical Workflows

The Video Notation Schema mirrors a professional video production pipeline. Below is an overview of the key schema sections, their intended use, and practical guidance for integrating them into your workflow.


1. Metadata (metadata)

Establishes foundational information about the project.

Key Fields:

  • title: Name of the project
  • duration_seconds: Overall duration
  • video_type: Format (e.g., advertisement, short film, educational)
  • aspect_ratio, frame_rate, resolution: Technical specs
  • language, target_audience, keywords: Contextual metadata
  • accessibility: Includes captions, audio descriptions, flashing/motion warnings

Tip: Define early, as it governs the global frame for AI interpretation.


2. Characters (characters)

Defines consistent, reusable character profiles for appearance, voice, and behaviour.

Key Fields (per character):

  • character_id, name, appearance, wardrobe
  • age_range, gender, ethnicity_descriptor
  • personality_traits, vocal_description, accent

Overrides: Scene-level details like wardrobe_override, action_details, or behaviour supplement or replace global character traits.

Tip: Define once, and reference throughout for consistent rendering.


3. Global Elements

Reusable definitions that simplify complex projects.

3.1. Props (global_props)

Physical items used across scenes. Override with prop_id_reference in scenes.

3.2. Audio Elements (global_audio_elements)

Reusable audio (ambient, sound effects, music loops). Reference with audio_id_reference.

3.3. Subjects (global_subjects)

Non-character elements (e.g., animals, scenery). Referenced via subject_id_reference.

3.4. Motion Graphics (global_motion_graphics)

Predefined motion graphics (e.g., infographic, animated_logo). Customise with overrides.

3.5. Lighting Presets (global_lighting_presets)

Reusable lighting styles (e.g., golden_hour_window). Referenced with preset_id_reference.

3.6. Tone Presets (global_tone_presets)

Reusable emotional tones (e.g., calm_inviting_tone). Also referenced via preset_id_reference.


4. Global Style (global_style)

Sets overarching artistic direction for the video.

Key Subsections:

  • shot: Camera settings (angle, motion, focus, lens)
  • subject: Styling for background or unnamed subjects
  • cinematography: Lighting/tone (can reference presets)
  • audio: Ambient, music, dialogue defaults
  • color_palette: Defines visual tone (hex codes or themes)

Tip: Define after characters to unify visual/auditory tone across scenes.


5. Scenes (scenes)

Defines an ordered array of narrative or visual moments.

Key Fields:

  • scene_id, narrative_role, duration_seconds
  • location, time_of_day, environment
  • subjects: Characters/objects present and their actions
  • visual_details: Props, motion graphics, timing breakdowns
  • audio: Dialogue, ambient overrides, sound effects
  • transition_to_next_scene: Style and logic of scene cuts

Overrides: Use shot_overrides, cinematography_overrides, and scene-specific settings to adapt global styles.

Tip: Compose scene-by-scene using previous definitions as a base layer.


6. Annotations (annotations)

Enables human-readable production and post-production notes.

Key Fields:

  • direction_notes: Director-style creative instructions
  • post_production: VFX, editing, sound mixing
  • production_assets: Linked image/audio references or concepts

Tip: Use for coordination across creative teams.


AI Video Platforms

  • The JSON can be use as prompt to any text-to-video AI model.

Schema Integration Tools

Editors & Validators

  • Visual Studio Code + JSON/YAML plugins
  • Online JSON Schema Validators

Libraries

  • Python: jsonschema, pydantic
  • JavaScript/TypeScript: ajv, zod, joi
  • Go: gojsonschema

Final Thoughts

The Video Notation Schema provides a structured, extensible way to craft AI video prompts with precision and reusability. From granular camera work to production notes, it enables creators to produce consistent, high-quality outputs at scale.