"Photo to video" covers several quite different techniques. One moves slowly across a photo. Another makes the person in it blink and turn. The results differ a great deal, and so do the risks. It helps to know the difference before choosing what suits your photos.
Three ways to make a photo move
1. Pan and zoom
The photo stays as it is; the frame slowly pushes in, pulls out or slides across. This is the staple of growth and memorial videos. The content of the photo is kept completely and nobody's face is altered. Quality comes down to speed, direction and how long each photo stays; see making a photo video that doesn't look like a slideshow.
2. Parallax (2.5D)
The person and the background are separated into layers that move at different speeds, giving a sense of depth. The person isn't distorted, only set apart from the background. It suits photos with a clear subject and a simple background, takes longer than pan and zoom, and is usually kept for the few most important photos.
3. AI-generated motion
AI imagines what happens next: hair moving, water flowing, a smile or a turn of the head. This is photo-to-video in the full sense and the most striking to watch. One thing to understand: the movement is generated. It wasn't in the photo.
Photos that suit AI motion
- Landscapes: clouds, waves, leaves, waterfalls, mist. Usually natural.
- Products and food: steam, a sweep of light, a slight camera move. Shapes and logos can still change, so each result needs checking; see using real products and logos in AI video.
- Illustrations and drawings: not realistic to begin with, so change is easier to accept.
- People at a distance or from behind: where the face isn't the point.
Portraits need particular care
When AI animates a person, these can happen:
- The expression stops looking like them; the smile is someone else's smile.
- Features shift slightly in motion, especially eyes and teeth.
- Warping where hands, hair and background meet.
- With low-resolution old photos, AI fills in detail that was never there, and not always correctly.
With strangers this goes unnoticed. For family, a father's smile that is slightly off is no longer his. For weddings, memorials and tributes to older relatives, we usually suggest pan and zoom, with parallax on a few photos at most, keeping the photos as they are. See preparing a memorial video.
Using someone else's photo also needs their consent; see using a real person's face in AI video.
Which to choose
- Family and friends, as a keepsake: pan and zoom, with parallax on key photos.
- Brand, product or scenery, for marketing: AI motion is worth considering, checked result by result.
- Few photos, a whole video to fill: mix all three, with text and AI-generated scenes.
Damaged or faded old photos are a separate matter of restoration, assessed through our repair page.
Before you hand it over
Send the original photo files (not screenshots, and not copies compressed by a chat app), and tell us what the video is for, which photos matter most and how much movement you'd like. If you aren't sure which method fits, send the photos and it can be discussed. See how to provide footage.
FAQ
Phone apps can animate photos. What does commissioning add?
Apps are fine for a fun effect on one or two photos. Making dozens of photos into a paced video with music and titles is about selection and editing, not just movement.
The AI result doesn't look like the person. Can it be changed?
It can be regenerated, but each attempt differs and a true likeness can't be promised. That's why we are cautious with portraits.
Can low-resolution photos be used?
Yes, shown at a suitable size. Enlarged too far they blur.
