MRI ads are the gantry still, not a talking radiologist

Photograph the bore. Image-to-video holds the magnet. A photoreal radiologist giving advice is a YPP kill.

An imaging-center clip that books a scan is the room you operate: the gantry, the table, the magnet already in the suite. It is not a constructed doctor walking a protocol, a contrast dose, or “what this finding means.” Those lines are health advice. A fake face delivering health advice is the YouTube pattern that closes Partner Program eligibility. Buy the facility plate. Do not buy a talking radiologist.

The gantry is the facility plate

A scanner is a SKU with a serial, a bore size, and a room that already exists. The buyer who calls you has been in a waiting room. They know what a magnet looks like when it is real. Text-to-video will invent a prettier suite: a generic tunnel, a cousin table, a logo that is not on your hardware. That is a different site.

Photograph the empty gantry. Same lens height you would use for a facility tour. Lights as they actually sit. No one on the table. No generated patient in the bore, no residual anatomy, no extra in scrubs. The still is the contract. Motion is allowed only after the still is true.

Image-to-video starts from that frame. Frame one is the file. The prompt is camera and physics — a slow push into the bore, a hold on the table edge, a slight parallax on the shroud. Re-describing “a 3T magnet in a cinematic hospital” is how the model starts drawing a second machine.

Do not generate a cleaner cousin of a competitor’s suite and present it as yours. A patient who books a magnet that does not look like the ad has a legitimate complaint. The plate has to be the room they will walk into.

Image-to-video holds the magnet

Seedance 2.5 is the still-to-motion family: four to thirty seconds, native audio, 720p on the image-to-video job. I2V is that job on this family. The published URL is the family page, opened from the image-to-video tool. Thirty seconds in one pass is for a hold that actually lasts thirty seconds. A gantry push is usually a four-to-eight-second idea. Do not pad to thirty because the slider goes there.

Native audio is on. Write “no speech, no invented clinician, room tone only,” or you will get a mouth you did not cast. A silent brief on a native-audio row still mints a soundtrack. Score later if you must. Do not ask the model to narrate a protocol.

720p is the ceiling on this family. Identity of the magnet matters more than a 4K invention of a different suite. Chase resolution on the same locked still, not by switching to text-to-video so the room can “look more cinematic.”

If you have a second approved still — closed table, then table in, same room — you can pin a last frame. That is control. It is not a default. A single hero still of the empty bore with a slow move is the usual ad.

A talking radiologist is a YPP kill

YouTube’s third inauthentic-content bucket is AI-generated personas that present as human experts on health, legal issues, finances, or politics. A constructed radiologist prescribing a scan, explaining a finding, or “what you should do with this report” is the same shape as a constructed doctor diagnosing a rash. Labelling the face AI does not leave the bucket.

A silent gantry is not that shape. A facility plate with authored type is not a human expert. The kill is persona plus prescription: a synthetic face styled as the person who reads the study.

Do not glue a talking white coat onto the magnet file inside one generate. If a protocol needs a voice, that voice is a real named credentialed person on a different day. Generation can still do the room. It cannot be the clinician.

Do not generate patients to “humanise” the bore. A photoreal body in the tunnel is a private figure you invented, or worse, one you referenced. Empty room. Empty table. Overlay the hours.

This is not the urgent-care clinician page

Urgent care clinics already have a practice page: wait times, a named clinician when someone must speak, a lobby people will recognise. Dentists are operatories and a chair. This listing is not those pages rewritten. Imaging is the magnet. The nearby audience is a clinician practice. The job here is a facility plate.

Hours, a phone, “evening slots,” a contrast instruction that is already on your intake sheet — those belong in type you typed after picture lock. Add a text overlay burns one authored line, whole clip, top / center / bottom. It does not transcribe. It does not invent a radiologist. Muted autoplay still needs a line on the file. Seedance will not letter the suite.

Write camera only. “Slow push into the empty bore, table static, no people.” Stop. “Dr. Maya explaining when you need contrast” is a second job. It is a private-person job if Maya is on payroll. It is a health-advice job if the line is a protocol. Either way it is not this generate.

FAQ

If I prompt “no medical advice,” can Seedance still put a radiologist in the suite?

The face is still a constructed presenter. Scan talk is still health. Hold the empty gantry. If someone must speak, put a real named credentialed person on camera, or do not prescribe.

Can I generate a patient on the table if I never show a face?

No. A photoreal body in a bore is still a generated patient. Photograph the empty table. Do not invent anatomy to prove the magnet is in use.

Do we start from a photo of our actual magnet?

Yes, when the suite has to be ours. That is image-to-video from the still, not a paragraph about a cinematic hospital. A stock tunnel is a different site.

Why not Veo 3.1 on every MRI post because it looks more “clinical”?

Veo is the dialogue and 4K row: native audio, official 4K tier, 4 / 6 / 8 second clips. A talking generate of a radiologist is the YPP pattern. Gantry motion on Seedance from a locked still is the imaging job.