Complete This Small Table Before Drawing Vertical Storyboards
| Item | Record |
|---|---|
| Shot priority | Ask yourself: What should the audience know after this shot?; Record: One main fact or character decision |
| Action space | Ask yourself: Who can reach the object, and who blocks whom?; Record: Relative positions of the door, counter, aisle, and people |
| Simultaneous reactions | Ask yourself: Which two things must remain visible together?; Record: Visible positions of the actor and recipient |
| Shot size | Ask yourself: Is the face, hand, or overall relationship most important?; Record: A frame wide enough to read that information |
| Playback obstruction | Ask yourself: Where will this delivery’s subtitles, badges, or controls appear?; Record: Occupied areas from the actual screenshot or template |
Solve the shot’s storytelling task before pursuing a “cinematic” look. A vertical frame has limited width. Packing every character side by side shrinks faces, hands, and key objects at once. Depth blocking creates room, but a foreground figure can still cover the very evidence the shot needs to reveal.
This is an editorial worksheet narrowed to the reader’s question, not the complete native storyboard format. A shot may center on emotion, waiting, or relational distance; it need not uncover new evidence. The scene determines the coverage.
Choose Shot Size by Information, Not by Habit
| Item | Record |
|---|---|
| Wide shot | Best for establishing: Entrances, three-person positions, and who blocks whom; Information at risk: Small expressions and text on paper cannot carry the main point |
| Medium shot | Best for establishing: Hand actions, interaction, and object handoffs; Information at risk: Confirm that the important object remains inside the frame |
| Medium close-up | Best for establishing: Speech and reaction with necessary gestures; Information at risk: A background character or exit can easily leave the frame |
| Close-up/detail | Best for establishing: A decision, decisive feature, or individual reaction; Information at risk: Spatial reference decreases, so viewers need to know where it belongs |
Find a working shot size that keeps the scene’s continuing interaction readable, then change it when attention shifts. If two people negotiate over a key on the counter, a medium shot can retain both hands and the key. Insert a detail only when an engraving or other small feature must be read.
Cutting to a talking head for every line loses the process of approaching, yielding, and blocking. Holding one wide shot throughout can make a decisive small action unreadable.
Original Three-Person Scene: Who Spots the Typo Before Engraving?
Story: At a pottery-shop counter, customer Wen Yue arrives to collect a personalized mug. Owner Qiao Wen offers the inscription proof for her to inspect, keeping hold of the sheet, while an apprentice picks up the blank mug and prepares to enter the rear workroom. Wen Yue sees that the requested name “江舟” has been written as “江周.” Both final characters are pronounced Zhou, but 舟 means boat; the written character is the evidence here. Qiao Wen immediately stops the apprentice so the proof can be corrected first.
Establish the space: Wen Yue stands on the customer side of the counter. Qiao Wen stands behind it. The apprentice waits by the rear-workroom door. Qiao Wen holds the proof; the apprentice holds the blank mug, which has not been engraved. The audience must read the order in which the customer notices the error, the owner gives the stop instruction, and the apprentice receives it.
| Item | Record |
|---|---|
| Establish the interaction | Composition: A vertical medium shot looks diagonally from the counter toward the rear. Wen Yue occupies the nearer side, Qiao Wen stands slightly deeper on the other side, and the apprentice remains visible by the door. Wen Yue’s shoulder does not cover the proof Qiao Wen extends.; What it accomplishes: Establishes who controls the proof and blank mug and that the apprentice is about to leave |
| Reveal the error | Composition: In a readable detail of the proof, show “江周.” Wen Yue’s finger stops beside 周: “The last character should be 舟, meaning boat, rather than 周.” Qiao Wen continues holding the sheet.; What it accomplishes: Makes the label error legible and gives a sufficient reason to pause engraving |
| Send and receive the instruction | Composition: Return to an angle containing the counter and workroom door. Qiao Wen calls to the apprentice. He stops before the doorway and turns back, still holding the mug.; What it accomplishes: Shows that the stop instruction was heard instead of filming only the person who gave it |
A later shot can show the corrected proof and the customer’s approval. These three shots demonstrate a shift of attention from paper to doorway; they do not prescribe three shots for every scene.
If the key lettering falls beneath the subtitles used in this delivery, adjust the detail framing until the proof remains readable. If a counter close-up removes the apprentice entirely, give him a visible response after the stop instruction. Do not avoid an overlay by moving the mug into the customer’s hand, and do not let a later shot suddenly contain a mug that has already been engraved.
The story supplies both “江周” and the intended “江舟” verbatim, and the text policy requires legibility. The storyboard places the surface, angle, and reading window rather than inventing another name. The text instruction and detail specification still need checking in the actual output; before generation, legibility remains a requirement.
Vertical Safe Areas: Confirm the Obstruction Before Reframing
A “safe area” must correspond to what will actually overlay this delivery. Ask the delivery owner for a page mockup, template, or screenshot, then confirm which bands are occupied, whether they remain onscreen, and what type of content covers them.
If subtitles sit at the bottom, inspect any paper text, handoff, or seated movement occurring there. If interaction controls occupy one side, check whether an essential gesture happens only beneath that strip. Without a confirmed layout, design for the image’s own readability and flag the overlay area for confirmation rather than inventing a universal percentage.
| What is obstructed | Preferred revision |
|---|---|
| Text on paper that must be read | Change the camera angle and framing, or add a readable detail shot |
| A decisive hand action | Preserve the real spatial relationship while changing the viewing angle so the action leaves the covered area |
| A character’s crucial reaction | Choose a better shot size or composition; retain the reaction instead of merely shrinking the subject |
| Unimportant background | Keep the composition; not every obstruction requires restaging the scene |
Overlay the player’s occupied areas on the storyboard sketch or keyframe. If the delivery page or subtitle style changes, recheck the relevant shots. Clearance decisions from an older layout do not carry over automatically.
How to Avoid Screen-Direction Jumps and Vanishing Characters
- Record the line between the characters and the side from which the camera primarily observes it.
- Write “left side of frame” separately from “the character’s left hand” so two meanings of left do not collapse into one.
- In shot/reverse shot, match eyelines. If the camera crosses to the other side or a character changes position, use visible movement or a new establishing view to restore orientation.
- When a third person remains present, decide which reaction viewers need now. Keep that person in the shared frame, give them a separate reaction, or withhold them deliberately; do not forget them merely because the edit has moved into two close-ups.
- An object handoff occurs once. Later detail shots show the new state rather than replaying the same yield or transfer.
Vertical composition can use occlusion, but it should serve the story. Concealing a character who has not yet been revealed is different from accidentally covering a hand viewers already need to see.
FAQ
Does vertical short drama require lots of close-ups?
Choose according to the fact the current shot must make readable. A close-up emphasizes reactions and details but temporarily removes spatial relationships. When a multi-person interaction, yielding, or blocking must remain visible, retain an appropriate medium or wide shot.
Can I simply crop a horizontal video into a vertical frame?
First check whether the crop preserves character relationships, essential actions, and evidence. If the two ends of an action sit on opposite horizontal edges, a simple crop removes one end. You will need to reframe, add coverage, or change the layout.
Do it with the skill
In Claude Code, say: “Use /short-drama-storyboard to design a vertical storyboard for this scene. Establish every character and object position first, then choose each shot size according to the information viewers must read. Mark obstruction risks using the player layout I provide, and preserve continuity of action and eyelines.” In Codex, use $short-drama-storyboard.
Read the method: Vertical groups, delivery overlays, and axes · Shot scale, source coverage, and boundaries
About Drama SkillsThe short-drama-storyboard skill on GitHub