A red scarf on an empty rehearsal chair can suggest a person before an actor enters the frame. For a small theatre company’s imagined autumn production, I would make that absence the centre of a short promotional film. The brief would be deliberately modest: twelve seconds, one room, one meaningful movement, and an invitation to discover the story on stage. The teaser should create curiosity without pretending to show footage from a finished performance.
The campaign would begin with an ordinary marketing question. What should a stranger remember after seeing this clip between two unrelated posts? My answer would be the scarf lifting from the chair as someone opens a door outside the frame. Everything else, including lighting, sound, camera distance, and the final title, should support that image. I would describe this as a proposed creative exercise, rather than claim that a production or an audience test has already happened.
For this kind of rehearsal-room concept, Seedance 3.0 gives the planning process a useful starting point: an AI video model supporting text-to-video, image-to-video, and reference-based generation with text, images, video, and audio. The attraction is the ability to explain a scene through different kinds of material. A written direction can describe the action while a cleared photograph establishes the room. Neither input removes the need to review the resulting performance of the shot.
Decide what belongs to the production
I would first separate the fictional film treatment from information that must be factual. The venue name, dates, booking address, accessibility details, and production title would come from the theatre team. Those details belong in a final graphic assembled during editing, where the wording can be checked precisely. I would not ask a generated poster on the wall to carry the booking information. A beautiful frame would have little value if viewers read the wrong opening date.
The visual materials would be equally selective. I would request a photograph of the actual rehearsal chair and an approved reference for the scarf, with permission to use both. A short video showing the intended pace of a door opening could communicate timing without introducing an actor’s likeness. If the team supplied music, I would check that its use was cleared for the campaign. Reference material should explain a creative choice, rather than become a miscellaneous folder of things that look impressive.
Give every reference one job
In a Seedance 3.0 brief, I would label the chair photograph as the composition reference, the scarf image as the colour reference, and the short motion clip as a timing guide. I would state that the room should remain simple and that the scarf should not change colour or become another object. This division matters because otherwise a reviewer cannot tell whether an unexpected result came from an ambiguous instruction or a conflicting reference.
The written prompt would describe one continuous shot: a quiet rehearsal room in late afternoon, a red scarf resting on a wooden chair, a slight movement of air from an unseen doorway, and a slow camera approach that ends before the scarf leaves the seat. I would avoid requesting a stage transformation, a close-up of a face, and a dramatic camera orbit in the same generation. Those might each be interesting ideas, but together they would obscure this particular invitation.
My first Seedance 3.0 comparison would keep the framing fixed and vary only the movement. One version could lift the loose end of the fabric briefly; another could let it slide a few centimetres along the seat. This would make the test about how much mystery the scene needs. If I changed the room, lighting, music, and camera direction at the same time, choosing a preferred version would reveal very little about why it worked.
Review the small movement before the atmosphere
I would inspect the Seedance 3.0 output first without music. The scarf should retain a believable relationship to the chair, especially where cloth meets wood. Chair legs should stay in place, the floor should not drift, and the light should remain consistent across the shot. Any frame in which the cloth appears to pass through the seat would weaken the central image. Atmospheric colour cannot compensate for a physical distraction at the exact point where the audience is supposed to look.
A second pass would focus on the story. Does the movement suggest an arrival, or does it simply look like a fan has been switched on? If the latter, I would revise the timing so the change begins after a short still interval. I might also add a restrained off-screen door sound in the edit. That adjustment would support the existing action instead of replacing the entire concept with another elaborate prompt.
I would then compare the beginning and ending frames side by side. A useful teaser needs a place to enter and a place to leave. If the final frame feels visually crowded, I would shorten the camera movement or hold a wider composition. The invitation graphic needs breathing room after the image, and the film should not demand that viewers solve an intricate visual puzzle before they notice the production name.
Build the invitation around the chosen shot
Once a Seedance 3.0 version met those checks, I would take it into an editing timeline with the approved campaign information. My proposed structure would reserve a brief opening beat for the empty chair, several seconds for the fabric movement, and a clean ending for the production title and booking instruction. Actual durations would depend on the result. I would favour a clear transition over squeezing an extra visual effect into an already short piece.
Sound would be designed with the same restraint. A distant footstep, a door latch, or a small breath of room tone could establish an unseen presence. I would avoid implying that a particular performer supplied a voice unless that contribution was real and authorised. For a silent social feed, captions or a short on-screen line could carry the invitation. The essential story should still be understandable with the speaker muted.
Before export, I would check both the vertical and horizontal crops. The scarf is the narrative object, so it must remain visible in each version. Text should stay away from interface overlays, and the booking instruction should be legible at phone size. A crop that cuts off the moving fabric would require a different edit or a new composition, not a hopeful assumption that viewers will infer what happened outside the frame.
Keep the creative decision visible
I would save the accepted Seedance 3.0 prompt, reference permissions, review notes, and final edit together. That record would make a later campaign revision easier: another date graphic could be changed independently, while a new scarf colour would require a fresh visual review. The distinction prevents small administrative changes from turning into unnecessary regeneration and keeps the theatre team in control of its public information.
The final question would remain simple: does this short film make the production feel worth discovering? Seedance 3.0 would be part of the image-making process, while the creative responsibility would stay with the people choosing the story. An empty chair, a moving scarf, and a carefully timed invitation can be enough. Leaving something unsaid is only effective when the few things shown are deliberate, coherent, and easy to remember.
