Racetrack ads lock the empty oval, then add motion; do not invent the field

First-last empty dirt to a consented race still. Text-to-video draws horses that are not entered.

A racetrack ad that books is the oval you operate, filling with a field you are allowed to show. It is not a paragraph that summons a pack from the average of every race clip a model has seen. Those horses are not in your program. Lock two stills. Interpolate. Do not invent the field.

Frame A is empty dirt; frame B is a consented race

Photograph the oval empty — same lens height, same rail, same infield board, same grandstand. That is the first frame. Photograph a race you have the right to show, or composite a still of this program's horses on that same geometry. That is the last frame. Numbers, silks, and faces in frame B have to be entered, consented, or yours. A stock gallop from another meet is a false card.

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 track, same lighting family, same camera. An empty dirt close-up morphing into a wide night card at a different plant is a warp, not a post time. Keep the oval.

This is the racetrack, not a riding lesson and not a generic hall. Horse riding schools sell an arena and a lesson. Event venues sell a room. This page is the dirt, the rail, the program. Mixing them is how a meet becomes a stock horse commercial.

Veo first-last interpolates; it does not draft the card

VEO First Last Frame is the catalog row: start frame and end frame, native audio, official 4K tier, 4 / 6 / 8 second clips, 40 credits. Audio is on — write the mix (a distant call, a quiet rail) or write silence. A silent brief still gets a soundtrack you did not choose. Do not let that soundtrack name horses that are not in frame B.

Prompt the travel, not the animals. The stills already contain the empty dirt and the consented field. Name the mechanism: hold on the empty chute, then the field enters frame; camera locked; no new grandstand. If the endpoints share geometry, the morph reads as post time. If they do not, you get a mushy crossfade you could have made in an editor without a video bill.

Create a first-and-last-frame transition 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. Missing the empty oval is a different job. Missing the consented last frame is how you interpolate toward a pack you did not enter.

Do not use a last frame from another track. The model will connect them. The connection is the artefact everyone who actually goes to your oval will notice.

Do not text-to-video the field; extend a keeper

Text-to-video has no plate. "Horses breaking from the gate at golden hour" decides whose horses. It will pick numbers, silks, and a rail from training data. That is a false program. Empty-to-field only works when both ends are true.

A loop (same still in both slots) is ambient motion on one plate — dust, a flag, a hold on empty dirt. That is a teaser, not a race. This offer needs two states.

When the 8-second landing is the landing you wanted, do not regenerate a new field to make it longer. Extend a video continues from where the original ends. You describe the continuation and pick a model. Credits are spent on the extension, not on throwing away a keeper to roll a new clip. Extending past the consented last frame is how new horses walk in. Prompt hold, dust, a pull off the rail — not "more of the pack."

If you needed a human line about post time, that is a different generate, or overlay, cut after the field is in. The oval shot does not have to talk. A constructed caller naming odds is a finance-adjacent mouth you do not need on a meet promo.

The test

Do you own an empty still of this oval, and a race still you are allowed to show, sharing geometry?

Yes: first-last. Veo 4 / 6 / 8s, 4K, 40 credits, native audio. Then extend the keeper if you need more seconds.

No last frame: you do not have a race. Shoot one, or hold on empty dirt as a loop. Do not ask text-to-video to invent a card.

No empty first frame: you are animating a race you already have. That is ordinary I2V, not this interpolation.

FAQ

Can I animate one photo of the empty oval and hope a field appears?

You can. The ending is then the model's guess. Racetrack ads that name a meet need the consented field as the last frame, or you are not selling that card.

Why Veo first-last instead of a talking-head host?

Because the product is interpolation between two approved stills, not a mouth. Veo first-last is 4 / 6 / 8 seconds, 4K, native audio, 40 credits. A presenter row will talk over an oval it invented.

Do the two photos have to be shot on the same day?

They have to share geometry. Same rail, same lens family, same lighting neighbourhood. A noon empty into a night card is a time-lapse you should prompt as one. A different plant is a morph.

If the 8-second clip is short, should I generate a new first-last at a longer duration?

Veo first-last tops out at 8 seconds. Extend continues the keeper. A new generate resets identity and can draft horses that were never in frame B.