How to Animate an Old Family Photo With AI

Updated November 2, 2026

A still family portrait can become a short video with a blink, a subtle smile, or a slight turn of the head. Start with a clear scan and request restrained movement; large gestures are more likely to change recognizable facial features.

Prepare the photograph

Scan the original or photograph it in even light, avoiding glare and deep shadows. Use the clearest available version and keep an untouched copy. Crop out distracting borders, but leave room around the head and shoulders so small movements are not cut off.

Choose an aspect ratio that suits the destination: 9:16 for a vertical story, 1:1 for a square post, or 16:9 for a widescreen family video. A close crop may remove people or details, so check the framing before generating.

Animate the portrait step by step

1. Open Oranzy and sign up for 100 bonus tokens. Choose the image-to-video workflow and upload your prepared portrait.
2. Select a video model and enter a motion prompt. Try: “The person blinks once, breathes gently, and gives a subtle smile. Keep their facial features, hairstyle, clothing, and the background consistent. Fixed camera, natural movement.” For a group photo, ask for small movements from each person without changing their positions.
3. Set the aspect ratio and, if a duration option is available, start with a short clip of around 5 seconds. Generate a first version. Keep the camera fixed and avoid adding several actions at once.
4. Watch the result at full size, especially the eyes, mouth, and edges of the face. Generate one comparison with either a revised prompt or a different model, then download the version that best preserves the people in the photograph.

For more ways to describe camera and subject movement, see the video prompt-writing guide.

Compare and refine the results

Change one thing per test so you can tell what helped. For example, keep the same image and prompt while comparing two models, or keep the model and replace “turns toward the camera” with “blinks gently.” Check both the first and final frames for changes in facial shape, clothing, and background objects.

If the selected clip looks soft at its intended display size, review the options in the AI video upscaling guide. Upscaling may improve presentation, but it cannot reliably recover a face that the animation has already altered.

Avoid common mistakes and wasted generations

• Overloading the prompt: Begin with one or two small actions. Running, waving, and dramatic camera moves can distort a still portrait.
• Using a damaged or blurry source: A clearer scan gives the model more visible detail to work with. Review the image before spending tokens on repeated tests.
• Ignoring the crop: Check that all faces fit in the selected aspect ratio before generation, especially in group portraits.
• Changing everything at once: Compare one variable at a time and save the prompt and model used for each promising result.
• Treating animation as historical evidence: The movement is generated, not recovered from the original moment. Label the clip accordingly when sharing it, and consider the privacy and wishes of the people pictured and their families.