Seedance 2.5 Letterbox Pop-Out 3D Illusion Prompting Technique
How to craft an out-of-frame 3D pop-out optical illusion in Seedance 2.5 where characters and props project over black letterbox bars using strict layer composi
A prompt engineering framework for generating a single-take, out-of-frame 3D pop-out optical illusion in Seedance 2.5—where characters and held objects break through top and bottom black letterbox bars—has been shared by AI creator Kōda (@aimikoda).

Image source: @aimikoda via X
Achieving a convincing out-of-frame 3D illusion directly from a generative video model without post-production compositing tools (such as After Effects or Nuke) requires strict prompt constraints. The model must be instructed on back-to-front compositing layers, immutable static screen graphics, foreground boundary penetration, and physical foreshortening across a continuous single-take comedy sequence.
Core Illusion Mechanism: Layer Compositing Order and Static Letterbox Rules
The primary mechanism that forces the generative model to render an out-of-frame pop-out effect is defining the scene hierarchy in a strict 'Back to Front' compositing order.
1. Back-to-Front Layer Hierarchy
The prompt structures the visual space into three distinct depth layers:
- Background (Back): A clean, pure white background confined strictly within the central picture window.
- Midground: Two broad, solid black letterbox bars fixed horizontally across the top and bottom of the screen.
- Foreground (Front): The animated character and the sequence of objects she presents to the camera.
While the white background is confined to the central picture window, the foreground character and her props are explicitly unrestricted. Their silhouettes continue uninterrupted across the boundary and render in front of the black letterbox bars with full color preservation.
2. Static Screen Graphic Constraints
To prevent the video model from deforming or moving the letterbox bars during physical character interaction, strict geometric invariance is specified:
- Immutable Screen Bars: Both black bars remain completely static screen graphics throughout the video, maintaining identical thickness, full canvas width, and straight horizontal inner boundaries.
- Overlap and Restoration: Foreground motion temporarily covers sections of the bars, but the visible edge segments on either side of any overlap remain perfectly aligned along the same horizontal axis. When an object is withdrawn, the original unchanged black surface is revealed beneath it.
- No Physical Reaction: Even when the character steps over the lower bar or touches the upper bar, the bars do not bend, slide, shake, stretch, open, or deform under physical contact.
Camera Setup and Strictly Nonverbal Performance Guidelines
To maintain visual clarity and comedic pacing across the single continuous take, specific staging and acting constraints are enforced:
Camera Staging and Framing
- Stationary Camera: A fixed wide-angle lens with level framing keeps the character close and prominent throughout the entire duration.
- Depth Staging: The character's face remains clearly visible within the central window, while her hands, props, and sneakers extend forward over the lower bar toward the viewer.
Nonverbal Acting and Implied Rejection
- Strictly Nonverbal Comedy: No spoken dialogue, whispering, narration, lip-syncing, vocalizations, subtitles, captions, or speech bubbles are permitted.
- Facial Dynamics: Continuous facial acting—including eye contact with the lens, natural blinking, eyebrow shifts, inquisitive head tilts, shifting smiles, and pouts—conveys the narrative.
- Implied Interaction: The viewer remains completely off-camera and silent. Every rejection is conveyed purely through the character's reaction after pausing to look into the lens.
Object Continuity and Audio
- Sequential Handling: Only one gift is presented at a time. Each object enters and exits through a side edge of the frame.
- No Morphing or Disappearing: Held items do not transform into subsequent props or vanish while in hand.
- Sound Design: Audio is restricted to subtle diegetic sound effects—light prop handling, fabric rustles, sneaker footsteps, and a single cat meow—with no background music or spoken vocals.
The 5-Stage Single-Take Gift Sequence
The continuous single-shot sequence progresses through an escalating five-stage comedic structure:
Stage 1: Flower (FLOWER)
The character smiles warmly, retrieves a single flower from beyond the side edge, and extends it toward the viewer. The flower, fingers, and sleeve project over the lower black bar. After an expectant pause looking into the lens, her smile fades. She nudges the flower forward encouragingly, accepts the implied refusal with a disappointed pout, and removes the flower offscreen.
