Meal prep is the box still plus diegetic fridge sound

Meal prep video is the box still plus diegetic fridge sound. It is not a talking chef and it is not a brand film. Lock the plated photo. Animate the close. Let the fridge door, the plastic lid, and the shelf hum come with the take. Change rows only when the printed label has to survive that close.

A meal-prep service ships a labeled container. Viewers already know the product. What they need from the clip is: this is the box, this is the fridge, this is the sound of a week of food landing on a shelf. That is an image-to-video job with native audio. It is not a script.

The close is a still with a mix

Crop the plated photo so the container fills the frame. If you would not print the still on a menu, do not buy motion.

Vidu Q3 Image to Video is the close: image required, audio on, 5s / 10s / 15s only, qualities 360p through 1080p, listed 16 credits and marked discounted. Features named: image-to-video, motion generation, high fidelity. Write the mix into the prompt — fridge compressor, lid click, shelf slide — because a silent brief still gets a score. Atmosphere lives here. A locked spoken line does not; that is a lipsync row after.

There is no aspect-ratio menu on this slug. Crop 9:16 or 16:9 before upload. The shortest legal clip is 5s. If the close fails in five seconds, you do not have a 3s cheaper retry on Vidu Q3. Fix the still.

The contract on this row is the still, not the slider. Fridge sound will not save a misspelled macronutrient line.

When the label has to hold, leave Q3

Vidu's job is the close with a soundtrack. If the SKU is the printed lid — calories, allergens, the brand mark a customer already holds — you need a row that treats the upload as a plate to extend, not a picture to interpret.

The image-to-video row on Seedance 2.5 is that plate: required still, first frame plus optional last frame, native audio, 4s–30s, 480p / 720p, listed 47 credits, resolution-based per second. Catalog copy: one frame into continuous motion without the drift or stitching of shorter multi-clip workflows. Put the approved lid on frame one. If the shot must land on a square-on label, feed that as the last frame.

720p is the ceiling. A 30s Seedance pass of a wrong lid is a long, expensive wrong lid. Approve the still the way you approve a print proof.

Image-to-video is the surface: upload is frame one; the prompt describes motion (door opens, box slides in, camera holds). Do not re-describe the meal.

Diegetic is not a presenter

A talking founder in front of a fridge is a different catalog row and a different disclosure problem. Meal prep content that works is the object doing object things: lid, shelf, condensation, the compressor in the mix. Native audio on Vidu or Seedance is that bed. It is not VO, and it is not captions.

If a claim has to appear ("12 meals, Sunday drop"), type it after. Timed text overlays burn lines you write, each with its own start and end. Do not ask Q3 to typeset the calorie line onto a moving lid.

Keeping a physical SKU believable is the label-fidelity gate. Meal prep is that gate plus a fridge.

FAQ

Why not generate the box from a prompt?

Because customers already own the box. Text-to-video will invent a lid. Image-to-video starts from the photo you would ship. Vidu Q3 and Seedance I2V both require an image. Use it.

When is Vidu Q3 enough?

When the job is a 5, 10, or 15 second close at up to 1080p with native fridge sound, and the label is either out of reading distance or already correct in the still. If the lid is the product, Seedance is the row.

Can I stitch three 5s Q3 takes into a 15s fridge story?

You can cut. You cannot pretend it is Seedance's single-pass 4s–30s hold. Q3 durations are only 5 / 10 / 15. Write the action to one of those buckets, or change rows.

Does native audio replace a caption?

No. Audio is the mix. Captions and offer lines are editor text. Transcribe speech if someone actually spoke; overlay the drop schedule if nobody did.