Planning Shots and Writing Video Prompts
Good AI video starts before generation, with a plan.
From idea to shot list
- Goal and platform: a 20-second Instagram reel? A 60-second product explainer?
- Script: what is said or shown, second by second.
- Storyboard: one frame per shot (sketches or generated stills).
- Shot list: each shot's content, length, camera movement and transition.
Video prompt structure
[Subject] + [action] + [setting] + [camera movement] + [shot type] + [lighting] + [style] + [pace]
Example:
A barista slowly pours steamed milk into a latte, forming a heart, in a sunlit café;
slow push-in, close-up, warm morning light through the window, shallow depth of field,
cinematic, gentle and calm
Camera language
| Term | Movement |
|---|---|
| Static | camera does not move |
| Pan / tilt | camera rotates horizontally / vertically |
| Dolly / push-in / pull-out | camera moves toward or away |
| Tracking | camera follows the subject |
| Orbit | camera circles the subject |
| Crane / drone | rising or aerial movement |
Keep each shot to one clear action and one camera move; complex instructions confuse current models.
Image-to-video for control
Generate or design the exact first frame you want (correct product, framing and colours), then animate it with a motion prompt. This gives far more control than text-to-video alone.
Exercise
Plan a 20-second promotional reel for a coffee shop: a script, a six-shot storyboard with stills, and a prompt for each shot including camera movement. Keep a log of prompts and settings.
Check yourself
- Why plan a shot list before generating?
- Why keep one action and one camera move per shot?
- How does image-to-video give more control?
Votre progression est enregistrée dans ce navigateur uniquement.
Envoyer pour relecture (10 points)