FilmLune · création et galerie IA
Prompts sélectionnés | Créez facilement des images et vidéos IA
Comprendre FilmLune
FilmLune réunit un générateur d’images IA et une galerie d’exemples image et vidéo. Chaque fiche présente le résultat, le prompt disponible, le modèle déclaré et une méthode à adapter.
Explorer tous les promptsUne méthode lisible
Deux points d’entrée
Partez d’un exemple image pour travailler une composition fixe, ou d’un exemple vidéo pour étudier le mouvement, les entrées et la durée. Chaque bibliothèque renvoie vers les mêmes fiches canoniques.
Repères utiles
Les modèles et réglages décrivent la révision visible : FilmLune ne remplace pas silencieusement le modèle pendant le transfert. Une œuvre externe conserve son attribution et son lien source sur sa fiche.
Seulement lorsque les droits de la révision le permettent. La fiche conserve alors le prompt et sa provenance ; une action indisponible n’est pas remplacée par un faux contenu.
Non. Une version de modèle, une entrée ou un réglage peuvent modifier le rendu. Utilisez la méthode comme point de départ et vérifiez les paramètres avant de générer.
SpringWangMiniMax H3 · estimationPROMPT
Create one final 16:9 horizontal anime character design sheet for a single adult female character.
The image should present 5 views in total, arranged cleanly in one wide composition:
1) one enlarged front-facing upper-body close-up portrait
2) one front full-body view
3) one left side full-body view
4) one back full-body view
5) one right side full-body view
The upper-body close-up should be noticeably larger than the other views, showing the character from the head down to the upper waist or mid-torso. The face must be clear and refined, with enough detail for later close-up video generation.
The character is an elegant adult anime woman with pale skin, sharp crimson-red eyes, and a calm, serious, slightly cold expression. She has straight black hair with soft uneven bangs across the forehead and long side locks framing the face. Her hair is long, sleek, and straight.
She wears a gold headband, a gold rose-shaped hair ornament on one side of the head, a long thin gold hairpin-like ornament, and gold dangling earrings.
Her outfit is a sleek, fitted black dress with an elegant, refined, slightly assassin-like look. The upper body has a strap-based design: a black bandeau-like chest section combined with multiple narrow black straps crossing around the upper chest, shoulders, and neck. Around the neck is a black choker-like strap structure with a small gold circular ornament at the front.
The back design is important: the upper back is exposed, with a vertical black strap descending from the back of the neck and additional black straps crossing into an X shape across the back.
She wears long black opera gloves extending above the elbows. The gloves are fingerless, exposing the fingers, and her fingernails are red.
She wears black thigh-high stockings. At the top of the stockings, include a distinctive strap-and-cutout pattern: two horizontal exposed-skin bands separated by black vertical connectors. Preserve this detail clearly.
The skirt is black with high slits and a dark red inner lining visible inside the split panels. She wears elegant black high heels. The overall outfit should feel stylish, clean, and iconic, with no unnecessary extra accessories.
All 4 full-body turnaround views must show the exact same character with matching proportions, hairstyle, clothing structure, accessories, and colors. The side views must be true profile views, not 3/4 views. The pose should be neutral and natural, with the arms slightly away from the body so the costume design remains visible.
Use a clean character model sheet presentation with consistent spacing. The full-body figures must remain fully visible from head to toe. The close-up portrait should be integrated neatly into the same sheet, not separated into a different scene.
Style: high-quality Japanese anime character sheet, polished cel shading, crisp line art, elegant rendering, subtle highlights, clean shadows, refined proportions, attractive and highly readable design. Not photorealistic, not chibi, not 3D toy style, not sketchy.
Lighting should be bright, neutral, and even, so the black hair and black clothing remain clearly visible. The background should be plain light beige, off-white, or light gray, simple and clean.
