最近、SeaArt Flow 2.0 という動画生成ツールが公開されました。
SeaArtでは創作コストが200コインの、今までの動画生成モデルでも破綻なしで動くものも多い中（ぜひこっちの、動画生成モデル比較記事で確認してみてください）80％オフでも1037コインというコストのデカさにビビるかもしれませんが、コストのデカさに見合うだけの理由があります。
それは『プロンプト指示を間違えなければ、動画・エフェクト・音楽・音声・効果音すら思い通りに一発で生成可能』という、とんでもない事実です。
カバー画像に使っている動画も、ノー編集。T2Vで一発出しです。
なので、今まで個人的に使って確かめたプロンプトの傾向や、どんなものが作れるのか、やらかしてしまったミスなど、自分なりの感覚を紹介します。
SeaArt Flow 2.0、どんなの作れるの？
T2V、I2V、ともに8秒の1280×720の横長16:9動画を生成可能です。
プロンプトの長さにも調べた限りでは制限がないですが、あまりにも長すぎるとそもそも8秒間の動画として収まりきらないです。
実験で試したプロンプトで3,100文字以上のものも問題なく生成可能でした。思った通りの動きはしてくれませんでしたが…
あと、なぜかプロンプト上でも生成物としても問題ないのにアダルト認定されて非表示にされる時もあります。
しかしエラーの場合はちゃんとスタミナ・コインは返還されます。安心。
T2V生成物1：西部劇風
画像クリックで大本の動画URLに飛ぶようにしています。（いかなかったらリンクURLからお願いします）
設定によっては、すべての動画でいきなり大音量が流れるのでご注意ください。
https://www.seaart.ai/ja/explore/detail/d0t0f15e878c739lqs3g
これがまず最初に試した生成物になります。プロンプトはこちら。
---
The searing high noon sun relentlessly beats down on the desolate, wind-swept main street of Redemption Gulch. Weathered wooden facades of a saloon and general store stand silently, casting sharp, elongated shadows. In the very center, two figures stand rigidly opposed, twenty paces apart, their silhouettes stark against the ***ing glare. A hardened cowboy, his face a mosaic of lines beneath a faded hat, holds his antique Colt Peacemaker steady, its barrel pointed directly at his adversary. His eyes, slivers of steely resolve, are locked on his opponent, a formidable figure in a pristine black hat and coat, whose ornate, nickel-plated revolver is held with cold precision. The only sounds are the mournful creak of a distant windmill and the gritty whisper of wind through parched tumbleweeds. A tense, minimalist acoustic guitar riff swells, punctuated by deep cello chords, underscoring the unbearable tension. The cowboy's low, gravelly voice cuts through the stillness: "This ends here. For good." The camera executes a rapid, tight zoom onto their hands, poised rigidly over triggers, then swiftly cuts to their eyes, locked in a death stare. A deafening CLICK-CLACK as both hammers are simultaneously ***ed. A split-second pause. Then, a ***ing flash of muzzle fire erupts from both barrels, accompanied by a deafening CRACK that echoes, followed immediately by the dull thud of a falling body.
---
脳内イメージではもっと人と人の間に距離があり（In the very center, two figures stand rigidly opposed, twenty paces apart, their silhouettes stark against the ***ing glareと指示した箇所）、あともう一つ、キャラクターは道を横断するように並ぶのではなく、道に沿って沿ってもっと離れた位置で銃を構えて決闘をして欲しかったのですが……
今思えば、撃って撃たれて流血の可能性があるので、初手のテスト動画として生成するものとしては不適切だった気がします。
BGM（A tense, minimalist acoustic guitar riff swells, punctuated by deep cello chords, underscoring the unbearable tension. ）とセリフ（"This ends here. For good." ）に関してに関してはうまくいっていますね。
T2V生成物2：中世ヨーロッパ風
西部劇に続いて、異なる時代の映像は生成できるのか？ と思って作ってみた中世ヨーロッパ風の動画です。
「木の下で傷ついた騎士のためにリンゴを持ってくる村娘」が脳内イメージ。
https://www.seaart.ai/ja/explore/detail/d0tk2pte878c73ak7vrg
プロンプトはこちら。
---
An ancient, gnarled oak tree provides dappled shade on the outskirts of a peaceful medieval village. Beneath it, a world-weary knight, his face scarred, slumbers fitfully, his greatsword resting beside him. The gentle rustling of leaves and distant bird chirps fill the air, accompanied by a soft, melancholic lute melody. A young maiden, the village lord's daughter, approaches with hesitant grace, carrying a vibrant crimson apple. She softly clears her throat. The knight stirs, his eyes opening, settling on the maiden who extends the apple. His scarred hand slowly reaches out, fingers brushing hers as he accepts it. A blush creeps across the maiden's cheeks as she casts her eyes downwards shyly, a small smile forming. The knight's rough voice offers, "My thanks, maiden," a flicker of warmth in his tired eyes. The lute swells slightly, underscoring the tender exchange.