Stage 2: Coffee (COFFEE)
She returns with a coffee cup held securely by its handle, her expression brightening with renewed confidence. She holds the cup close to the lens, overlapping the lower bar without spilling. When met with another silent rejection, her smile falters, her shoulders drop in resignation, and she sets the cup offscreen.
Stage 3: Jewel (JEWEL)
Inspired by a sudden idea, she presents a large, sparkling gemstone on an open palm. The jewel and her hand project visibly over the lower bar. Confident of success, she waits proudly, but upon rejection, widens her eyes and lifts her eyebrows in silent disbelief before placing the jewel offscreen.
Stage 4: Cat (CAT)
She brings in a small, fluffy cat supported with both hands, holding it toward the camera so its front paws and her hands clearly overlap the lower black bar. The cat blinks and moves naturally. After another implied refusal, her delight turns into exaggerated frustration. She gives the cat a comical, gentle backward toss, and the cat lands on its paws and trots away.
Stage 5: Herself (HERSELF) — Full Out-of-Frame Pop-Out
Her frustration shifts into a mischievous realization. She raises an index finger in a 'wait one moment' gesture, overlapping the upper black bar. She lifts one knee and extends her sneaker over the lower black bar, with the shoe appearing magnified through foreshortening. Shifting her weight forward, she steps entirely out of the picture window into the viewer's space—her coat overlapping the lower bar and her oversized pointed magician hat overlapping the upper bar simultaneously, while the letterbox bars remain stationary beneath her.
Full Seedance 2.5 Prompt
The complete prompt text as published by Kōda is provided below:
Use @[char ref] as the exact character and style reference. Preserve her long pale-blue hair, oversized blue-and-orange pointed magician hat, orange-and-blue coat with its circular sleeve emblem, blue knee-high socks and chunky orange-and-blue sneakers. Keep the same angular painterly anime rendering, sharp graphic shapes and character proportions. Replace the background with clean pure white.
Create a playful, completely nonverbal comedy in one continuous shot. She offers increasingly surprising gifts to the viewer, reacts to each implied rejection through expressive facial acting, then decides to step out of the picture herself.
MAIN VISUAL GOAL:
An unmistakable OUT-OF-FRAME, POP-OUT effect. Her offered objects, hands, sleeves, hat and finally her legs and body visibly project OVER the black letterbox bars toward the viewer.
COMPOSITING ORDER, BACK TO FRONT:
1. White background and scene.
2. Two broad black letterbox bars at the top and bottom.
3. The foreground character and the objects she presents.
The white background is confined to the central picture window. The foreground character and offered objects are NOT confined to that window. Their silhouettes continue uninterrupted across its boundaries and onto the black bars. Their original colors remain fully visible when overlapping black.
STATIC BARS, MOVING FOREGROUND:
Both bars are completely static screen graphics. Keep their equal thickness, full width and exact screen position unchanged throughout. Their inner boundaries remain perfectly straight and horizontal.
Foreground movement only temporarily hides portions of the bars. The underlying bars never move or deform. Visible edge segments on either side of an overlap remain aligned on the same horizontal line. When an object withdraws, reveal the original unchanged black surface beneath it.
Keep black visible beside the overlaps. No bending, sliding, shaking, stretching, opening or impact reactions. The bars remain unchanged even during the final step out of the picture.
CAMERA AND PERFORMANCE:
Stationary camera, fixed wide-angle lens and level framing. She stays close and prominent throughout. Keep her face clearly visible inside the central picture window while her hands and gifts extend over the lower bar.
Her facial acting remains continuously active: eye contact, eyebrow changes, natural blinking, questioning head tilts, shifting smiles, pouts and expressive cheek movements. Connect every reaction to the current offer.
