← Tutti i casiDrinks & Products · 30 min

Spring Matcha: A 15-Second Motion Visual

Use AI stills, three shots and simple motion to finish a playable portrait-format piece.

Cosa realizzerai

One 15-second silent MP4, a shot list and a reusable editing prompt.

Ti servirà

  • The theme base image from this series, or your own image
  • An editing tool with keyframes and a timeline, such as CapCut

Passaggi

  1. 00–05 s: Show the whole scene

    Show the matcha milk in a clear glass with a slight push-in to establish the overall scene. Don't add rotation, flashes and complex transitions all at once.

  2. 05–10 s: Move in on detail

    Crop the same base image into a close-up and use one slow pan to emphasize texture, color or light. This movement comes from editing keyframes; it doesn't mean the objects in the image actually move.

  3. 10–15 s: Return to the theme

    Return to a fuller view and hold on one steady final frame. You can add “First Sip of Spring” at the end, but the demo clip stays text-free so you can adapt it.

  4. Check, then export

    Export the portrait practice video at 720×900, 24fps. The current demo has no audio; if you need music, use audio you have the appropriate rights to, and check rhythm and volume before publishing.

Errori comuni

  • Mistaking image push/pull for object motion — This exercise practices editing and shot pacing. For real movement, you'll need additional video generation or filmed footage.
  • Shots are too fast and the image edges show — Shorten pan distances, slow down keyframes, and check the frame edges shot by shot before exporting.

Modello di esercizio

Design a 15-second motion poster for {subject}. Color palette: {palette}; mood: {style}; setting: {scene}. Use only one still image and complete it with editing keyframes: 0–5 s, a gentle push-in on the wide shot; 5–10 s, a detail pan; 10–15 s, return to the theme and fade out. Output per-shot crop, keyframe and duration suggestions, and make clear this is still-image editing that doesn't assume the subject itself moves.

This clip is an edited demo of GPT still images with zoom, pan and fades, not generated by a video model; no audio track.