Cross-Mapping
Cross-Mapping is the translation layer between how Fluxx describes itself and how standard pipeline conventions describe the same (or adjacent) concepts. Two audiences need it: engineers who already know AYON, ftrack, or ShotGrid and need the rosetta stone; counsel who needs to understand which Fluxx terms have prior-art collisions versus genuinely new ground.
The Fluxx-native primitives (see Pipeline Architecture) compose into a coherent platform that does not exist as a unit anywhere else. Each Fluxx primitive maps to a standard concept partially, fully, or not at all. The divergences are intentional; the alignments come from porting AYON patterns where they solve the right problem (see AYON Port / Reject).
Rosetta table
| Fabrixx description | Standard pipeline assumption | Divergence note |
|---|---|---|
| Fluxx Vault (content-addressed, multi-tenant, provenance-attached) | AYON Folder / Product / Version / Representation entity model | Aligns conceptually (hierarchical with provenance plus version semantics). Diverges on substrate (content-addressed cloud object storage versus file system paths) and on schema extensions (ComposiTrack sub-frames, stage configuration, AI provenance fields, finish-track assignment all Fluxx-native). |
| ComposiTrack three sub-frames (ChromaTrack, ContentTrack, CompensationTrack) | ShotGrid Representation | Contradicts. ShotGrid’s data model has no place for a three-sub-frame capture. AYON’s entity model can hold the data, but the upstream pipeline producing the sub-frames is not something AYON’s host integration knows how to drive. |
| Pipeline-owned VM lifecycle | AYON Launcher long-lived host | Contradicts. AYON’s launcher assumes a long-lived artist workstation; Fluxx assumes short-lived VMs as the unit of compute. No AYON analog for vm.provisioned or vm.teardown events. |
| Multi-tenant scoped at every entity | AYON single-deployment-per-studio | Contradicts at the data-model tier. AYON’s tenancy is “run multiple instances.” Fluxx scopes tenant at every entity, every event, every cost ledger entry. |
| Near-Time PostViz on-stage checkpoint (30 to 60 minutes at quarter resolution, overnight high-res to final pixel) | Flow Capture next-day dailies | Contradicts. Traditional dailies assume overnight cut → review-tomorrow. Fluxx moves the loop to same shoot day before strike. The whole dailies-tool category is structurally obsolete for Fluxx productions. |
| ftrack on-prem integration substrate | Autodesk Flow Production Tracking cloud-only | Contradicts at deployment model. Cloud-only structurally rules out Flow PT for on-prem-security-sensitive productions; ftrack on-prem is Fluxx’s integration substrate per Q-G15. |
| Per-shot / per-finish-track / per-tenant cost attribution | Deadline worker-pool utilization | Diverges. Render-farm managers track worker utilization, not per-shot dollar cost. Fluxx layers attribution above whichever job-queue substrate ships. |
| Vault publish (Pyblish-shaped, Vault-aware) | Pyblish validate / extract / integrate | Aligns at the publish-stage tier (verbatim port). Diverges at the integrate step (Vault content-addressed commit versus file system). |
| Leecher / Processor / Transmitter for ftrack integration | AYON’s three sibling tracker addons (ftrack, ShotGrid, Kitsu) | Aligns. Three-service decomposition is the right pattern. Fluxx ports verbatim, tracker-agnostic at the contract tier so Kitsu integration is structurally identical. |
| Cloud-VM deployment topology | NVIDIA Omniverse Cloud | Aligns at topology (DCC runs in cloud VM, artist remotes in, storage on object store). Contradicts at substrate ownership (Nucleus is NVIDIA-owned; Vault is Fluxx-owned). |
| Vault AI provenance schema | LoftApp standalone provenance tracker | Aligns at field set (prompt, seed, model, model_version, LoRA, generation_timestamp, generation_tool, parent_asset_id, C2PA). Contradicts at framing (LoftApp is a vertical SaaS the production has to wire in; Vault absorbs the schema as the system of record). |
| Agentic Sequence Assistant sequence-propagation | No incumbent equivalent | Contradicts by absence. No incumbent pipeline has a primitive for sequence-scoped agentic propagation across shots. Fluxx Layer 3 native. |
| ComposiTrack (canonical product name) | ChromaTrack (patent filing) | Same artifact; two names. ComposiTrack is the canonical product name across all Fluxx surfaces per Q-G88. ChromaTrack appears only in the original patent filing as the legal record. Glossary alignment: use ComposiTrack everywhere except patent context. |
| Velvet (UI library) | “Brand Master UI” or “Fluxx UI” (deprecated naming) | Velvet is the canonical UI library name across all Fluxx surfaces per Q-G71 and Q-G72. |
Divergence index (intentional divergences)
Five places where Fluxx structurally diverges from standard pipeline assumptions, ordered by leverage.
