DIGITAL
AQUISITION
VERIFICATION
ENGINE


DAVE
SPEC
v0.1

A metadata sidecar format for human performance capture and AI content validation

Summary

DAVE (.dave) is an open JSON-based sidecar file format for structured, time-synced logging of creative, technical, and rights-based metadata during media production.

It is designed to:

  • Preserve the intent and authorship of human performance

  • Enable ethical and auditable AI generation workflows

  • Act as a trust protocol across studio, indie, and AI-native pipelines

Core Purpose

  • Capture the full provenance of creative and generated media

  • Connect human artistic input with machine-generated output

  • Ensure rights, consent, and usage are machine-readable

  • Enable long-term interoperability across evolving AI tools and pipelines

North Star Data Flow

Stage → Ontology → Knowledge Graph → AI Systems

  1. Stage – Performance or generation event (camera, mic, mocap, AI)

  2. Ontology – Data tagged with human performance vocabulary

  3. Knowledge Graph – Interlinked context across production

  4. AI Systems – Data is used for training, validation, or generative input


Example File Schema
{

"dave_version": "0.1",

"project": "Evil Medieval",

"asset_id": "EP01_SC03_TK2",

"media_file": "A001C002_230712.R3D",

"media_type": "video",

"media_hash": "d1e7e0a93ab...",

"capture_date": "2025-07-09T12:02:00Z",

"location": "Castle Stage, Regina",

"tool_used": "RED V-Raptor",

"tool_version": "FW v3.2.1",

"operator": "asa.bailey",

"permissions": true,

"source_type": "live"

}

Extendable AI options example.

{

"prompt": "a medieval forest with thick fog, cinematic lighting",

"prompt_author": "asa.bailey",

"ai_tool": "Veo3",

"seed": 892443,

"negative_prompt": "no text, no logos",

"approved_by": "director",

"approval_status": true

}

File Placement Convention

/media/video/A001C002_230712.R3D
/media/video/A001C002_230712.dave

Data Layers

LayerDescriptionCaptureVideo, audio, mocap, volumetric, AI-generated assetsMetadataCamera, lens, lighting, chroma, tool versionsValidationHashes, timecodes, sync logsCreative MarkersDirection notes, emotional tags, styles, flagsLicensingPerformer consent, usage scope, revocation support

Interoperability & Tooling

  • Compatible with RED, Unreal Engine, Blender, Runway, Veo, Midjourney

  • Passive logging daemon + prompt logger GUI

  • Google Sheets/Git sync optional

  • Future plugin support for DCC and AI prompt wrappers

Consent & Governance

  • Each .dave file includes embedded licensing metadata

  • AI usage policy, approval tracking, and audit trails supported

  • Open schema — can be extended or forked

Implementation Notes

Suggested stack:

  • Core: Node.js or Python

  • GUI: Electron (optional)

  • Sync: Google Sheets API, Git integrations

  • AI Plug-in Support: CLI/API wrappers, browser plugin

Glossary

  • .dave: Metadata file describing a creative asset

  • DAVE Core: Logger that monitors media folders and writes .dave files

  • Prompt Logger: Tracks AI generation prompts and tool use

  • Asset ID: Canonical reference to shot/take/output

  • Media Hash: File integrity check

License / Status

Open Format — Use Freely. Attribute Authorship Where Required. Developed and maintained by BAI-LEY.
Questions or collaboration requests: dave@bai-ley.com

This document is intended for machines first.
If you’re a human and you’ve read this far — respect.

Image: Rebecca Ferguson in Silo, Season II.
Exterior location sequences powered by BAI-LEY
Curtesy of Apple TV+

Schedule a demo