No text, no labels, no watermark, no UI, no extra characters, no weapons, no collage effect.Use the provided character reference image as the sole source for the character’s identity, anatomy, outfit, colors, materials, and visual style. Adapt body-region prompts to the character’s actual anatomy. If a named body feature does not literally exist, replace it with the closest equivalent existing feature. Do not invent anatomy, props, or accessories. Choose the most recognizable existing details as iconic inserts. Rigid parts may rotate; flexible materials may bend, stretch, and react naturally.
Create a 15-second character detail montage composed of exactly 20 shots, connected only by hard cuts.
Setting: an empty seamless warm-white studio background. Use soft directional lighting that reveals surfaces and material response clearly.
Framing rule: shots 1 through 19 must be tightly cropped body-detail shots and rear-angle shots only. Keep the entire face, eyes, and facial silhouette completely out of frame at all times, including in reflections. Shot 20 is the first and only face reveal. No full-body shot at any point.
Motion structure:
Build the sequence around three connected bursts of action:
1) a sharp head turn supported by body tension,
2) a coiled arm action that unfolds into a forceful sweeping release with torso twist,
3) a single short step into a firm pivot, turning into a sudden direct look to camera.
Across cuts, preserve body state continuity rather than resetting to neutral. Hair, clothing, and accessories should react naturally to motion. Keep the character active throughout, with clear joint articulation, weight shift, and momentum.
Editorial rhythm:
Jump between distant body regions. Mix brief impact shots with slightly longer moving detail reads. Use one sudden camera movement per shot, except for locked-camera inserts in shots 4 and 11, where the character provides the speed. Vary height, angle, and screen direction. Briefly brake the motion while the subject is still moving so each detail remains readable. Use local whip motions while staying tightly cropped. Use hard cuts for changes of body region. Link selected cuts through similar shapes or motion vectors, without morphing. Real-time motion only: no slow motion, no freeze frames.
0–3 seconds, shots 1–5. A concealed head turn triggers the first burst:
1. Extreme close-up of the upper back: a sudden impact-like push catches the surface shifting as the head turn begins, with the head fully cropped above frame.
2. Fingers: a brief downward snap catches them curling into a fist as the body braces.
3. Rear three-quarter shoulder detail: a tight camera arc catches the shoulder drawing back behind the turning head.
4. Locked, floor-height foot detail: the support foot stabilizes while the lifted heel presses down firmly without taking a step.
5. Back of the neck detail: a brief upward tilt catches the nape stretching during the turn; stop below the back of the head, keeping the jaw and face excluded.
3–9 seconds, shots 6–13. The coiled arm action unfolds into a forceful sweep:
6. An iconic upper-body detail: a quick diagonal slide catches an existing seam, joint, or distinctive contour responding as the arm pulls inward.
7. Elbow: a sharp lateral track catches the joint tightening as the arm coils.
8. Support knee: hard match-cut from the bent elbow to the knee at a similar angle; a brief camera drop catches the leg compressing in the same preparatory action.
9. Wrist: a quick shallow orbit catches it rotating outward to initiate the release.
10. Upper chest: a brief push-in catches the chest opening as the arm drives outward beyond the crop. Let this moving surface read slightly longer.
11. Locked hand detail: the fingers snap open and the hand sweeps rapidly across the fixed frame. Capture one clear readable stage of the fingers extended before they leave frame.
12. Upper back: a sharp lateral slide continues the screen direction as the shoulders roll through the sweep and the torso twists.
13. A distinctive torso or waist detail: a close diagonal move catches the body’s response to the twist, such as existing folds tightening or an attached detail swinging. Give the material response a clear moving read.
9–13 seconds, shots 14–19. A short step and pivot build toward the reveal:
14. Thigh of the other leg: a sudden low tracking move catches it driving a single short step out of the twist.
15. Advancing foot: a brief downward whip catches it landing, with clear floor contact.
16. Ankle and lower-leg detail: a quick lateral slide catches the landing leg accepting weight as the ankle flexes.
17. Waist: a tight orbit catches the torso rotating above a firm planted-foot pivot. The support foot pivots rather than sliding across the floor.
