In short: shoot 4–6 photos in soft daylight, cut the product out with Any subject, put it on a brand colour, add gentle 3D motion, a 3-sentence voiceover and three words of text per shot, and sync the cuts to music.
The shot list
Buyers want to answer three questions: what is it, how big is it, what's it like to use. Plan one photo for each, plus a hero:
- Hero — the whole product, front three-quarter angle.
- Detail — texture, stitching, the finish.
- Scale — in a hand, next to a familiar object.
- In use — on a desk, being worn, in a kitchen.
- Packaging or variants — optional, great as the last shot.
Shooting on a phone
- Soft light from a window, never direct sun or the phone's flash. A sheet of white paper on the shadow side brightens it.
- Plain, contrasting surface — makes the cut-out cleaner.
- Use the phone's 2× lens if it has one; wide lenses distort products.
- Leave space around the product so you can crop to any shape later.
Cutting out the product
The standard people segmenter doesn't know what a mug is. For photos, VDOAPP offers a second model: choose Any subject under What's the subject? and it uses BRIA's RMBG-1.4, a background-removal model trained on products, animals and objects (about 42 MB, downloaded once).
- Add the photos, select the first one and open AI effects · selected clip.
- Set What's the subject? to Any subject.
- Under Behind the subject, choose Colour with your brand colour, or Image for a lifestyle backdrop. Keep the in-use shot with its real background — context sells.
Licence note: BRIA publishes RMBG-1.4 for non-commercial use, and commercial use needs an agreement with BRIA. If your product video is for a business, check BRIA's licence first — or photograph the product on a plain green or blue sheet and remove it with the chroma key in the PiP tab, which has no such restriction.
Adding motion
Still photos feel like a slideshow; subtle motion makes them feel like video. Use 3D motion (Zoom in or Orbit at 40–50%) on the hero and in-use shots, and a plain Zoom in keyframe preset on cut-out shots. Set each photo to 2–4 seconds.
Voice, text and music
- Voiceover: write three sentences — what it is, why it's better, what to do next — and generate it with text to speech. Heart or Michael suit most products.
- Text: three to five words per shot: "Hand-thrown stoneware", "Holds 350 ml", "Dishwasher safe". Use the Box style for legibility.
- Captions: caption the voiceover too; most feeds play muted.
- Music: an upbeat instrumental, low under the voice, with beat sync on Every 2 beats.
A 20-second recipe
| Time | Shot | Treatment | Text |
|---|---|---|---|
| 0–3 s | Hero | Cut-out on brand colour, 3D zoom | Product name |
| 3–7 s | Detail | Zoom in keyframes | Key material or feature |
| 7–11 s | Scale | Original background | Size or capacity |
| 11–16 s | In use | 3D orbit | The benefit |
| 16–20 s | Hero again | Cut-out, neon outline | Price or "Shop now" |
Export at 1:1 or 4:5 for marketplaces and feeds, and 9:16 for Stories and Reels — change the canvas and export again; it takes a minute.
Frequently asked questions
Can I cut out a product in a video clip?
Not with Any subject — that model is for photos. For product video, film on a plain background and use the chroma key in the PiP tab with a green or blue backdrop.
Can I use the AI voice in an ad?
The Kokoro voice model is Apache 2.0 licensed, which allows commercial use. Check the rules of the platform you advertise on.
Will marketplaces accept the video?
Most accept MP4 at 1080p. Check each marketplace's length and size limits before exporting.