Barn doors interpolate hung hardware to rolling closed
Frame A is the strap hinge hung, frame B is the same opening rolled shut: Flux 3 first-last 5–20s at 9cr; a generated barn invents different hardware than the catalog track.
Frame A is the strap hinge hung, frame B is the same opening rolled shut: Flux 3 first-last 5–20s at 9cr; a generated barn invents different hardware than the catalog track.
Barn-door content that books is hardware travel on this opening: strap hinge or hanger, the catalog track, the same jamb, open then shut. The buyer is matching roller count, strap profile, the face of the slab they already quoted. Text-to-video will invent a cousin barn. Ordinary image-to-video will start on your still and improvise a different track by second four. First-last has to land. You already own both stills. Interpolate.
Frame A hung; frame B rolled shut
Photograph the unit on the opening it ships for — same lens height, same casing, same light family. Strap hinge hung, slab stacked open or parked on the track, hardware readable. That is the first frame. The same rectangle, same track, slab rolled shut across the opening is the last frame. Composite the closed state only if the geometry is still this opening. A different doorway is a morph.
First-last frame takes those two stills and generates the motion between them. Ordinary image-to-video knows where it begins and improvises the rest. First-last has to land. The arc of the shot is decided by you.
The two frames have to be plausibly connectable in the time available. Same opening, same camera, same lighting neighbourhood. A wide of a barn morphing into a macro of a strap is a warp, not a close. Keep the rectangle.
A loop (the same still in both slots) is ambient motion on one plate. This offer needs two states. Do not use a last frame that is a different SKU. Flat-track sealed into a box-rail close is a morph of the hardware. The model will connect them. The connection is the artefact every installer notices.
Flux 3 interpolates the travel
Create a first-and-last-frame transition video is the named agent job: two photos in, one transition out. The agent passes them as first and last frame and picks a first-last-capable model. Cost is priced per model, shown before you confirm. This is not the plain “turn one photo into a video” capability.
Flux 3 First Last Frame to Video is the catalog row for that interpolation: start frame and end frame, native audio, 720p or 1080p, durations from 5 seconds through 20. The listed rate is 9 credits per second. Audio is on — write the mix (roller travel, a quiet pad) or write silence. A silent brief still gets a soundtrack you did not choose.
Prompt the travel, not the objects. The stills already contain the strap, the track, and the opening. Name the mechanism: hold, then the slab rolls shut; camera locked; no new house, no new hanger. If the endpoints share geometry, the morph reads as a close. If they do not, you get a mushy crossfade you could have made in an editor without a video bill.
Start at 5s or 6s. A 20-second miss is a long miss of the same warp. Twenty seconds is the ceiling on this row. Do not replace it with a 30s atmosphere clip because you wanted “more rustic.” The product is the close. The close is in frame B.
A generated barn invents hardware
Text-to-video has no plate. “Rustic barn door sliding shut on a timber cabin at golden hour” decides whose track, whose strap, whose slab. Average hardware is not this catalog row.
Image-to-video from one open still will invent the close. Barn OEM is the shut face on that opening, on this track. Pin the closed still as the last frame or you are not selling the travel. One image is open-ended motion. Two images is a bounded landing.
Do not use a last frame from a competitor’s track “because the close looks cleaner.” Do not prompt “heavier strap” or “more rollers” while interpolating — that is a new SKU. Do not start from a paragraph about “a barn door rolling closed on a farmhouse.” That rebuilds the hardware.
A generated barn is a different building. It will grow a cousin hanger and a strap count that was not on the sell sheet. Buyers who already quoted the track notice that before they notice your logo. If you would not put either JPEG on the bid, you are not ready to buy the interpolation. Native audio will add a bed. Do not write a voiceover into the interpolation prompt. Spec as type you typed, after lock, if a series string must appear.
What this is not
It is not a talking installer. It is not a glass-crew film. It is not a single-image wiggle that leaves the slab sitting open for eight seconds of dust.
Garage door companies are the nearby installer page: emergency springs, same-day open, curb-appeal upgrade on a house like the buyer’s. A barn-door crew that copies that playbook is filming the call they actually take — a house, a truck, a before and after of this opening. That is not the OEM catalog loop.
OEM motion starts from the unit stills. A generated “farmhouse at golden hour” that almost matches the catalog is a different SKU: wrong hanger, wrong rail. Two stills of the same rectangle. Interpolate.
The test: pause the last frame. Is that the track you sell, slab shut, strap still the strap? Yes: this job. No, a roller appeared or vanished: you did not own both photos.
FAQ
Can I animate one photo of the hung door and hope it rolls shut?
You can. The ending is then the model’s guess. Barn OEM is the closed face on that opening, on this track. Pin the closed still as the last frame or you are not selling the travel.
Why Flux 3 instead of a talking-head model?
Because the product is interpolation between two approved stills, not a mouth. Flux 3 first-last is 5–20 seconds, 720p/1080p, native audio, 9 credits per second. A presenter row will talk over an opening it invented.
Do the two photos have to be shot on the same day?
They have to share geometry. Same opening, same lens family, same lighting neighbourhood. A noon still into a night still is a time-lapse you should prompt as one. A different house is a morph.
Is this the same as uploading one image to image-to-video?
No. One image is open-ended motion. Two images is a bounded landing. The agent notes that this capability needs first frame and last frame. Missing the closed slab is a different job.