1. Cloud-VM-and-object-storage versus LAN-and-file-system
Centerpiece contradiction. Patterns AYON solves (DCC bridge, addon SDK, publish/load, event spine) are right. The substrate AYON assumes is wrong. Port the patterns; reject the substrate. File system paths cannot represent cloud object storage; AYON’s load plugins would need to be rewritten end-to-end to handle multi-tenant cloud-storage URLs. Fluxx’s Vault translates the storage URL at the substrate tier so DCC plugins see local cache paths.
2. On-set Near-Time PostViz versus next-day dailies
Traditional review topology assumes review-after-the-fact. Fluxx collapses the dailies loop to before-strike same-day. Camera-to-cloud ingest pattern is fine; destination is Vault, not Frame.io. The dailies-tool category structurally obsoletes for Fluxx productions.
3. AI provenance as Vault schema extension versus as separate vertical product
LoftApp’s window closes between now and Frame.io shipping C2PA-aware metadata as workflow (probably 12 to 18 months). Fluxx absorbs the field set inside Vault. The provenance graph becomes a substrate property of the platform; productions meet EU AI Act Article 50 disclosure obligations by default. Q-G18 plus Q-G50 govern the policy layer above the schema.
4. Fluxx-owned Vault versus NVIDIA Omniverse Nucleus runtime dependency
Omniverse is the closest reference topology but Nucleus binds production to NVIDIA’s roadmap and USD opinions. Port the topology, reject the substrate. The architectural rail “no third-party at the core” is written about this divergence.
5. Per-shot cost attribution layer versus render-farm worker-pool-only
Multi-tenant JV economics require per-shot, per-finish-track, per-tenant cost attribution. Fluxx layers this above any job-queue substrate (Deadline, Tractor, Royal Render, or Fluxx-native). The substrate choice is open per Fluxx_Pipeline_Architecture_v0.1.md §10; the attribution layer is Fluxx-owned regardless.
Where Fluxx has no equivalent in the standard
These are pure Fluxx-native primitives with no standard pipeline analog:
- ComposiTrack three sub-frames at capture time
- Two-track finish through one workflow (Near-Time plus Near-Time VFX)
- Pipeline-owned VM lifecycle with
on_vm_provisionedandon_vm_teardownhooks - Per-shot finish-track assignment as a Vault field
- Agentic Sequence Assistant sequence-propagation
- Pre-comp recipe as a serializable artifact passed from CompViz to Fabrixx
- Stage configuration events (
stage.wall_repositioned,stage.lighting_recipe_applied)
Where the standard has no equivalent in Fluxx
Standard pipeline concepts Fluxx does not adopt:
- Local file system as primary asset addressing
- Long-lived artist workstation as the unit of compute (Fluxx uses ephemeral cloud VMs)
- Single-tenant data model (Fluxx is multi-tenant from v0.1)
- Next-day dailies as a separate review tier (Fluxx delivers Near-Time PostViz on-stage)
- Render-farm worker-pool utilization as the primary cost view (Fluxx attributes per-shot)
- AI provenance as a separately purchased product (Fluxx absorbs the schema in Vault)
Source dependencies
02_Working/Pipeline_Research/Cross_Mapping_v0.1.md(primary)02_Working/Pipeline_Research/Fluxx_Pipeline_Architecture_v0.1.md(Fluxx-native primitives)02_Working/Pipeline_Research/AYON_Architecture_Deep_Dive.md(AYON reference)04_Fluxx_Shared/_Master_Canon/Q-G/Q-G88_ComposiTrack_Canonical_Name.md(naming canon)02_Working/Strategy/Company_Overview/Glossary_Source_2026_05_03.md(Fluxx term canon)
Synthesis gap
Per Cross_Mapping_v0.1.md §5: virtual stage primitives in the pipeline data model is the highest-priority Kevin gap-fill. The ComposiTrack three sub-frames are scoped (Q-G3 plus Q-G38); Warp camera-layer inputs, RapidWall configuration state, and programmed lighting recipes are not yet schema-defined in canon. The stage configuration event family is named (stage.configuration_loaded, stage.wall_repositioned, stage.lighting_recipe_applied) but the underlying entity model has not been written down.
Audience visibility
industry, vendor, counsel, investor.