How to Turn Product Photos Into a Short Video Without Letting AI Redraw the SKU

If you already have accurate product photos, you do not always need an AI video model to make them move. For a short social advert, a simple pan, zoom and text sequence can create enough motion to hold attention while keeping the photographed SKU exactly as it appeared in the approved still.

This is useful for jewellery, packaging and other products where generative image-to-video can create a new problem: the bracelet looks correct in the first frame, then a clasp, engraving, stone or label changes as the model invents the movement.

Adobe Express currently lets users animate photos with effects such as Zoom and Pan, add animated text and export the finished result as MP4. Adobe also says its free plan includes core photo-editing tools, effects and free templates. The workflow below uses that ordinary image animation rather than asking a generative model to reconstruct the product.

Step 1: choose three approved product photos

Do not begin by generating more assets. Start with photographs or AI-assisted stills that have already passed your product-accuracy check.

A simple three-shot set works well: one hero image that identifies the product immediately, one detail image that shows a selling point, and one lifestyle or scale image that gives context. A jewellery seller might use a clean bracelet hero, a close-up of the engraving and a worn wrist image. A skincare seller could use the bottle front, a texture or pump detail and a bathroom lifestyle image.

Keep the untouched masters separately. The video project is another derivative, not the new source of truth for the SKU.

Step 2: build the video in the format you actually need

For a full-screen TikTok-style placement, start with a vertical 9:16 canvas rather than making a horizontal video and cropping it later. TikTok's current ad specifications list 9:16 as the video ratio and require more than 540 by 960 pixels.

Import the first approved image and position it deliberately. Do not stretch it to fill the frame. If the photograph is wider than the canvas, choose the crop that keeps the product complete and leaves space for any headline or call to action.

If important jewellery, packaging text or product hardware sits close to an edge, move the image rather than cropping through it. The video format should adapt around the product, not the other way around.

Step 3: use restrained pan and zoom instead of synthetic motion

Select the photo and open the animation controls. Adobe Express currently provides photo-motion options including Zoom and Pan, with controls for the movement direction or starting point.

For product imagery, subtle movement is usually the safest choice. A slow push toward an engraving, a gentle pan across a necklace or a small zoom out from a gift set can make a static asset feel like video without creating an unseen side of the object.

The important distinction is that ordinary pan and zoom animation moves or scales the existing image layer. It does not need a generative model to invent a new back view, changing clasp or moving gemstone for every frame. That makes it a useful option when preserving the exact photographed product matters more than creating complex motion.

Step 4: turn the three stills into three clear beats

Use the hero image first and let the motion draw attention toward the product. Then switch to the detail image and make the movement support whatever the customer should notice. Finish with the lifestyle or scale image and the action you want the viewer to take.

For an engraved bracelet, the sequence could be: hero bracelet with a slow push-in, engraving close-up with a gentle sideways pan, then the bracelet on a wrist with a short line such as “Personalise yours.” The photographs are doing the factual product work. The animation is only controlling attention.

Avoid making every image spin, bounce or move in a different direction. Product videos often feel more premium when the motion language is consistent and restrained.

Step 5: add text as an editable layer, not inside the product image

Adobe Express also lets you animate text and other elements independently. Use that for the headline, feature callout and call to action rather than asking an image generator to bake words into the product photograph.

Editable text is easier to proofread, reposition and reuse. It also avoids a common generative-image failure where the scene looks strong but the brand name, engraving or feature text contains a wrong character.

Keep the copy short enough to read on a phone. The product should still be understandable when the viewer ignores the text entirely.

Step 6: export as MP4 and run one final product check

Adobe says animated-photo projects should be downloaded as MP4 to preserve the animation. Before exporting, watch the video once for pacing and then review it again specifically for product presentation.

Check that no zoom makes the source image visibly soft, no crop removes a clasp or label, no text covers a critical detail, and the apparent product size does not become misleading. For jewellery, make sure the engraving, chain, stones and finish remain readable enough in the frames where they matter.

The advantage of this workflow is that there should be no frame-to-frame product drift caused by generative motion, because the same approved still is being moved on the canvas. You can therefore spend your review time on crop, scale, readability and sequence rather than checking whether AI redesigned the SKU halfway through the clip.

When this is better than AI image-to-video

Use this method when the job is presentation: add movement to a product hero, create a social advert from existing catalogue images, show several details in sequence, or test a new hook without another shoot.

Use real video when the selling point depends on physical behaviour. If the customer needs to see a clasp opening, a bracelet flexing, a lid turning, fabric moving or a mechanism working, film the real product. A pan across a still cannot prove those actions.

Generative image-to-video becomes more useful when you genuinely need movement that did not exist in the source. It can create a camera move, model motion or changing environment that simple 2D animation cannot. The trade-off is that the product must be checked across frames because the model is synthesising new visual information.

Where Lustra Studio fits

If you use Lustra Studio to create jewellery scenes, model images or consistent colour variants, approve those stills first. Then use ordinary animation for campaigns where you only need motion and pacing. This keeps the product-generation stage and the video-editing stage separate, which makes it easier to know where an error was introduced.

The takeaway

A product video does not always need generated movement. If you already have accurate stills, a simple vertical sequence with pan, zoom and editable text can create a useful social asset while preserving the approved product pixels.

Start with three strong images, animate the camera attention rather than the SKU, keep text separate, export to MP4 and use generative video only when the creative actually requires new motion. Sometimes the safest AI-era workflow is knowing which part of the job does not need generation at all.

Sources: Adobe Express, “Add Photo Animations to Images for Free,” reviewed 12 September 2026; Adobe Express animated-image guidance, reviewed 12 September 2026; TikTok Ads creative specifications, reviewed 12 September 2026.