---
リンゴが手の平で消えてしまったことと、騎士が起きるスピードが速すぎることだけがマイナスですが、そこはプロンプトに「騎士はリンゴを握って微笑む」だったり「町娘の声に気づいてゆっくりと目を開ける」などを指示すべきだったなと思います。
騎士を示すために（his greatsword resting beside him）と指示をしたため、剥き身の剣を枕にしている恐怖がありますね。「鞘に入った剣」と指示するべきでした。
でも、BGMとセリフの指示である（The knight's rough voice offers, "My thanks, maiden," a flicker of warmth in his tired eyes. The lute swells slightly, underscoring the tender exchange.）は問題なく通っています。
T2V生成物3：舞台風
次は、場面転換を指定したくて生成したものです。
舞台内容としては「復讐のために魔王に身を落とした主人公を救おうとするヒロインのシーン」をイメージしました。
緊迫感があるシーンを見ている観客越しのステージ→観客の息をのむ横顔→ステージへの画面転換、を指示してみたかったのですが。
https://www.seaart.ai/ja/explore/detail/d0vhme5e878c739e8hh0
プロンプトはこちら。
---
The opulent, gothic-inspired set of a grand theater stage is bathed in dramatic, crimson and violet stage lighting. Center stage, a towering figure, portraying the Demon King, stands cloaked in intricate, shadowy ***, a horned helm framing his commanding presence. He wields a m***ive, obsidian staff, its tip glowing with an ominous internal light. A soaring, epic orchestral score, a villain's theme, swells through the theater, punctuated by thunderous timpani. The Demon King raises the staff high, a single, piercing spotlight locking onto him, casting long, dramatic shadows. His voice, powerfully amplified and resonating with a sinister, theatrical echo, booms across the auditorium: "Now is the time for vengeance upon this world!" The camera rapidly sweeps from the imposing figure to capture the hushed, captivated faces of the audience, their expressions a mix of fear and awe, then dramatically pivots back to the stage. A young woman, ethereal in a flowing, simple white gown, rushes forward, her face etched with profound anguish. She throws her arms around the Demon King, clinging to him desperately. The grand orchestral score abruptly softens, transforming into a tender, pleading string melody. Her voice, choked with raw emotion and amplified by the stage's hidden mics, cuts through the sudden stillness: "Stop it! It's enough, you don't have to be evil anymore!" The camera focuses on the stark contrast: her pleading face buried in his ***, his rigid posture slowly softening, the glowing staff still held aloft but now wavering slightly.
