Vanguard
Own-likeness character study · 2026 · Higgsfield, Premiere
The brief. Can one person’s face survive a superhero suit, a helmet, a hundred camera angles, and a city-wide action sequence without turning into someone else?
▶ Cinematic trailer · 1:02
PROOF OF CONCEPT (1:02)
Landing, launch, flight, close-ups. Stills generated first, then animated, then cut in Premiere. Personal piece.
Approach. Turnaround sheets first: front, back, three-quarter, helmet on and off, until the suit was locked. Then environment plates (a Golden Gate under siege), then the sequence: landing, launch, flight, close-ups. The 62-second proof of concept is the output. The hundred-frame turnaround set is the real work.
What it shows. The still-first discipline that makes AI motion usable: lock the character, lock the world, then move.
Lock the character. Suit lock, three-quarter: one suit, one face, held across every frame before anything moves.
Alcatraz siege. Golden Gate behind, ships overhead. Personal piece.
Lock the world. Then move. Bay Bridge, dusk.
Transamerica, night.
Downtown, straight down.
Freefall, rain on the visor.
Top of the city.
Under the bridge, low and fast.
Off duty. Rooftop. Glazed.
Aisle seven. Helmet rides in the cart.
Threat level: avocado toast.
Still on call.
Deck. Espresso. Before sunrise.
Fixes his own gear.
Drop-off. Lunch handed over.