Create a 30-second Japanese sword-action comedy: reference-style 2D anime characters composited into

Prompt

Create a 30-second Japanese sword-action comedy: reference-style 2D anime characters composited into photorealistic Shibuya. Synchronized dialogue, music and effects. Use the selected aspect ratio. REFERENCES A = black-haired heroine from @Image 1; her voice uses @Audio 1. B = white-haired heroine from @Image 2; her voice uses @Audio 2. All enemies use the design from @Image 3. Preserve exact faces, hair, costumes, accessories, proportions and reference art styles. Images define identity, not poses or sheet layouts. Equal heroine heights, excluding ornaments. A and B are allies; ONE of each. Replicate only enemies. Audio references define timbre and natural pitch, not scripts. Never replay recorded words or swap voices. Enemies are silent. STYLE AND LOCATION A, B and enemies retain their reference illustrated styles, clean lines and cel shading. Never turn them into live-action people or 3D figurines. Photorealistic Shibuya Scramble Crossing at night after rain: diagonal zebra stripes, station frontage, curved commercial buildings, LED screens, wet asphalt and realistic glass, concrete and steel. Screens show abstract colors, no readable brands. Integrate the illustrated characters through matching light direction, soft contact shadows, wet-road reflections and colored light on their cel shading. Keep their faces and linework sharp. Buildings, airborne debris, dust and destruction have realistic materials, weight and lighting. Sword energy remains vivid anime-style light. Only heroes and enemies occupy the streets; no civilians or moving traffic. Establish Shibuya in the first frame before the destruction. OPENING SPECTACLE AND CONTINUITY Begin directly with a combined ultimate attack; no title or long setup. Both swords are already charged in the first frame. TEN THOUSAND enemies fill the crossing and roads in deep layered ranks. A's teal slash and B's violet slash visibly merge into ONE gigantic crescent, expanding into a pressure wave that blasts enemies AND buildings outward. Each heroine has one right-hand silver katana and a waist scabbard. Left hands may support the grip. Keep blades rigid and clear of each other. Show contact before displacement. Enemies stay down. Close combat uses three to six attackers, preserving @Image 3. Sparks and knockdowns; no blood or exposed wounds. SHOT 1 — COMBINED ULTIMATE ATTACK, 0–3 seconds Low wide view across the crossing: A left, B right, side by side, facing a vast enemy formation. Their charged blades and wet road reflections glow teal and violet. A slashes diagonally upward, B downward, in separate spaces. Energy trails cross ahead and fuse into one white-gold crescent. Bodies and blades never collide. Around 0.1–0.8 seconds, A uses @Audio 1 and says: 「いくよ!」 Around 0.9–1.6 seconds, B uses @Audio 2 and replies: 「あわせる!」 These are the only lines; accurate lip sync. At 1.6 seconds the crescent launches the front ranks. Show merger and impact clearly. SHOT 2 — TEN THOUSAND BLOWN AWAY, 3–8 seconds High oblique view over the crossing and roads. Both heroines stay grounded at the origin; debris flies away from them. The pressure front lifts successive enemy ranks and drives thousands of silhouettes backward with spinning weapons. It also strikes buildings along both roadsides, blasting them away from the heroes. Show a traveling chain reaction, not an instant disappearance. At wave contact, windows shatter, signs tear away, and buildings rip from their foundations and fragment. Huge concrete sections and steel frames fly OUTWARD with the army. Show collapse spreading building by building, never an unrelated explosion. Briefly slow the airborne peak, then resume speed. Illustrated enemies and photorealistic building fragments fill the sky. Keep silhouettes readable through dust. SHOT 3 — BACK-TO-BACK GUARD, 8–11 seconds Low medium-wide: the heroines slide back-to-back, feet grounded. The army stays down. Keep buildings ruined throughout. A fresh @Image 3 squad enters over distant rubble; show their arrival, no respawning. The heroines exchange a glance and face them. SHOT 4 — LOW/HIGH TEAMWORK, 11–16 seconds Full-body lateral tracking. A advances into three attackers, ducks an incoming sword and sweeps low. Show blade contact, sparks and the attackers losing balance. B leaps above A's crouched shoulder, clearly separated, and strikes another attacker. Show contact and its backward fall. B lands ahead as A rises. Hold leap and landing continuously, keeping blade paths distinct. SHOT 5 — MUTUAL PROTECTION, 16–22 seconds Wider three-quarter view. An enemy attacks A's rear. B steps in and visibly parries above A's shoulder. A pivots beneath the clash and strikes the attacker approaching B's side. They turn back-to-back and simultaneously slash outward into opposite sectors. A short teal-and-violet shockwave knocks down the remaining close attackers. This follow-up is smaller than the opening ultimate attack. Dust clears. The main army remains down, the reinforcements are defeated, and ONE last enemy is standing several meters ahead. Both heroines lower their blades. SHOT 6 — TOO POLITE TO FINISH, 22–27 seconds Stable wide three-shot on the crossing: A left, B right, last enemy ahead. Stop the action music. A offers the last enemy with her empty left palm, right-hand katana lowered. B returns the gesture. Both stay serious, looking at each other. The enemy quietly lays its sword on a zebra stripe, backs away and runs out of frame. Show this continuously. A and B turn back together and discover only the abandoned sword. Hold an awkward pause. No dialogue explains the joke. SHOT 7 — THE FAMILIAR FACE, 27–30 seconds Frontal close-up of anime A against a blurred photorealistic ruined Shibuya skyline. Preserve her exact face and hair. Wide eyes, small pupils, raised brows and lips slightly parted once, then held still: her familiar stunned, uncomprehending expression. A looks toward camera. Push in; hold the final full second, sword out of crop. No blush, crying or speech. End on A. SOUND AND MUSIC After the two opening lines, no dialogue, narration, grunts, gasps, screams, laughter or vocal breaths from anyone. Instrumental drums, strings and brass drive combat. Synchronize swords, impacts, collapses and landings. Opening: huge bass impact and cascading crashes. Cut music at Shot 6. Hear the sword, retreating steps, settling debris and wind. No choir, vocals, comedy sting or canned laughter. No text, subtitles, logos, watermarks, HUD, hero clones, mixed identities, extra limbs, fused bodies, morphing swords, costume changes, persistent afterimages, uncontrolled camera spins, realistic human faces, 3D character redesign or chibi.

Reference Images

@tokyo_Valentine19

You may also like