Video models learned from millions of real clips, and those clips were described using film vocabulary. That means the same words a director would use on set are the ones a model understands best. You don’t need film school; you need about ten terms.
Moves that go toward or away
Push-in (dolly in)
The camera moves steadily closer to the subject. It builds attention and emotion, and it’s the single most useful move for portraits and products. Prompt words: slow push-in, dolly in, camera slowly moves closer.
Pull-out (dolly out)
The opposite: the camera backs away to reveal more of the scene. Great for endings and for reveals, like a lone hiker who turns out to be standing on a huge glacier. Prompt words: slow pull-back revealing, dolly out.
Moves that go sideways or around
Pan
The camera stays in place and turns left or right, like turning your head. Good for landscapes and rooms. Prompt words: slow pan left, panning across.
Tracking shot
The camera travels alongside a moving subject, keeping pace with it. It makes motion feel fast and immersive. Prompt words: tracking shot following, camera moves alongside.
Orbit
The camera circles around the subject. It is the classic product-video move and works well for statues, cars and anything with an interesting shape. Prompt words: slow 360 orbit around, camera circles the subject.
Moves that go up or down
Tilt
The camera stays in place and tips up or down — from a climber’s feet up to the summit, for example. Prompt words: tilt up to reveal, camera tilts down.
Crane and drone shots
The camera rises or flies over the scene. These give scale and a sense of place, and they suit travel and real-estate clips. Prompt words: aerial drone shot, crane shot rising above, bird’s-eye view.
Shots where the camera doesn’t move
A static shot or locked-off tripod shot keeps the frame still and lets the subject do the moving. It’s underrated. For delicate motion — steam, candle flames, a blink — a still camera lets the detail breathe. If a render looks wobbly, adding “static tripod shot” often calms it down.
Handheld is the other extreme: a slightly shaky, documentary feel. Use it sparingly; it can turn into blur.
Framing words
- Extreme close-up: an eye, a drop of water, a fingertip.
- Close-up: a face or a single object filling the frame.
- Medium shot: a person from the waist up.
- Wide shot: the whole subject and its surroundings.
- Establishing shot: a very wide view that sets the location.
Combining them
Use one move per clip. “Slow push-in” plus “close-up” is a combination; “push-in, then orbit, then pull out” is three shots and will likely come out messy. If you want a sequence, render each move separately and cut them together.
Try a few on the text to video tool, or apply them to your own photo with image to video.