---
観客のど真ん中に舞台があるので脳内イメージとはちょっと違いましたが、これはたぶん（Center stage, a towering figure, portraying the Demon King, stands cloaked in intricate）の指示がダメだったのかなと思いました。
（The camera rapidly sweeps from the imposing figure to capture the hushed, captivated faces of the audience, their expressions a mix of fear and awe, then dramatically pivots back to the stage.）という、部分はうまく通ってないですね。
そして変わらず音に関しては（A soaring, epic orchestral score, a villain's theme, swells through the theater, punctuated by thunderous timpani.）（His voice, powerfully amplified and resonating with a sinister, theatrical echo, booms across the auditorium: "Now is the time for vengeance upon this world!" ）（ Her voice, choked with raw emotion and amplified by the stage's hidden mics, cuts through the sudden stillness: "Stop it! It's enough, you don't have to be evil anymore!"）がちゃんと通っています。（The grand orchestral score abruptly softens, transforming into a tender, pleading string melody. ）の音声指示だけが、8秒の動画としては難しい指示だったようです。
あと、字幕が変なことになってますね。
こういう字幕がつくぐらいなら「字幕なし」とかを指示したほうがいいのか？と思ったりします。
T2V生成物4：SF風
今度は動きが激しいものを試してみたかったので、動きが早いSFのスペースシップとコックピットの様子を指示しました。
爆発のようなエフェクト要素もどうなるか気になったので、それも確認したくて、激しい攻防戦、スペースシップ、コックピットで慌てて叫ぶ男性というイメージだけでプロンプトを作りました。
https://www.seaart.ai/ja/explore/detail/d100qd5e878c73adtaq0
プロンプトはこちら。
---
The inky blackness of deep space is violently torn apart by streaks of emerald laser fire and ***ing crimson energy blasts. Sleek, agile fighter crafts engage in a chaotic dogfight amidst a backdrop of swirling nebulae and distant star fields. Explosions bloom silently like violent, beautiful flowers, sending glittering debris tumbling through the void. Inside the cramped, glowing ***pit of a lead fighter, a male pilot, sweat-streaked and grim-faced, aggressively yanks at a joystick, his fingers flying across holographic displays. Blaring alarms and the hiss of straining hydraulics fill the confined space, underscored by a pulsating, urgent sci-fi orchestral score. His voice, strained with adrenaline and amplified by his comms, slices through the chaos: "They're on my tail! I can't shake them!" Suddenly, a m***ive enemy capital ship looms into view, unleashing a barrage of thunderous plasma cannons. The ***pit shudders violently, showering sparks, as the pilot throws the craft into a desperate barrel roll, punctuated by a deafening BOOM that shakes the entire vessel to its core.
---
ビームの色（The inky blackness of deep space is violently torn apart by streaks of emerald laser fire and ***ing crimson energy blasts.）と爆発（Explosions bloom silently like violent, beautiful flowers, sending glittering debris tumbling through the void.）は問題なし。
男性の行動の（Inside the cramped, glowing ***pit of a lead fighter, a male pilot, sweat-streaked and grim-faced, aggressively yanks at a joystick, his fingers flying across holographic displays.）は、レバーをガチャガチャは通っていますが、ホログラムディスプレイを撫でる、は省かれました。
セリフ（slices through the chaos: "They're on my tail! I can't shake them!"）は問題なし。
I2V生成物1：SF風
このプロンプトならI2Vでもいけるのではないか？と思ったので、1フレーム目に使えそうな画像を作りました。
それがこちら。
コックピットから始まってほしかったので、画像プロンプトは以下の通りです。
---
A dynamic, close-up shot of a male pilot's sweat-streaked, grim face inside a cramped, futuristic fighter ***pit, illuminated by the green glow of holographic displays. His hand is clenched tightly around a joystick, eyes wide with intense focus and adrenaline, reflecting the unseen chaos of a space battle outside. Sparks subtly fly in the background, hinting at recent damage. The lighting is dramatic and tense, with sharp shadows and highlights, emphasizing the high-stakes situation and a sleek, metallic sci-fi aesthetic.
---
そしてこれを使い、I2Vしたものがこちら。
https://www.seaart.ai/ja/explore/detail/d101bdte878c73f3acb0
プロンプトはほぼ一緒で、コックピットについての指示と宇宙空間で戦闘している指示を入れ替えただけです。
相変わらずホログラムは指でなぞってはないし、最後になぜか機体が下からやってくることに笑ってしまいましたが、しっかり画像を認識して前半と後半の人物にブレがないのは素晴らしいと思いました。
今までの生成動画モデルは最初と最後のフレームで同じ画像を指定しない限り、ここまで同一のキャラを最後に使うことはあまりなかった気がします。
T2V生成物5：和風
激しい人以外の動きがいけるのならば、当然、人の激しい動きとカメラの動きも試したかったので、和風で、侍が対峙している日本映画の脳内イメージを作ってみました。
和風っぽい音も、どんな形で生成させるのかも気になったので。
https://www.seaart.ai/ja/explore/detail/d1014mle878c73ec2hf0
プロンプトはこちら。
---
A desolate, ancient Japanese landscape shrouded in mist, with towering bamboo groves flanking a narrow, overgrown path. In the foreground, a lone samurai, his back to the camera, stands silhouetted against the ethereal gloom, clad in a tattered, weathered kimono, a katana ?ed at his hip. Beyond him, piercing the mist, stands a majestic, ancient five-story pagoda. The only sounds are the distant, mournful call of a cuckoo and the soft rustle of wind through bamboo. A solemn, traditional Japanese flute (shakuhachi) melody begins, slow and deep. The camera executes a rapid, dramatic pan-up, swiftly ascending the pagoda's height to reveal the same samurai now perched precariously on its topmost spire, his tattered cloak ? in the wind. A subtle, mischievous smirk plays on his lips. With a sharp SHIING of steel, he draws his katana from its ?, the blade glinting ominously. He raises the sword high, then, with a powerful leap and a *** battle cry, launches himself directly towards the camera, blade leading, the air whistling with the sound of his descent before the screen is filled with the ***ing flash and SWISH of his downward strike.
