- Trace the volume's signal chain from video server through nDisplay to the LED cabinets.
- Bring an Unreal environment up on the wall and explain what camera tracking and the inner frustum are doing.
- Run first-line LED cabinet troubleshooting without making anything worse.
- Say concretely how designing for a camera differs from designing for a headset.
- Pitch a final project in three minutes: concept, venue, scope, and what you'll cut first.
2:00 · Brief The world, photographed
In VR, the world wraps around one viewer. In the volume, the world exists for exactly one eye: the cinema camera. Everything about virtual production follows from that inversion — and operating the room is a craft in itself, one that studios hire for by name.
The volume, end to end
Our stage is a chain, and you should be able to draw it from memory: the Unreal project runs on the video server; nDisplay splits the rendered world across render nodes and maps it onto the LED cabinet grid; the camera tracking system feeds the physical camera's position back into Unreal, which renders a higher-fidelity inner frustum — the patch of wall the camera actually sees — in correct perspective for the lens (this is ICVFX, in-camera VFX). When something looks wrong on the wall, you debug the chain in order, not by panic-clicking.
LED cabinets are hardware, and hardware fails
The wall is a grid of individual LED cabinets, each with power, data, and its own opinions. The first-line troubleshooting discipline: identify which cabinet, not "the wall"; check the obvious physical layer first (power, data connections, seating); know the reset procedure for a single cabinet; and log what you did. A dead cabinet mid-shoot is normal life on a volume — the crew member who stays calm and methodical is the one who gets hired back.
Designing for camera vs. headset
A headset scene must survive scrutiny from any angle at 90 fps. A volume scene must be flawless from one moving viewpoint at 24 — which means you can spend everything where the lens looks and cheat everywhere else. Depth reads differently too: the wall provides parallax only via tracking, so composition, lighting, and atmospheric depth do the work your stereo eyes did in VR. If your final targets both venues, you are effectively art-directing two different pieces from one scene.
Futuresource counted ~200 LED ICVFX volumes worldwide (end-2023 data) and forecasts 18% annual growth in installs through 2028 — driven by small and medium stages like ours, not mega-volumes. UE 5.8 backs this week directly: Live Link Hub and Movie Render Graph are now Production-Ready, with nDisplay support. Discussion: map the VP crew roles — which one is your entry point?
2:45 · Guided lab Crewing the volume
We rotate through crew positions in the VP Studio: server, wall, camera, and talent. Everyone touches every station. The agents come with us — the video server project is a UE project like any other, and small environment fixes happen live between setups.
- Walk the signal chain
Physically. Start at the video server, follow the path to the render nodes, the cabinet data runs, and the tracking system. Sketch the chain in your notes — this diagram is your debugging map for Week 15.
- Open the nDisplay configuration
On the server, open the stage's nDisplay config asset and identify how the cabinet grid is described: viewports, render nodes, and the stage's real-world dimensions. Don't edit it — read it until it stops being magic.
- Bring an environment up on the wall
Launch this week's demo environment through the studio's standard launch procedure. Note the checklist order: server project open, render nodes connected, then content live on the cabinets.
- Verify camera tracking
With the tracked camera live, move it and watch the inner frustum follow on the wall. Then break the illusion on purpose: look at the frustum from where the camera isn't. Understanding why it looks wrong from your eyes is understanding ICVFX.
- Cabinet troubleshooting drill
One cabinet gets "failed" for the exercise. Practice the discipline: identify the cabinet, check power and data at the physical layer, follow the studio's reset procedure, and write the one-line incident log entry.
- Shoot a plate
Put a classmate in front of the wall, frame a shot, and judge it on the camera monitor — not with your eyes. Adjust environment lighting until the talent sits in the scene. This is the moment camera-vs-headset stops being theory.
- Live environment notes via agent
Between setups, use Claude Code on the environment project for small directed fixes — "raise the ambient level in the alley by a stop," "scale down the props against the back wall." Review the diff before anything relaunches on the wall: on a stage, an unreviewed change costs everyone's time, not just yours.
The studio is a shared instrument. Nothing gets plugged, unplugged, or updated without the station lead knowing. Every change to the server project goes through version control — the stage machine is never the only copy of anything.
4:00 · Crit Project 3: Project of Choice / VP Collaboration
Ten minutes per project. Choice-track pieces run on their target headset; VP-collaboration pieces run on the wall, where they were built to live. We crit against the production rubric — craft, concept, interactivity, process — and your AI development log is open on the projector alongside the build. Be ready to answer for any node in your graph: that's the standing deal.
5:15 · Crit / share Final project pitches
Three minutes each: what you're making, whether it lives in a headset, on the wall, or both, what "done" looks like at exhibition, and the first thing you'd cut if the sprints run short. The room asks one hard question per pitch. Approved pitches lock this week — the sprint weeks are for building, not deciding.
Homework — due before Week 13
| Task | Deliverable |
|---|---|
| Revise your pitch from tonight's feedback into a locked one-page brief. | Brief committed to a fresh final-project repo: concept, venue, scope, cut list. |
| Plan the sprints: what ships in Week 13, what ships in Week 14, what's exhibition-only polish. | Sprint plan in the repo — three goals, each demoable in a sentence. |
| Gather your raw material before production starts: references, assets, scans, sounds. | Reference board + assets staged in the project; AI log entry #1 for the final. |
Resources
- nDisplay overview — the official map of what you saw on the server today.
- In-Camera VFX overview — Epic's ICVFX documentation, inner frustum included.
- Futuresource LED volume report — the market you'd be crewing into.
- Virtual production crew roles — find your entry point before crit discussion.