Camera and lens
Lesson 5 of 22 in the course “AI Video, No Flops”. Free, with animations and a quiz.
Shot size
Shot size is how much of the person you see. A wide shot shows where we are and where we’re running. A medium shot shows the body and clothes; it’s the main shot for a hoodie ad. A close-up shows emotion on the face. A detail shows fabric, laces, the logo. Models understand these terms: wide shot, medium shot, close-up, extreme close-up. Cutting from wide to detail reads easily, as if the viewer is walking closer.
Camera height
The same model looks different depending on where the camera stands. From below she looks strong and confident, which is why sports brands love low angles. At eye level it feels honest, like real life. From above a person seems smaller and more vulnerable. Straight down, the city turns into a map. For the hoodie ad Mira picks a low camera: “low angle” in the prompt.
Lens in millimetres
The number on a lens is the focal length. The smaller it is, the more of the world fits in the frame. 16–24mm is wide angle: lots of space, motion feels faster, but a face up close gets distorted. 35–50mm is roughly how the eye sees. 85mm and up is portrait: the background is blurred and seems “pressed” against the hero. Models understand “24mm lens” or “85mm portrait lens” and change the picture exactly that way.
Camera moves
The camera is an actor too. Static: the camera holds and all attention goes to the action. Pan: it turns in place. Dolly in or out: it moves toward or away from the hero, adding tension or revealing the place. Tracking: it moves alongside the running model, the best option for Mira’s video. Crane: it rises or lowers smoothly. Handheld: natural shake, the “shot on a phone” feel. In studios like Higgsfield many of these moves come as ready-made presets.
One move per shot
A common beginner mistake is cramming a dolly, an orbit, a zoom and shake into one shot. The model gets confused: the background “melts”, walls bend, the face changes. Pros use one move per shot and name it plainly: “slow dolly in on the face” or “camera tracks from the right at waist height”. Want a complex fly-through? Cut two simple shots together in the edit.
Depth of field
Depth of field is how much of the space is in focus. Shallow depth: the model is sharp and the background turns into soft blobs of light. That separates the hero from a busy street and makes the shot look expensive. Deep focus: everything is sharp, from the face to the signs, which feels documentary. In the prompt, write “shallow depth of field, soft bokeh” or “deep focus”. Bonus: a blurred background hides small glitches behind the hero.
Camera to match the genre
The camera tells the viewer what they’re watching. UGC, meaning “video from a regular person”: handheld, vertical, a wide phone-like angle around 24–28mm. Premium ad: a smooth gimbal and 50–85mm. Film: static shots and slow dollies on 35mm. Sport and action: tracking on a wide angle. Mira makes two versions: premium for the website and UGC for TikTok, because in the feed “shot on a phone” often feels more native.
Check yourself
- Mira needs to show the hoodie fits perfectly. Which shot is the main one?
Correct answer: Medium: body and clothes. A medium shot shows silhouette, fit and fabric. Wide and detail shots are good extras but won’t show the fit. - In the clip the camera orbits, zooms and shakes at once, and the walls bend. What helps?
Correct answer: Keep one move and name it plainly. Several moves at once overload the model. One named move keeps the geometry stable; a complex fly-through is built from two simple shots. - Signs in the background of Mira’s clip keep glitching. Which trick hides it fast?
Correct answer: Shallow depth of field: soft background blur. A blurred background makes letters unreadable, so the glitch just isn’t visible. And the hero looks more premium too. - For TikTok Mira wants the video to look like a friend shot it. What does she pick?
Correct answer: Handheld, vertical, wide angle around 24–28mm. That’s exactly what phone video looks like: a little shake, vertical framing and a wide angle like a smartphone camera.