Why one clear action beats five vague ones when you are prompting video.
Describing motion is fundamentally different from describing a still. A photo can hold a dozen simultaneous details; a few seconds of video can only really deliver one clear movement before it starts to feel chaotic or muddy.
One movement, clearly stated
"She turns to face the camera and smiles" gives a video model one thing to execute well. "She turns, smiles, laughs, reaches for her coffee and glances out the window" asks for five things in the time there is only room for one - the result usually blurs all five into something unconvincing.
Direction, not just description
When you need more control than one free-running action gives you, use keyframes: set a first frame and a last frame and let the motion fill in between. That turns a vague request into an exact instruction about where a shot starts and ends.
- 1Pick the single movement that matters most in this clip.
- 2State it as one clear action, not a list.
- 3If you need precise start and end points, set keyframes instead of describing the whole arc in words.
If a generated clip feels chaotic, the fix is almost always to cut the request down to one action, not to add more descriptive words.
Motion has its own grammar; sound has another. The next lesson covers directing voice, music and sound effects with the same precision.