---
動画の始まりのイメージ（In the foreground, a lone samurai, his back to the camera, stands silhouetted against the ethereal gloom, clad in a tattered, weathered kimono, a katana ?ed at his hip.）としては、指示通りで、脳内イメージとほとんど一緒でした。
BGM（A solemn, traditional Japanese flute (shakuhachi) melody begins, slow and deep.）も、だいたいイメージ通りです。
一番脳内想像と異なったのは（He raises the sword high, then, with a powerful leap and a *** battle cry, launches himself directly towards the camera, blade leading, the air whistling with the sound of his descent before the screen is filled with the ***ing flash and SWISH of his downward strike.）でした。
塔のてっぺんに立っている侍が、カメラに飛び掛かって刀を振り下ろす、というシーンを作りたかったのですが、うまくいっていません。
指示としても、日本語から英語に翻訳しているので細かいニュアンスが通ってないだけの可能性がありますが、カメラに何かをする、というのはもしかして学習されたないのかな？
I2V生成物2：和風1
T2Vで難しい希望通りに動かすのが難しい脳内イメージならば、I2Vならどうだろう？と思い、試行錯誤した結果をお見せします。
個人的感覚として、I2Vの方が制御が難しい気がしました。
まずは、脳内イメージに合うような画像を生成します。
https://www.seaart.ai/ja/explore/detail/d101ec5e878c73f3euh0
T2Vの出力を見た限り、真後ろからのカメラアングルだと背中を向けている侍の顔を描写できないので、ちょっと斜めで出力されたものを採用。
プロンプトはこちら。
---
A solitary samurai's back view, clad in a tattered, weathered kimono, stands silhouetted against a mystical, mist-shrouded ancient Japanese landscape. He holds a katana ?ed at his hip. In the background, a majestic, towering five-story pagoda pierces through the swirling mist, its intricate wooden architecture subtly lit. The atmosphere is solemn and ancient, evoking a sense of *** vigilance, with a deep, moody color palette ***ted by cool blues, greys, and muted greens, accentuating the dramatic silhouette and ethereal mist.
---
そして、T2Vと同じプロンプトを使って生成したものがこちらです。
これ、リンクから確認してもらったらよく分かると思いますが、指定した1フレーム目が一瞬で違うシーンに切り替わっています。
間違いなく、T2Vのプロンプトの出だしである「A desolate, ancient Japanese landscape shrouded in mist, with towering bamboo groves flanking a narrow, overgrown path. 」が、指定した画像の説明と異なるからだと思います。
そしてカメラアングル、塔の上の侍の動きなどは大体T2Vと一緒でした。
I2V生成物3：和風2
T2Vの情景説明を省き、カメラの動きなどを指定するようなプロンプトに書き換えてみて生成したのがこちら。
https://www.seaart.ai/ja/explore/detail/d104bqle878c73ajt7fg
---
Dynamic effects: The camera quickly pans vertically upwards, revealing the highest tier of the pagoda emerging from the fog. The same samurai stands atop its pinnacle, a defiant, twisted smirk slowly spreading across his face. The samurai slowly, yet decisively, draws his katana from its scabbard, its dark, gleaming blade casting a dull gleam. As he powerfully raises the sword high above his head, he boldly leaps towards the camera, defying gravity as he brings the katana down with full force in mid-air. Scene development: A shift from a serene introduction to an intense action sequence, highlighting the samurai's defiant presence and building towards a ***. Style: Hyperrealistic, cinematic, Japanese historical drama style, dark fantasy, Bushido, majestic. Audio: {environmental sound effects: sudden roar of wind breaking the deep mountain's silence, faint rustling of moving fog, the samurai's defiant laugh, sharp metallic sound of drawing the sword, a thunderous thud as he pushes off the ground, intense whooshing sound as the blade cuts through the air; background music: shifts from quiet, profound Japanese ambient to an increasingly tense and grand Japanese orchestra, with heavy percussion resonating at the ***} Camera: Starting from behind the samurai, the camera quickly cranes up (pans up) vertically to capture the full view of the pagoda and the samurai standing on its highest point. The camera slowly zooms in on the samurai's face, emphasizing his defiant smirk. Subsequently, as the samurai raises his sword, the camera slightly pulls back, then dynamically moves to follow him as he leaps and brings down his sword, maximizing impact and immersion.
