Models
Seedance 2.0: plan motion from a product photo
Use a product photo to plan a Seedance 2.0 shot in Muvi. Separate input roles, write one motion cue, and review shape, labels and framing before your next edit.
Sources checked:
This guide is available in English.
Decide what the reference must preserve
A product photo gives a video a clear starting composition. Ask for a large camera move, though, and the shot may reveal surfaces your photo never showed.
Use this workflow to brief a small movement while keeping a visible checklist of product details. ByteDance describes Seedance 2.0 as accepting text, image, audio and video inputs; the controls exposed by a particular Muvi mode determine which of those you can use here. See the official Seedance 2.0 overview.
Use the image as a starting frame
Open Seedance 2.0 image-to-video for this exercise. Check the selected model and mode, then upload a clear product photo you may use. A starting image defines the opening composition; an additional reference is a different input role and should only be used where the selected mode exposes it.
- Keep the full silhouette and any details you need to inspect inside the crop.
- Use the image types, dimensions and file-size limits displayed by the upload control.
- Choose a duration and aspect ratio offered by that exact model. A provider's maximum is not a promise for every Muvi selection.
- If you have no photo and want a concept scene, use Seedance 2.0 text-to-video. Do not expect it to reconstruct an exact product from its name.
Request a motion you can isolate
This is an example prompt, not a generated sample or evidence of output quality:
Use the uploaded blue ceramic mug as the starting composition. The mug stays on the table, with its handle visible on the right. The camera makes a slow, small push-in and stays at the same height. Soft window light, plain background, one continuous shot. Keep the mug's silhouette and blue glaze consistent. No new objects or text.The intended result is a closer view with the same mug and framing direction. If the source contains a label, add its exact appearance to your review checklist rather than assuming the prompt can guarantee its spelling. For a second version, keep the camera still and request a gentle lighting change; do not combine both experiments at once.
Compare the clip with the source photo
- Check the opening frame against the photo: handle position, rim shape, glaze colour and crop.
- Pause in the middle and at the end. Look for a handle that thickens, a rim that bends or a new object entering the background.
- Check whether the camera moved closer without circling the mug. A new view may invent unseen product geometry.
- Reject a version that changes the product, even if its movement looks appealing. Reduce the requested motion and keep the original photo for comparison.
Separate shots do not guarantee identical appearance. Keep a record of the source image, prompt and selected settings when assembling multiple clips. Check a pending job's status before starting another request.
Questions about inputs and limits
Can I add an audio or video reference in this image mode?+
Only if that mode's visible controls support the input. The provider's multimodal capability does not mean every Muvi mode accepts the same attachments. This exercise uses one starting image.
Will this preserve a logo or label exactly?+
That requires checking the output. Compare small text and edges with the source photo, and add approved graphic elements in your editing tool when exact reproduction matters.
Build the first shot
Start with the Seedance 2.0 image workflow. Once the product survives the frame review, use the product-video shot plan to decide how the clip fits an edit.
Sources checked 8 September 2026: ByteDance's Seedance 2.0 overview for input modalities; the linked Muvi Studio selections for the text/image paths. Examples are editorial starting points, not measured comparisons with other models.
Seedance 2.0: plan motion from a product photo
Use a product photo to plan a Seedance 2.0 shot in Muvi. Separate input roles, write one motion cue, and review shape, labels and framing before your next edit.