↘ Technologies R&D Patent PCT/FR2025/000021 France 2030 laureate

The tools to step in together.

Born from our productions, open to other creators: XR shows for a hundred people, live body capture and characters that talk with the audience.

100+spectators in one world

0markers, internet, infrastructure

73 msfrom gesture to screen with LIFE

4streams: body, face, hands, voice

01 — ⌜lft

Large Field Theater.

The all-in-one platform for multiplayer XR shows. Dozens of spectators share one virtual world in one space, deployed from road-ready flight cases.

↘ Hardware

Pre-wired flight cases

Server, network, charging and per-slot headset management. Modular in kits of 15 headsets. PICO 4 Ultra headsets, OpenXR standard.

↘ Software

A control room in the browser

Operations, fleet, mixed-reality calibration and supervision from a local web interface. Two shows side by side on the same hardware.

↘ Creation

An SDK for creators

Unity plugin, group management, avatars and sync. Canterville and Gaijin are our first productions on the platform.

LFT Mini

The whole platform, in a suitcase.

The same software that runs 100+ headset installations, in two cabin-size cases. From 3 to 15 headsets, for touring, demos and small venues.

LFT Mini — The kit — ↘ The kitEverything you need, in cases you carry.

LFT Mini — Server Case — ↘ Server CaseControl screen, server, network. 3 headsets.

LFT Mini — Device Case — ↘ Device Case4 headsets, charged and synced.

LFT Mini — Travel — ↘ TravelCabin size, on wheels.

LFT Mini — Inside — ↘ InsideOne slot, one light per headset.

↘ Built to tourIP67, TSA locks.

LFT Mini — Scale up — ↘ Scale upAdd cases, up to 15 headsets.

Worldwide distribution

02 — ⌜life

LIFE mocap.

Live Interactive Full Embodiment. The body, face, hands and voice of one or several performers, cleaned and streamed live into Unreal Engine. Developed with Dark Euphoria.

Dual rig in Unreal — Canterville character

Average latency: 73 ms
Performers: 3 validated, 10 targeted
LIFE Studio: Arcueil, OptiTrack
LIFE Mobile: 12U rack, set up in 45 min
Support: France 2030

↘Discover LIFE

03 — ⌜morph↗IA

morphIA.

Characters that listen, understand and answer. morphIA is our voice conversational agent, built into the XR engine: audiences talk with expressive virtual characters, with natural voice, lip sync and synchronised gesture.

Character studies for Canterville — Canterville characters

Voice: listening, answer, lip sync, gesture
Audience: several visitors at once
Runs: on site, offline
Integration: open-source Unity SDK, LFT
Guardrails: real-time moderation

In Ovid’s Metamorphoses, Morpheus takes human form to guide people through their dreams. morphIA does the same: a character who welcomes, tells, answers and adapts to each visitor, rather than a plain NPC.

Already on stage

Kephri

In the immersive Tutankhamun exhibition, Kephri the scarab walks visitors through. He tells the pharaoh’s story, places it in ancient Egypt, amuses with his airs, moves with his nostalgia, and answers everyone.

Victor Hugo

Let’s talk literature with Victor Hugo: the writer’s avatar welcomes visitors into his house and his work, guides them, answers their questions and reacts to what they do.

Mediation
A historical figure or an artist who adapts their words to each visitor.
Works that answer
The installation speaks and reacts: contemplation becomes dialogue.
Accessibility
Voiced content, with no screen or interface, for every audience.
Collective
With the Large Field Theater, several visitors talk in the same space.

↘ 04 — In the lab

What’s next.

The modules we develop with our partners, for our productions and for the platform.

LFT
Syncspace

Virtual venue scouting: the space modelled in 3D, visited remotely, the show laid out before arrival.
LFT × Olfy
Scent immersion

Three to five scents synced to the show timeline, for each spectator.
LFT × Thales · SACD
Security and rights

Secure analytics with Thales, authors’ royalties tracked with SACD.
LIFE
Auto-Rig

A humanoid rig in one click from any 3D model: 70% less preparation time.
LIFE
Spatial Relationship Engine

Gaze, distances and interaction zones between several performers, in real time.
LIFE
Unreal ⇄ Maya dual rig

A pure data layer and a polish layer, synced both ways.

↘ 05 — Research

With academia.

Our stages serve as a testing ground for scientific work with leading engineering schools. Atelier Daruma is a patron of the Digital Transformations Institute at Mines Paris – PSL.

Research project · Mines Paris – PSL · 2025

Mareva: physical space assignment and audience trajectory optimisation in a multiplayer immersive experience

Salomé Gobbi, Juan Pablo Duque Escobar¹

¹ Mines Paris – PSL · Supervised by Alexis Paljic, Professor, Centre for Robotics, Mines Paris – PSL · Funded by Atelier Daruma

Abstract

In Large Field Theater, around a hundred headset-wearing spectators move through the same physical space at once while following a linear story spread across several virtual rooms. This work proposes a tool to intelligently manage the trajectories of the headset fleet. The stage is modelled as a fully connected graph of sub-rooms; each room carries a population function over time, and the cost of a trajectory is the integral of occupancy across the rooms it visits. After comparing shortest-path (Dijkstra, A*) and assignment algorithms (Hungarian, Hopcroft-Karp), which do not suit time-varying weights, the authors adopt a permutation search optimised by a nodal representation that avoids repeated computation, producing a bipartite assignment graph at each time step. Tested on fifteen groups and ten sub-rooms, the model balances space occupancy; visualisation tools plot each room’s population and animate group trajectories.

Keywords — multiplayer audiences · space assignment · combinatorial optimisation · graph theory · collective virtual reality

↗Request the full PDF

Fig. 1 — Formalisation: the stage becomes a fully connected graph; the weight of an edge into a room is the integral of its population, and a trajectory’s cost is the sum of the weights it crosses

Fig. 2 — At each time step t, a bipartite graph links groups to free sub-rooms

Fig. 3 — Each group is given a free physical sub-room for the virtual room it is in, spreading use of the stage

Fig. 4 — The chosen algorithm: a permutation search that keeps the minimum-cost assignment

Fig. 5 — Authors’ simulation: group trajectories across the ten sub-rooms (mode 0)

Fig. 6 — Authors’ simulation: assignment variant (mode 1)

Fig. 7 — Results: population functions of the nine sub-rooms over the experience (fifteen-group test)

↘ Intellectual property

PCT/FR2025/000021

International patent for Large Field Theater, priority date 20 February 2025.

LFT^®Registered trademark
LIFE^™Trademark
morphIAProprietary technology

↘ The patent in brief

Transportable virtual or augmented reality multimedia system.

The invention protects the core of Large Field Theater: a fleet of headsets, a wheeled storage container that charges them and links them to its electronics, and a computer unit in a flight case. On site, the unit deploys its own wireless network, scans the space, generates the game environment from that scan, then connects and synchronises every headset within it. No fixed infrastructure: the theatre fits in two cases.