---
T2Vより動きと音の迫力は出たのですが、カメラの動きの指示は完全に無視されました。
I2V生成物4：和風3
変に暴れたりしないように、背中を向けていた侍にこっちを向かせて喋らせてみるというプロンプトに書き換えました。
また、せっかくなので日本語を喋ってもらおうとした結果がこちら。
https://www.seaart.ai/ja/explore/detail/d1055pte878c73ej5300
---
A ronin in tattered attire stands quietly with his back to an ancient five-story pagoda in a misty mountain valley, his katana faintly glowing. As he slowly turns back over his shoulder, resolve settles in his eyes, ***ering, "覚悟は…決まった," and his katana is fully drawn with a crisp "shiiiing" sound. Mist drifts solemnly, distant trees sway. The camera zooms into the ronin's face, then shifts to the pagoda's peak, focusing on ***ther samurai who smirks defiantly, declaring, "見せてみろ、お前の力を" He then powerfully leaps from the pagoda, brandishing his sword mid-air, bringing it down towards the camera. The video style is breathtakingly solemn, Yugen cinematic Japanese period dark fantasy. Audio features environmental sounds of mist, wind, the ronin's murmur, sword draws, and slashes, complemented by a mysterious and grand musical score with shakuhachi, ***o, taiko, and strings building to a ***. The camera begins behind the ronin, moves to a close-up on his face, focuses on the samurai on the pagoda's top, and transitions to a dynamic, subjective viewpoint tracking the jump and downward sword strike.
---
動きの指示は個別よりも文章の方がいいのか？と思い、再度文章としてのプロンプトに変更しました。
日本語の読みがおかしく、字幕も変になっていますが、ただ、動きは希望の動きに近づきました。
I2V生成物5：和風4
もしや、カメラに向かって切りかかったりするのは難しいのか？と判断し、そもそものプロンプトを書き換えました。
ゲームっぽく映像スタイルを何個も切り替えるようにして、カメラや動きなどを個別に指示するプロンプトに戻し、セリフも英語に戻しています。
男が走る→カメラが男に近づく→男がセリフを呟く→水彩、水墨画で描かれたように動画スタイルが変わっていく、としたかったのですが…
https://www.seaart.ai/ja/explore/detail/d10v17te878c73fricb0
プロンプトはこちら。
---
In a quiet, deeply misty mountain setting, an ancient five-story pagoda looms fantastically. The air is cold and clear, and moss-covered ground and puddles that dimly reflect the light around the samurai's feet spread out with a realistic texture. Filling the frame, a lone samurai stands with his back turned, clad in tattered, ancient attire, firmly clutching a drawn katana in his right hand. The blade reflects the misty light dully, carrying a gleam of quiet resolve. The overall tone is a bluish, subdued light, creating a mysterious and solemn atmosphere, initially depicted with high-definition, realistic detail.
Action: After quietly surveying his surroundings, the samurai takes a deep breath and slowly but powerfully re-positions the katana in his hand. His eyes are fixed on a single point as he takes a resolute step towards the five-story pagoda shrouded in mist, carefully avoiding the puddles. As he takes this first step, the entire video gradually loses its realistic texture, transitioning into a soft, bleeding watercolor or ink wash painting style with visible brushstrokes and subdued colors. In sync with the samurai's movement, the flowing mist in the background and the distant mountains also transform into a soft-touch Japanese painting style, with blurred outlines yet emotionally rich depiction.
