First-Person
FPV charactersubjective camera
A first-person camera makes the lens function as a character's bodily viewpoint, with a body that owns it: hands entering the frame, breath, tools, locomotion. It is subjective seeing as continuous embodiment, not a single optical insert.
Narrative function
The audience is the character for a stretch of time. Lady in the Lake tries to tell a detective story this way.
Hardcore Henry runs it as an action body. Enter the Void keeps the camera inside a consciousness that still has a path through rooms.
Hands, a steering wheel, a gun, a door handled at the edge: those are the tells that a person, not an invisible observer, is looking. A Point of View shot can be one cut of what they see, clean of the owner.
First-person keeps the owner as a felt body. Object POV parents the camera to a prop instead.
SnorriCam looks at their face from their chest, which is the opposite ownership.
How it works
Parent the camera to the character's head or upper chest with a human field of view (roughly 24mm to 35mm spherical, not a tiny action-cam unless that is diegetic). Allow hands, shoulders at the extreme edge, or a tool to enter.
Add breath and footfall as camera motion only as much as a body would. Do not show the owner's face except in glass.
Prompt: first-person camera as [Subject] would see, hands entering frame, breath, a body that owns the lens.
When to use it
Use it when the audience must be the character, not merely glance as them. Concrete beat: running a wet overpass, gloved hands hitting the rail, breath in the cold, no face of the runner.
Do not use it for a two-shot argument; you will lose the listener.
Compared with similar shots
Point of View is the optical slice of the seen world, often an insert after a look, owner body forbidden. Object POV is a ball, a bullet, a phone, a vehicle.
Over-the-Shoulder Coverage keeps us behind someone we can still see. Fourth Wall has them look at us.
SnorriCam bolts the camera to the torso and stares at their face while the world slides.
In film
Hardcore Henry keeps the body in the lens: hands, weapons, impacts. Enter the Void rides a consciousness through space after the body is gone, still as a moving first-person path.
Lady in the Lake (1947) attempts extended first-person detective storytelling, other characters addressing the lens as Marlowe. Use those as embodiment grammar.
Prompt it
Require a body at the edges. Quote: First-person camera as [Subject] would see, hands entering frame, breath, a body that owns the lens.
Video: First-person view moving as [Subject] moves, hands or tools entering the edges. Add: no owner's face except a reflection, eye height of that person, not a drone, not an OTS shoulder.
What usually goes wrong
The usual wrong cousin is a clean Point of View plate (no hands) or a Follow Shot behind the actor. Tell: a back of a head, or a perfectly disembodied lock-off of a room.
Fix: hands or a tool in frame, motion as a walk or a turn of the head, owner invisible as a face. If the camera is bolted to a car hood with chrome in frame, you wanted Object POV.
Questions
What is first-person camera in film?
First-person makes the camera a character's body: we see as they see, with hands, breath, or tools entering the frame. It is continuous embodiment, not only a glance.
First-person vs POV
A POV shot is what they see, often a single insert with no owner body. First-person keeps a body owning the lens over time: hands, locomotion, breath.
When should I use first-person?
Use it when the audience should inhabit a body for a passage of action or perception. Do not use it for ordinary reverse-shot dialogue.
How do I prompt first-person for AI image or video?
Write: First-person camera as [Subject] would see, hands entering frame, breath, a body that owns the lens. For video, move as they move, tools at the edges, no owner's face.
Melies can generate this First-Person
Melies is the AI filmmaking workspace. This First-Person opens with the angle and prompt loaded, so the generator starts from the viewpoint instead of a generic portrait.