Motocross: the bike in the whoops has to be the rider's bike
Lock number and livery. Veo first-last empty jump to a consented bike still. A generated wreck of a machine you do not own is a false ad.
Lock number and livery. Veo first-last empty jump to a consented bike still. A generated wreck of a machine you do not own is a false ad.
A motocross clip that sells a rider or a track is this bike: number plate, plastics, the livery you run. Interpolation is allowed. Inventing a cousin machine — or a wreck of someone else’s — is a claim you cannot point at a photograph for.
Number and livery are the bike
Sponsors and stewards read the same pixels: plate digits, colourway, the graphics kit in the pits. Text-to-video of “a motocross bike blasting through whoops” invents a number and a plastic set. That is a different bike.
Photograph the machine you have a grant for. Square-on or three-quarter in the pits, number sharp, livery uncropped. Do not prompt “clean up the graphics.” Erasing a tear redraws the plastic behind it. Treat another rider’s livery the way you treat another brand’s blister: leave it off the prompt.
If the only photo includes a rider, you need a release that covers generation, or you crop to the bike. “Anonymous rider, no face” is still a body. Empty bike still, or a consented still. Those are the two honest plates.
Empty jump to a consented still
The ad is a state change: empty face of the whoops or the jump lip, then the consented bike in that dirt. Two stills of one line.
Photograph the empty jump or whoops — same lens height, same berm, same light family. That is frame A. Photograph the rider’s bike on that line, or a pit still that shares geometry with the landing. That is frame B. A different track 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 two frames have to be connectable. Same face, same camera, same lighting neighbourhood. A pit close-up of a number plate morphing into a wide empty jump is a warp, not a pass. A loop (the same still in both slots) is ambient motion on one plate. This offer needs two states. Empty-to-empty is not the rider’s bike. A different number in B is a different machine.
Veo first-last is the interpolation
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.
VEO First Last Frame is the catalog row: first and last frame, native audio, official 4K tier, 4s / 6s / 8s, listed 40 credits. Image required. Audio is on — write a pack rumble, dirt, no speech, or silence. A silent brief still gets a soundtrack you did not choose.
Prompt the travel, not the objects. Name the mechanism: hold, then the machine enters the whoops; camera locked; no new number, no crash. If the endpoints share geometry, the morph reads as a pass. Do not use a last frame that is a stock bike “because the jump looks cleaner.” The connection is a number that changes mid-air.
Image-to-video from one photo is open-ended motion. Fine for a slow push across a locked pit still. It will not land on the empty jump unless that jump is a last frame you attached. One still of the whoops will invent a bike. That invented bike is not the rider’s. Start at 4s until the number holds. 4K does not fix a wrong plate.
A wreck you do not own is a false ad
Veo 3.1 is native audio, official 4K tier, 4 / 6 / 8 second cinematic clips. It is a bad way to mint a crash. A paragraph that says “motocross wreck in the whoops, rider down” will still draw a machine and still grow a body. Those plastics will not be the livery you run. That crash did not happen. Publishing it as this rider or this track is a false ad.
Do not text-to-video a crash of a bike you do not own, first-last a pit still into a generated wreck “for drama,” or swap another team’s number into frame B. Worn is a different machine if you did not photograph the wear. If a real crash still exists, it is a photograph you have rights to, or it is not in the ad.
Inspect the midpoint. Ends are held by construction. If the middle grows a second number, a cousin fork, or a body that was not in either still, reroll. Do not regenerate both stills “to match the motion.” Pause the last frame. Is that the rider’s bike — number, livery, the machine in the pits? Yes: this job. No: you asked the model to invent a crash. Do not write a commentator into the interpolation prompt. Lips do not belong on this dirt.
FAQ
Can I animate one photo of the empty whoops and hope the rider’s bike appears?
You can. The ending is then the model’s guess. The product is this number and this livery. Pin the consented bike still as the last frame or you are not selling that machine.
Why Veo first-last instead of Veo 3.1 text-to-video?
Because the product is interpolation between two approved stills, not a paragraph about a jump. VEO First Last Frame is 4 / 6 / 8 seconds, 4K, 40 credits, native audio, image required. Veo 3.1 text-to-video will invent a bike and, if you ask for heat, a wreck you do not own.
Do the two photos have to be shot on the same lap?
They have to share geometry. Same face of the whoops or the same jump, same lens family, same lighting neighbourhood. A pit still into a night race still is a time-lapse you should prompt as one. A different track is a morph. A different number in frame B is a different bike.
Is this the same as uploading one bike photo 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 consented bike — or missing the empty jump — is a different job.