Audio: {Environmental Sound Effects: Distant bird calls, sound of cold wind blowing quietly, irregular dripping water sounds, footsteps of the samurai, subtle metallic sounds of the sword, sound of deep breath; Dialogue: "…Beyond this mist, there must be an answer." spoken in a low, resolute voice; Background Music: Starting with quiet and mysterious Japanese instrumental music, then powerful and grand strings and shakuhachi join in after the dialogue, and as the style transitions to watercolor, the music also shifts to a more fantastical and ethereal sound}
Camera: Starting with a medium shot from behind the samurai (realistic texture), the camera slowly moves forward, allowing a slight glimpse of the samurai's profile. As the samurai slowly re-positions his sword, the camera slowly circles to his side, capturing a close-up of his stern expression and determined gaze, emphasizing the blade's gleam. As the samurai takes his first step, the camera slowly pans up from his feet, capturing the entire screen as it transforms into a watercolor style. During the dialogue, the camera remains fixed on a close-up of the samurai's face, emphasizing the change in his expression. After the dialogue, the camera pans to follow the samurai's gaze towards the five-story pagoda in the depths of the mist, capturing the flowing mist, while also capturing his strong forward strides from a low angle. In sync with this panning motion, the background pagoda and mountains also appear to flow as if painted with watercolor brushstrokes.
Style: Initially high-definition realistic, seamlessly transitioning to watercolor, ink wash, and Japanese painting styles as the video progresses. Bushido aesthetics, profound and majestic atmosphere, a style merging the gravitas of a period drama film with art animation.
---
これが今回で一番長いプロンプトです。
思ってたスタイル変更ではなく、完全に違うシーンが差し込まれるようになってしまいました。
これはこれで面白いのですが、思った通りのものが出力できていないままです。
I2V生成物6：和風5
きっとプロンプトが長すぎたに違いない。
シンプルに纏めて、セリフも日本語に戻し、スタイル変更に特化させようと思ったら、予想外のところで想像と違う出来上がりになったのがこちら。
https://www.seaart.ai/ja/explore/detail/d1170ste878c73c8vkpg
プロンプトはこちら
---
Action: The samurai quietly surveys his surroundings, takes a deep breath, and re-positions his katana. His eyes are fixed on the pagoda. The next moment, he begins to run towards the five-story pagoda. Simultaneously with this "running" action, the entire video gradually loses its realistic texture, turns monochrome, and finally transitions into a faint, brush-stroked expression reminiscent of ink wash painting.Audio: {Environmental Sound Effects: Distant bird calls, sound of cold wind blowing quietly, irregular dripping water sounds, strong footsteps of the samurai, subtle metallic sounds of the sword, breathing; Dialogue: "……この霧の先に、必ず答えがあるはずだ." spoken in a low, resolute voice; Background Music: Quiet and mysterious Japanese instrumental music, centered on shakuhachi and strings, with a growing sense of rhythm as he starts running, and shifting to an ethereal sound with the ink wash style transition}Camera: Fixed in a medium shot from behind the samurai. As the samurai starts running, he moves slightly within the frame, but the camera angle itself does not change significantly. During the dialogue, the focus remains on the samurai's back and slightly visible profile, maintaining a composition that conveys his resolve.Style: Initially high-definition realistic. As the video progresses, the visual expression seamlessly transforms from realistic to monochrome, then to an ink wash painting style, creating a "painting-like motion." Bushido aesthetics, profound and majestic atmosphere, weighty like a period drama film.
---
「女性ボーカルで歌え」なんて指示は全くしていないのですが、なぜか侍にしゃべらせるはずだった（……この霧の先に、必ず答えがあるはずだ）がアニメOPのような曲になりました。
ただ、スタイルの変更自体はうまくいっている気がします。もっと完全に水墨画のようにしたかったのですが。
T2V生成物6：アニメOP風
曲が出力されるなら、アニメOPも作れるんじゃないか？と思ってT2Vに戻って生成したのが、この記事のカバーで使っている動画になります。
ちなみにこれ、以前のSeaArtのイベントで「冥界探検部」というものがあったのですが、その時にエントリーして作った「獄卒ヘブン」という設定からOPを作っています。
https://www.seaart.ai/ja/explore/detail/d126nbte878c73a55j40
プロンプトはこちら
---
A vibrant, dynamic, and clean Japanese anime art style explodes onto the screen, drenched in electric blues, purples, and neon pinks. A futuristic, neon-lit infernal stage, adorned with glowing magma channels and gothic-cyber elements, pulsates with electric energy. An incredibly upbeat and catchy J-Pop idol anthem blasts, driven by driving synthesizers and powerful drums. Clear, high-energy female vocals sing exhilarating Japanese lyrics, promising liberation and dreams: "胸に響くリズム刻んで、地獄の扉開け放て！光り輝く未来へ、私達が導くよ！" ("Carve the rhythm in your heart, throw open the gates of Hell! Towards a shining future, we will lead the way!"). The overwhelming roar of a m***ive, ecstatic crowd cheering and waving glowing infernal-pink lightsticks fills the audio mix. Center stage, Akari, the charismatic leader with twin red pigtails and sharp red horns, belts out the chorus, her infectious smile radiating pure joy as she executes powerful, gravity-defying leaps. Beside her, Yami, cool and elegant with flowing blue hair and subtle blue horns, performs precise, captivating moves, her gaze sharp and intense, radiating a calm allure. The camera dynamically sweeps across the stage, emphasizing their synchronized power and dazzling stage presence.
