# The camera-first note I use before image-to-video generation

> Published 2026-08-26 · https://www.promptzone.com/yan_alex_f238fa660e011774/the-camera-first-note-i-use-before-image-to-video-generation-i31

I have been getting more predictable results from image-to-video tools when I write the camera note before the style note. It sounds small, but it changes how I judge the first draft.

The short version of the checklist is:

1. Start with what should stay stable in the image.
2. Add one camera movement, not three.
3. Say how long the motion should feel.
4. Describe the ending frame before adding mood words.
5. Keep a separate line for things that should not change.

For example, I usually write the subject lock first, then a slow push-in or sideways drift, then the final frame. After that I add lighting or texture. When I test a still image in [Image to Video AI](https://imagetovideogen.com/), this order makes it easier to see whether the prompt failed because of motion, continuity, or style.

A prompt that has worked better for me lately looks like this:

Stable subject: keep the face, clothing, background layout, and main colors close to the source image.  
Camera: slow push-in, no rotation.  
Duration: short clip, calm motion.  
Ending frame: subject still centered, background not rearranged.  
Avoid: extra limbs, sudden zooms, text artifacts, heavy scene changes.

The habit is useful because it keeps the review step simple. If the subject changes, I fix the stability line. If the clip feels messy, I fix the camera line. If the ending frame drifts, I make the final-frame instruction more specific.