Photo Motion LabAnimate a photo, or remake it
Get AI Video

Make a still photograph move, or remake it entirely

The expression actions: the five that keep the most of your photo

Waving, laughing, surprised, cheering and a peace sign ask the model to invent the least, which is why they are the safest first thing to try and often the best result you will get.

By The Photo Motion Lab team · 8 October 2026 · 5 min read

Preview still for the Waving action, as shipped inside AI Video.

Most of what people ask an animated photo to do is the hardest version of the job: a dance, a sport, a scene with a dinosaur in it. The five actions in the expressions category ask for almost nothing by comparison, a hand that moves, a face that opens into a laugh, a shoulder that lifts. That smallness is the whole appeal. The less a model has to invent, the closer the result stays to the actual person in your photograph, and these five invent less than almost anything else in the library.

Why small wins

Every action in the catalogue is the same basic request: start from this still frame and generate a few seconds of plausible motion around it. The more that motion has to cover, legs that were never in frame, a background that has to rebuild itself as a body turns through it, the more room there is for the result to drift from the photograph you gave it. Waving, Laughing, Surprised, Peace Sign and Cheering all stay almost entirely in the face and the upper body, which means almost everything the model needs was already visible in your source photo. It is not inventing a scene. It is extending one.

This is also why these five are the best place to start if you have never generated a clip before. A disappointing result from a dance action can come from a dozen different places, the crop, the floor, the clothing, the background. A disappointing result from Waving almost always comes from one thing: the face. That makes it a far easier action to learn from.

What each one actually does

Waving raises one hand, waves it toward the camera a few times, then settles into a warm smile with a slight head tilt. It is the most forgiving action in the entire library, not just this category, because a hand already visible in the photograph gives the model something real to extend rather than something to invent from nothing.

Laughing opens a quiet smile into a full, sustained laugh, eyes creasing, shoulders lifting, head tipping back slightly before settling again. It asks more of the photograph than Waving does, because the whole arc depends on reading the starting expression correctly, a face that is already smiling has less distance to travel than one caught completely neutral.

Surprised snaps the face open, eyes widening, brows lifting, mouth falling open, one hand flying up to the chin, before softening back into a smile. It is rated slightly harder in the catalogue than the other four, because the expression it asks for is the furthest from a resting face, which gives the model more to invent even within a small frame.

Peace Sign lifts one hand into a peace sign beside the cheek, holds it, then lowers slightly into a grin down the lens. Like Waving, it leans on a hand gesture that is easy to read and easy to hold steady, which keeps the result settled.

Cheering throws both arms overhead, opens the mouth into a cheer, a small bounce onto the toes, then brings the arms back down into a satisfied grin. It is the biggest movement of the five, both arms rather than one, a shift in posture rather than just the face, and it is rated harder for exactly that reason.

The photograph they need

All five want the same basic shape of picture: head and shoulders, close to front on, eyes open and clearly visible, with even light across the face. That is a narrower requirement than it sounds, because it rules out two of the most common photo problems before you even get to the action itself. Heavy overhead light that shadows the eyes, and a face turned much past three-quarters toward profile, both work against every action in this group in the same way, by taking away the one thing the model is reading most closely.

Waving and Peace Sign specifically benefit from a hand already in frame, even resting at the side of the body is enough, because it means the gesture is being extended from something real rather than conjured out of nothing. For Cheering, a little more room around the subject helps, since both arms are about to lift overhead and a tight crop gives them nowhere to go.

None of the five need a particular background. Because the movement stays close to the body, a cluttered room behind the subject is far less of a liability here than it would be for an action where a whole body turns and drags the scene around it. The category note the site already follows puts it plainly: a face fills this kind of clip, so shoot it that way.

Where they genuinely fall short

Two of these, Surprised and Cheering, carry a slightly higher difficulty rating in the catalogue than the other three, and it is worth saying plainly rather than letting a first attempt feel like a fluke. Surprised asks for a bigger jump from a resting expression, and Cheering asks for a second thing to move correctly, the arms, on top of the face. Neither is unreliable, but neither is quite as safe a first attempt as Waving.

More than one person in the frame is the other real limit. Cheering and Laughing can carry two people close together and similarly sized reasonably well, a shared reaction reads naturally as a shared moment. Past two or three faces, the same arithmetic that affects every other action in the library applies here too: each additional face gets a smaller share of the frame and the model’s attention, and a wide group shot that works fine as a still photograph tends to disappoint as a clip.

Said plainly

The model is generating a short, plausible extension of one still frame, not recovering anything that was filmed. These five actions ask it to extend the smallest possible distance from where your photograph already starts, which is exactly why they are the most reliable corner of the whole library.

If you are choosing between a photo for one of these and a photo for something bigger, the general rules in choosing a photo that will animate well cover the rest of the decision. And if the only copy you have of the right face is an old print with real damage, our sister site has a practical guide to restoring an old or damaged photo before you animate it, since these small actions still carry whatever is wrong with the original straight into the result.

Common questions

Which expression action should I try first?
Waving. It is the most forgiving of the five, it reads clearly from a waist-up photo, and a hand already visible in the source picture gives the model a real starting point instead of an invented one.
Why do these look better than a dance or sports clip from the same photo?
Less invention. A small movement in a face the model can already see clearly stays closer to the person you started with. A full-body movement asks it to invent limbs, a floor and often a background at the same time.
Can I use these on an old or faded photograph?
Yes, more reliably than most of the library, because the crop they need is small and the movement is small. A genuinely damaged print still carries its damage into the clip, so repair first if the face itself is affected.
Do any of these work for more than one person in the photo?
Cheering and laughing hold up reasonably well with two people close together and similarly sized. Past that, every extra face is a smaller share of the model's attention, and a wide group shot usually disappoints here the same way it does everywhere else.