Dryer sheets: Ideogram the box face before any tumble motion
The scent name is the product. Type-lock the carton, then maybe I2V.
The scent name is the product. Type-lock the carton, then maybe I2V.
A dryer-sheet listing is a box face: the scent line, the sheet count, the brand lockup, the color that already lives on the shelf. It is not a tumble of anonymous fabric, a generated laundry room, or a talking parent folding towels. If the letters on the carton are wrong, the motion is a commercial for a cousin SKU. Get the type right in a still. Spend on video only after the face would survive a freeze-frame.
The scent line is the SKU
Shoppers do not buy "dryer sheets." They buy the named scent and the count on the carton they already reach for. A video model that redraws "Spring Linen" as "Fresh Meadow" sold a different box. A video model that melts the count into a decorative numeral sold a different pack size. Identity lives in letters. Motion is optional.
Photograph the carton square-on, or generate the face only when you do not have a pack shot and the job is a poster-accurate still. Do not start in a dryer drum. The drum has no type. Once the camera is inside the tumble, the model has to invent the box that never entered frame, and it will invent a plausible cousin.
Ecommerce brands is a nearby catalog audience. That page is a general product loop: still, then motion, then a shoppable clip. This page is narrower. The dryer-sheet job is the carton face first, because the scent string is the differentiator. A generic "laundry product orbit" is the ecommerce playbook used on the wrong object. Keep the box large. Keep the scent large. Crop the utility room before you buy seconds.
Do not generate a person inhaling the dryer. Do not generate a child in a basket. The object is the carton. A lifestyle plate is a second still you shot, or it is not in the ad.
Ideogram owns the carton type
Ideogram V4 is the text-to-image row for this face: 6 credits, posters and pack shots, accurate letter rendering, 1:1 / 16:9 / 9:16 / 4:3 / 3:4. It is an image model. It does not tumble. It does not speak. That is the compliment.
Prompt the lockup you would print. Name the scent exactly. Name the count exactly. Name the carton color family you already ship. Ask for a square-on pack shot, sharp type, no extra badges, no invented "new" flashes unless legal already owns them. Generate an image from text is the named still job. Approve the letters at 100% zoom. If a character is wrong, you are still on an image meter. That is cheap compared with discovering the typo in a ten-second tumble.
A video row that "also does text" is not this job. Video will move the box and renegotiate the type. The scent line will drift a letter, then you have a commercial for a scent you do not sell. This page is Ideogram because the carton is a poster with a SKU on it. Do not Omni a new laundry film around a box you have not locked.
If you already have a photograph of the shipper carton, skip the generate. Crop it. That file is the face. Ideogram is for when the photograph does not exist or the type is unreadable. It is not a style filter you run on a good pack shot for luck.
Motion is optional and later
Once the face is signed off, image-to-video is allowed to do a short, boring thing: a slow push on the carton, a slight parallax on the sleeve, a hold on the scent line. Prompt camera and physics. Do not re-describe the scent in the video prompt. Re-describing is how the model starts drawing a second carton.
A tumble clip is a different claim. Fabric in a drum does not prove the box. If you need both, they are two shots: locked face, then a real tumble you photographed, cut together. Generating "sheets tumbling, scent blooming, box on the lid" in one pass is how the lid becomes a different brand. The AI product video generator is the product-motion door after identity holds. It is not a substitute for the still.
Keep the clip short enough that the type does not have time to crawl. A long laundry mood piece spends the video meter on steam instead of on the letters you paid Ideogram to hold. If the feed is muted, burn the scent as overlay; do not ask the video model to restamp the box.
What this is not
It is not a talking laundry expert. It is not a text-to-video "fresh dryer sheets in a sunny laundry room" paragraph that rebuilds the carton. It is not a single-image wiggle of a blurry phone snap you hoped would sharpen in motion.
Do not I2V a box that is still wrong. Motion will not correct a bad letter. It will animate the error.
Do not treat a dryer-interior generate as the hero. The hero is the face a shopper would recognize in a freeze-frame. If that freeze-frame would not pass legal, the video will not pass the shelf.
Pause on the scent line. If you cannot read it, you are not done. Type-lock. Then maybe I2V. Tumble last, and only from a plate that already tells the truth.
FAQ
Can I skip Ideogram and animate a phone snap of the box?
If the snap already shows the real scent and count, sharp, square-on, yes — that snap is the still. Image-to-video from a readable pack shot is the later job. A blurry snap will not grow a legal lockup in motion.
Why not start in the drum and put the box in later?
Because the drum has no type, and the model will invent a carton when you ask for one. The scent name is the product. Lock it in a still first, or you are advertising a cousin.
Do I need a tumble clip at all?
Only if the brief is fabric motion. The listing usually needs the box face. A short push on a type-locked carton is enough for a feed ad. A generated tumble is a second claim.
What if the carton has a window showing the sheets?
Photograph or Ideogram the window as it actually prints. Do not ask a video model to "reveal the sheets" through the window. That is a new interior, and it will not match the stack you ship.