Balenciaga on visionOS
Balenciaga’s runway shows on Apple Vision Pro: a VR180 player and a lookbook, designed for clients wearing a headset for the first time.
- Lead Product Designer
- Atomic
- 2 years
- Apple Vision Pro
- Vision Pro sold slowly, so the app moved from the public to Balenciaga’s top clients, in boutiques and at private events. Most had never worn a headset.
- The budget moved from 3D accessories to the footage: better cameras, a higher bitrate, shows pre-downloaded on boutique headsets. The interface is pressed by hand, at arm’s reach, and the latest show plays in one tap.
- 7 showsCouture and Ready-to-Wear, released on the same foundation
- 50 → 150 MbpsVideo bitrate: public streaming profile → local playback in boutiques
- 100% offlineBoutique playback: every show is pre-downloaded on the headset, none waits on the network
From living rooms to boutiques
The app is an immersive media player and a lookbook. Windows to browse shows, looks and accessories by season; a fully immersive space to watch a show in VR180.
It launched as a public app. When the headset sold more slowly than expected, it became a tool for the brand’s top clients, in flagship boutiques and at private events. The user changed from an owner who knows the device to a guest who has worn it for a few minutes.
Before. The budget went to interactive 3D accessories, to look at in your own living room, and to video that held its quality on a small home connection.
After. The budget went to the filming: upgraded cameras and higher-quality footage, played from the headset itself.
I led product design at Atomic for two years, on the headset from day one. Four decisions shaped the app.
Design for someone who has never worn a headset
Constraint. A sales advisor opens the app, then hands the headset over. The client has never worn one, and the headset is not calibrated to their eyes.
Decision. Everything is pressed by direct touch. Windows open within arm’s reach, so every control is hit with a fingertip and never aimed at with the eyes: eye calibration is skipped entirely. Controls sit close to the hands, scrollable areas are large. The home window is one cover and one Play button: the latest show, playing in one tap.
Trade-off. Accessories and music moved to a side menu, one step further away.
Immersion as a second step
Constraint. Apple’s guidelines existed; shipped apps to learn from did not. Comfort and distance had to be judged in the headset.
Decision. Browsing happens in windows, with the room still visible. Immersion is gradual: one tap plays the show, a second opens the full VR180 session. There, the playback bar and the look strip are anchored below the horizon line, under the runway: the head stays level on the show, a glance down reaches the controls with no strain on the neck, and nothing covers the models.
Trade-off. One more tap before the full show, so no one lands in a 180° video before they are ready.
Brand in the type, controls from the system
Constraint. Balenciaga’s identity is brutalist: condensed capitals, raw black and white. A client on their first headset needs controls that look pressable, with room for a fingertip in the air.
Decision. The brand stays in the type and the black. Glass, depth and feedback are the system’s, so every target reacts like the rest of the headset.
Trade-off. Less of the brand’s raw look in the controls. It shows in the content and the typography instead.
Spend on the footage, play without a network
Constraint. The public app streamed to homes, so its video was tuned for small connections. In a boutique the show has to play, whatever the network does.
Decision. One app, two ways to deliver a show. For the public, streaming: we worked with a network engineer on a delivery architecture for low and high bitrate video, tested on a wide range of networks. For boutiques, the headsets pre-download the shows and play them with no connection at all.
Trade-off. The budget came out of the interactive 3D accessories.
One skeleton, seven shows
Seven Couture and Ready-to-Wear shows have been released on the same skeleton. The handoff to developers is automated: after each show we ask the client for the same set of assets. My part is the specification: the spatial layout schema, the aspect ratio of each asset, the design specs. Video and pipeline engineers wrote the ingestion scripts that apply it: naming, transcoding, cropping and conversion to spatial images.
What cannot be compressed is the video itself: treating and encoding each show’s 8K VR180 footage from the RAW files.