Two sensors in one housing: a long lens that identifies a face at 40 m and reads plates, and a wide one that keeps the whole scene. Solves the oldest problem in zoom, which is losing the context.