She turns slowly from the mirror toward the camera, lets the loose strap slide a little further down her arm and reaches back to the half-open zip, holding the lens with a slow, knowing smile. She murmurs, "Help me with this?" and bites her lip. The vanity bulbs glow warm, the camera pushes in gently.
He rises from the marble bench as the steam curls around him, runs a hand back through his wet hair and walks slowly toward the camera, water running down his chest, a slow confident smile. Hiss of steam, dripping water echoing off marble. The camera holds low and tilts up as he comes closer.
The woman in the white shirt draws her partner closer by the strap of her camisole; they share a slow, lingering kiss and smile against each other's lips while the morning light shifts across the sheets. Soft laughter, birdsong outside the window.
The woman from the first image stands in the fitting room from the third image, wearing the emerald silk robe from the second image over a black satin slip. She ties the sash loosely, turns to the gilded mirror, lets the robe slide off one shoulder and glances back at the camera with a slow smile. Soft rustle of silk, a lamp hums.
Night, a candlelit bedroom doorway. The woman from the reference image, in a black satin slip, leans back against the door, turns the key in the lock behind her, looks into the camera and whispers in the voice from the reference audio: "Lock the door. Nobody is coming up tonight... and I am not in any hurry." A slow, sly smile at the end. Close-up, shallow depth of field, flickering candlelight.
Inside a mirrored, brass-lined hotel elevator at night, the bearded man from the first image loosens the tie of the man from the second image, steps in close and backs him gently against the mirror; they hold each other's gaze and almost kiss as the doors begin to slide open. The elevator chimes softly.
In the boutique fitting room from the fourth image, the woman with long black hair from the first image tries on the emerald silk robe from the third image over a black satin slip and turns slowly in front of the gilded mirror. The woman with short curls from the fifth image ties the robe's sash for her. The bearded man from the second image, in his dark suit, leans in the doorway, smiles and says, "It's perfect. We're late." Both women laugh. Warm lamplight, a soft rustle of silk, slow push-in.
カメラに向かって一言（image-to-video、1080p）。 黒髪のウェーブヘアの女性が、黒いサテンのスリップドレスを着て、化粧用の電球が並ぶ三面鏡の前にいるフレームから。背中のジッパーは数センチ下りています。彼女はカメラに振り向き、緩んだストラップをさらに滑らせ、ジッパーに手を伸ばして「Help me with this?」と言い、唇を噛みます。顔、ドレス、部屋はフレームのままで、16:9 のフレームはちょうど 1920 x 1080 で返ってきました。
蒸気の中から歩いてくる（image-to-video、縦長）。 無精ひげの男性が白いタオル姿で大理石の浴場のベンチから立ち上がり、濡れた髪をかき上げて、カメラに向かって歩いてきます。9:16 のフレームは 720 x 1280 で返り、音声には蒸気のシューッという音と水滴の音が入っています。
借りた声で一言（reference-to-video、9:16、9秒）。 長い黒髪で赤い口紅の女性のポートレート1枚と、音声クリップ1本。ろうそくの灯る戸口で、彼女は鍵を回し、「Lock the door. Nobody is coming up tonight... and I am not in any hurry.」とささやきます。台詞はプロンプトに書いてあり、音声クリップは声を決めるだけです。
image-to-video はフレームに従います。ある16:9のフレームは、540pで 952 x 544、720pで 1268 x 724、1080pで 1906 x 1088、4kで 3804 x 2172 で返りました。別のフレームは1080pでちょうど 1920 x 1080、2:3 の縦長は540pで 586 x 880 でした。reference-to-video では、16:9 の540pが 960 x 528、9:16 の720pが 720 x 1280、1:1 の2kが 1920 x 1920 でした。クリップは指定した長さより100分の数秒ほど長くなります（5秒が5.04秒で返りました）。
クリップは短いシーンとして書きます。誰がいて、何を着て、どこにいて、クリップの間に何をして、ダブルクォートで囲んだどんな台詞を言い、どんな音が聞こえ、カメラがどう動くか。image-to-video では次の瞬間を書いてください。顔、服装、部屋はフレームにすでにあります。reference-to-video では、各参照を順番で指し（「the woman from the first image」「the robe from the second image」）、画像もその順番で送ります。登場人物は全員成人にしてください。
She turns slowly from the mirror toward the camera, lets the loose strap slide a little further down her arm and reaches back to the half-open zip, holding the lens with a slow, knowing smile. She murmurs, "Help me with this?" and bites her lip. The vanity bulbs glow warm, the camera pushes in gently.
The woman from the first image stands in the fitting room from the third image, wearing the emerald silk robe from the second image over a black satin slip. She ties the sash loosely, turns to the gilded mirror, lets the robe slide off one shoulder and glances back at the camera with a slow smile. Soft rustle of silk, a lamp hums.
Night, a candlelit bedroom doorway. The woman from the reference image, in a black satin slip, leans back against the door, turns the key in the lock behind her, looks into the camera and whispers in the voice from the reference audio: "Lock the door. Nobody is coming up tonight... and I am not in any hurry." A slow, sly smile at the end. Close-up, shallow depth of field, flickering candlelight.
Inside a mirrored, brass-lined hotel elevator at night, the bearded man from the first image loosens the tie of the man from the second image, steps in close and backs him gently against the mirror; they hold each other's gaze and almost kiss as the doors begin to slide open. The elevator chimes softly.
弱い例： 「参照音声の声と言葉でささやく」（音声がまったく出ませんでした）や、3人のキャストなのに服装が1人分しか書かれていないもの。強い例： 台詞そのものをダブルクォートで書き、そのあとに「in the voice from the reference audio」を続け、シーンの全員に服装を指定する。台詞はクリップ内で言える長さに収めてください。5秒のクリップでは短い質問1つ、9秒のクリップでは短い文3つが入りました。