The viewer remains unseen and silent. Every rejection is IMPLIED entirely by her own reaction after looking into the lens. No viewer hands, camera nodding or external rejection signals.
STRICTLY NONVERBAL:
No dialogue, spoken words, whispering, narration, lip-sync or human vocalizations. No subtitles, captions or speech bubbles. She never mouths sentences. Mouth movement is limited to natural nonverbal expressions.
Follow this exact order: FLOWER, COFFEE, JEWEL, CAT, HERSELF.
FIRST OFFER, FLOWER:
She smiles warmly and retrieves a single flower from beyond a side edge of the screen. Extend it toward the viewer, bringing the flower, fingers and sleeve visibly over the lower black bar.
After a brief expectant look into the lens, her hopeful smile fades. One eyebrow rises, her head tilts and her eyes search the viewer’s face. Give the flower one small encouraging forward nudge, then accept the implied refusal with a tiny disappointed pout.
Retract the flower and visibly place it outside the side edge of the screen before retrieving the next object.
SECOND OFFER, COFFEE:
Reach offscreen and return with a coffee cup held securely by its handle. Her face brightens with renewed confidence.
Offer the cup close to the lens. Its silhouette and her sleeve clearly overlap the lower black bar while her face remains visible above it.
Her smile falters. Look at the cup, then back into the lens with raised eyebrows and pursed lips. A small resigned shoulder drop follows.
Pull the cup back and set it offscreen without spilling.
THIRD OFFER, JEWEL:
Her eyes brighten with a sudden idea. Produce a large sparkling gemstone and present it in an open palm. The gemstone, fingers and sleeve project over the lower black bar.
Look proudly from the jewel to the viewer, expecting success. After the implied rejection, widen her eyes, lift both eyebrows and briefly part her lips in silent disbelief. Give the lens an incredulous stare, press her lips together and put the jewel offscreen.
FOURTH OFFER, CAT:
Retrieve a cute small cat from offscreen, supporting its body with both hands. Present it toward the lens, bringing its front paws, fluffy body and her hands visibly over the lower black bar.
Her expression becomes delighted and persuasive. The cat blinks and moves its paws naturally.
After another implied rejection, her delighted expression collapses into exaggerated disbelief. Look at the cat, then at the viewer, then back at the cat. Squint slightly and form a frustrated pout.
Turn slightly and give the cat a light, comical backward toss. The cat lands on its paws behind her and trots away. Keep this quick and playful.
FINAL OFFER, HERSELF:
Turn back to the viewer. Her frustration changes into a mischievous idea. Raise one index finger in a clear requesting-a-moment gesture, with the fingertip overlapping the upper black bar.
Lift one knee toward the lens and extend her sneaker OVER the lower black bar. The shoe grows large through foreshortening. Her leg crosses the picture boundary in front of the unchanged black surface.
Shift her weight forward and bring the rest of her body after it, as though stepping out of the framed image into the viewer’s space. Her coat overlaps the lower bar while her oversized hat overlaps the upper bar.
The bars remain perfectly straight and stationary beneath her. Her body simply renders in front of them. No opening, bending or displacement of either bar.
Finish close to the lens with open palms presenting herself, lifted eyebrows and a cheeky hopeful smile. Keep natural blinking, breathing and small facial movements through the ending.
OBJECT CONTINUITY:
One offered gift at a time. Show each retrieval, presentation and removal clearly. Objects enter and leave through a side edge of the output. No objects transforming into the next gift or disappearing while still held. Preserve consistent object shapes and natural hand contact.
Brisk natural real-time pacing with readable reactions. No slow motion, frozen expressions, distant retreats or camera movement. No warped bars, duplicated hands, costume changes or text.
Audio consists only of light prop sounds, fabric rustles, sneaker steps and a brief cat meow. No dialogue, human vocalizations or background music.
Original source
- Kōda Official X (@aimikoda): What Can I Offer You? (Seedance 2.5)