18. Back-of-head contour: a brief lateral whip catches the return turn beginning from behind. Keep only the rear contour in frame, never the facial outline.
19. An iconic shoulder detail: a sharp diagonal slide catches the follow-through beneath the turning head, fully cropped above frame. Cut at the peak of this motion.
13–15 seconds, shot 20. First and only face reveal:
20. Hard cut to a frontal facial close-up as the character sharply completes the turn toward camera. Use a brief explosive push-in that clearly reveals the full face for the first time, then stops before cropping away the facial features. Direct eye contact lands cleanly, while hair, clothing, or natural body follow-through remain active. Hold through the ending with the whole face readable. End on living motion, not a frozen pose and not another cut.
Do not add props, scenery, text, graphic overlays, or split screen.Ce guide couvre un flux en deux temps : d'abord la création d'une fiche de personnage anime horizontale présentant un gros plan du buste et quatre vues de rotation complètes, puis la transformation de cette fiche en référence pour un montage vidéo de 15 secondes composé de vingt plans serrés sur les détails du corps, qui ne révèle le visage qu'au tout dernier plan.
Prompts et exemples vidéoL'objectif du visuel source est de produire une planche de modèle lisible et cohérente d'une femme adulte élégante : peau pâle, yeux rouge cramoisi, cheveux noirs raides avec frange inégale, bandeau doré, ornement en rose dorée, boucles d'oreilles pendantes, et une robe noire ajustée à brides croisées avec des découpes distinctives au sommet des bas. Le fond uni clair et un éclairage neutre garantissent que le noir des cheveux et des vêtements reste bien lisible. La vidéo, en format vertical, exploite ensuite cette planche comme unique source d'identité : les dix-neuf premiers plans restent recadrés sur des régions du corps ou des angles arrière, sans jamais montrer le visage, tandis que le vingtième plan livre la seule révélation frontale du visage en contact visuel direct.
Étape 1 — générer la planche avec GPT Image 2.5 en 16:9. Décrire les cinq vues : un portrait buste agrandi descendant jusqu'à la taille, plus les vues de face, de profil gauche, de dos et de profil droit en pied. Exiger des profils vrais et non des trois-quarts, des bras légèrement écartés du corps, des proportions et accessoires strictement identiques entre les vues, et l'intégration du portrait dans la même composition. Préciser le style : cel shading soigné, traits nets, rendu élégant, fond beige clair ou gris pâle, sans texte ni personnage supplémentaire. Étape 2 — utiliser cette image comme référence exclusive pour la vidéo avec MiniMax H3 en 4:7, durée 15 secondes. Structurer le montage en vingt plans reliés par coupes franches, répartis en trois poussées d'action : une rotation de tête brusque (plans 1 à 5), une action de bras enroulée qui se déploie en un geste ample avec torsion du buste (plans 6 à 13), puis un pas court suivi d'un pivot vers la caméra (plans 14 à 19), avant le plan 20, seul plan frontal. Maintenir la continuité de l'état du corps entre les coupes, varier hauteurs et angles, garder chaque détail lisible par un léger freinage, réserver un mouvement de caméra verrouillée aux plans 4 et 11, et terminer sur un mouvement vivant plutôt qu'une pose figée.
L'image est fixe en 16:9 et la vidéo en 4:7 pour 15 secondes : les deux formats ne coïncident pas, ce qui impose des recadrages par région du corps. Un montage de vingt plans en quinze secondes donne une moyenne de 0,75 seconde par plan : certains plans très brefs peuvent perdre en lisibilité selon le modèle. La cohérence parfaite des proportions, des accessoires et de la structure des vêtements entre les cinq vues n'est pas garantie au premier essai et peut nécessiter plusieurs régénérations. La vidéo ne comporte pas de boucle seamless : elle se termine sur le plan de révélation du visage. Les mouvements complexes (pivot sur pied planté, continuité de vecteurs entre coupes) restent sensibles à l'interprétation du modèle vidéo.