VFX · Pipeline
Building a quiet pipeline: how we render 240+ shots without crunch
A field report from inside our LA + London comp pipeline — what we keep small, what we automate, and why our crit room beats our render farm every time.

01 · The premise
We ship around forty projects a year. None of them get crunched. That is the rule — not the rough average, the rule. The pipeline I'm going to describe in this essay is the system we built specifically to keep that promise.
This is not a manifesto, and it isn't a tools list. It is a description of decisions we have made and re-made over five years of running a VFX house at the seam between film and series work. Some are technical. Most are about people.
"You can't render your way out of a critique problem."
— Jonas Riese, VFX Supervisor
02 · Pipeline anatomy
Strip away the brand names and our pipeline is small, deliberately boring, and built from the same handful of tools every other VFX house in the city already runs. None of it is exotic.
What sets the work apart is not the software. It is the discipline around the boundaries — the rule about who hands work to whom, in what state, and by when. Get the boundary right and the tooling almost doesn't matter.
The first rule is simple and non-negotiable: every shot has exactly one owner. That owner is the only person who closes the shot. Producers don't override owners. Directors don't override owners.
03 · The crit room
The reason none of this software actually matters is the room we do it in. We have a standing crit at 10am LA / 6pm London every weekday. It runs for forty-five minutes. Everyone shows up. We watch shots, take notes, and end the session.
The crit room is the only place where direction happens. We have learned, painfully, that anything decided outside the crit room ends up needing to be re-decided inside it.
04 · When it breaks
The pipeline fails maybe four times a year. The failure mode is always the same: a shot owner gets bypassed. The thing that fixes it is naming it out loud. When it happens, we name it the next morning, and we re-route the work.
05 · Closing notes
The render farm is a beautiful machine. It is not the engine of our studio. The engine is the room — the people in it, and the discipline of letting them be the only people who decide.