Suddenly, the scene shifts with a crisp cut. The camera frames the profile of a young male protagonist, clad in modern human clothes, standing in a dim, dusty corner of this bizarre Hell. His eyes, wide and sparkling with unadulterated awe and burgeoning hope, are fixed intently on Akari and Yami on stage, the vibrant stage lights reflecting brightly in his dilated pupils.
The camera then dramatically pulls back, revealing the entire, wildly energized infernal stage and the vast, cheering demon audience—a chaotic sea of glowing lightsticks. Akari and Yami are captured at the *** of their performance, their voices soaring in perfect harmony against the dazzling backdrop of pyro and light effects. As the music reaches a powerful, triumphant crescendo, a bold, stylized ***le card, shimmering with fiery effects and bright neon glows, flashes across the screen in crisp Japanese kanji: 「獄卒ヘブン」 (Gokukoku Heaven), before fading out as the idols hit a final, iconic, power pose, leaving the viewer with an overwhelming sense of excitement and adventure.
---
もともとの自分の設定では5人のキャラクターがいるのですが、全員を指示するとプロンプトが長くなりそうだったので2人のキャラクター設定だけを入れています（Center stage, Akari, the charismatic leader with twin red pigtails and sharp red horns, belts out the chorus, her infectious smile radiating pure joy as she executes powerful, gravity-defying leaps. Beside her, Yami, cool and elegant with flowing blue hair and subtle blue horns, performs precise, captivating moves, her gaze sharp and intense, radiating a calm allure. ）が、そのキャラクターはちゃんと出力されていて驚きました。なぜか記述していない3人目（動画左側の子）が生成されたのは、2人だとバランスが悪いと判断されたか？
それと、舞台風のT2Vを作った時にも思ったのですが「最初に生成されたシーンを眺めている人の横顔のカット」（Suddenly, the scene shifts with a crisp cut. The camera frames the profile of a young male protagonist, clad in modern human clothes, standing in a dim, dusty corner of this bizarre Hell. His eyes, wide and sparkling with unadulterated awe and burgeoning hope, are fixed intently on Akari and Yami on stage, the vibrant stage lights reflecting brightly in his dilated pupils.）というのは生成が難しいのかな？と思いました。指示が悪いだけかもしれませんが。
歌詞の日本語（胸に響くリズム刻んで、地獄の扉開け放て！光り輝く未来へ、私達が導くよ）と、日本語タイトルロゴ（flashes across the screen in crisp Japanese kanji: 「獄卒ヘブン」 (Gokukoku Heaven)）はうまくいってません。しかし、全言語対応してくれというのは贅沢ですね。
このように、コストは高いですが出来上がりは想像以上のものとなっています。
ぜひとも一度は使ってみて、驚きを感じてほしいです。
注意点1
なぜかわかりませんが、SF風動画を作る時に作った、どうしても生成がエラーになり成功しなかったプロンプトを載せておきます。
---
Scene: Vast cosmic space, countless stars ***ling, a colossal nebula stretching in the distance. Spaceship debris drifts. Inside a tense spaceship ***pit, numerous monitors flicker, and instruments glow with red warning lights. Characters: Multiple futuristic combat spaceships (sharp design, blue engine glow, some showing damage and wear), a male pilot in his 30s (tanned skin, resolute facial features, sweating profusely with disheveled hair, eyes showing strong determination and urgency, wearing a functional dark gray pilot suit with a communication wrist device). Actions: Numerous spaceships engage in high-speed dogfights, powerful laser cannons firing, leading to m***ive chain explosions upon hitting targets. The male pilot intensely operates the joystick, his body swaying violently, grimacing and shouting. Light: The silent glow of stars in space, sharp flashes of laser fire, vibrant light from spaceship propulsion engines, dazzling flames and afterglow from m***ive explosions, flickering instrument lights in the ***pit, red flashing warning lights indicating emergency, and the monitor light illuminating the face. Audio: {environmental sound effects: silence of space, laser firing sounds, spaceship engine hums, metal creaking, m***ive explosion sounds, ship's warning alarms, ***pit creaks, pilot's heavy breathing; dialogue: "Male Pilot: ???? it! I can't hold out much longer! Requesting backup! Now!"; background music: epic sci-fi action BGM, tension-building and thrilling electronic music, heavy percussion; emotions: despair, urgency, determination, tension, anger}. Camera: Dynamic shots tracking spaceships at high speed, transitioning from wide shots capturing the overall scene to close-ups of intense dogfights, with rapid cut transitions during laser exchanges. A quick pan into the ***pit for a tense close-up of the sweating male pilot's face, focusing on his hands operating the joystick, briefly interspersing a first-person view (POV) of space from the ***pit. Style: Sci-fi movie style, high-definition, dynamic, realistic, heavy VFX.
---
シーン：広大な宇宙空間で、無数の星々が瞬き、遠方には巨大な星雲が広がる。宇宙船の残骸が漂う。そして、緊迫した宇宙船のコックピット内、無数のモニターが点滅し、計器類が赤い警告灯を伴い光る。
キャラクター：多数の未来的な戦闘宇宙船（***なデザイン、青いエンジン光）、30代の男性パイロット（汗だくで髪が乱れている、目に強い決意と焦り、機能的なパイロットスーツ着用）。
動作：多数の宇宙船が高速でドッグファイトを繰り広げ、強力なレーザー砲が飛び交い、ターゲットに命中すると巨大な爆発が連鎖的に発生する。場面が切り替わり、男性パイロットが操縦桿を激しく操作し、体を大きく揺らしながら、叫び声をあげる。
光：宇宙空間の星々の静かな光、レーザー砲の鋭い閃光、宇宙船の推進エンジンが放つ鮮やかな光、巨大な爆発によるまばゆい炎と残光、コックピット内の計器の明滅光と緊急事態を示す赤い点滅。
音声：{環境音効：宇宙空間の静寂、レーザーの射撃音、宇宙船のエンジン音、金属の軋む音、巨大な爆発音、船内の警告音、コックピットの軋む音、男性の荒い呼吸音；対話：“男性パイロット：くそっ！このままじゃ持たない！援護を頼む！”；背景音楽：壮大なSFアクションBGM、緊張感とスリルを煽る電子音楽；感情：絶望、焦燥、決意、緊迫感}。
カメラ：高速で宇宙船を追跡するダイナミックなショット、広角で全体像を捉えるショットから激しくドッグファイトする宇宙船のクローズアップへ、レーザーが飛び交う場面では素早いカット切り替え。コックピット内の男性パイロットの顔に緊迫したクローズアップ、操縦桿を操作する手元にフォーカス、一人称視点（POV）での操縦席からの宇宙の眺め。シーン間は素早いカットで場面転換。
スタイル：SF映画風、高精細、ダイナミック、リアル。
---
英語、日本語、ともにエラーでした。
どこかにNGワードが含まれていたのかもしれませんが、ちょっと判断がつきませんでした。
注意点2
何度も言いますが、生成コストがかなり高いです。プロンプトの入力をミスったらかなりもったいないので注意しましょう。
和風をI2Vしている時、和風の画像を入力したのにプロンプトの入力をしないまま生成ボタンを押してしまい、もともとのサンプルプロンプト（Desolate post-apocalyptic wasteland, ruins of skyscrapers looming through sandstorms, abandoned machinery and rusted metal ***tered across barren earth, sky displaying eerie orange-red hue, a solitary signal tower flickering red light among ruins, a figure in protective gear standing on high ground, she removes her helmet, saying hoarsely: "They can delete my memories, but they cannot erase the truth in my heart"）のまま和風画像を生成してしまいました。
そうすると、ガビガビ音声のこんな感じのものが出来上がりましたので、特にI2Vを試すときはプロンプトがちゃんと画像にあっているかを確認しましょう。
