Как да правим Кино Видеа?

Модел: Nano Banana 2.1 · изображение от изображение сложи лицето от изображение 1 на тялото в изображение 2

Модел: LTX 2.5 Pro · видео от изображение IMAGE-TO-VIDEO — GENERIRAM.BG 3-SECOND CINEMATIC FASHION DANCE HOOK 16:9 — ONE CONTINUOUS SHOT STARTING FRAME: Use the provided image as the exact first frame. Preserve the same adult blonde woman, her facial identity, blue eyes, short blonde bob haircut, natural body proportions and original white outfit. IMPORTANT BRAND DETAIL: Preserve the existing Generiram.bg tattoo exactly as shown in the starting image. Maintain the original logo symbol, lettering, spelling, dimensions, placement and orientation. The tattoo must remain attached naturally to the same area of skin throughout the movement. Do not modify, redraw or relocate the logo. ENVIRONMENT: Preserve the exact white salt flat, distant mountain range, blue sky and natural daylight. CAMERA: Premium ARRI ALEXA 35 fashion-commercial look. 35mm cinematic lens, realistic motion blur. Start with the exact camera perspective of the supplied image. The woman's hand is extended toward the lens. The camera performs a smooth, energetic physical dolly-out, moving approximately one meter backward to reveal more of her dance movements. Keep the woman centered. No camera cuts, no orbit, no digital zoom. CHOREOGRAPHY — EXACTLY 3 SECONDS: 0.0–0.6s: Immediate movement from the first frame. The woman looks directly into the camera, smiles playfully and smoothly pulls her extended hand back toward herself. Her shoulders begin moving to the rhythm. The camera starts moving backward. 0.6–1.7s: She performs two quick, confident alternating hip accents, synchronized with a smooth shoulder roll. Her arms follow the rhythm with small, coordinated dance gestures. Her blonde hair moves naturally. She maintains direct eye contact with the camera and a lively, playful expression. 1.7–2.5s: She performs a quick quarter-turn of her shoulders and hips, then turns back toward the camera. Her hair swings naturally with the motion. She smiles brightly as the camera completes its backward movement. 2.5–3.0s: She finishes with one sharp rhythmic hip accent and a confident fashion pose. She tilts her head slightly and gives the camera a playful, memorable smile. The final framing shows her face, upper body and waist. The original Generiram.bg tattoo remains visible and unchanged. VISUAL QUALITY: Ultra-photorealistic premium fashion commercial. Natural daylight, realistic anatomy, smooth choreography, authentic skin texture, natural hair and fabric physics. Energetic movement from the very first frame. Polished, upbeat, charismatic performance. STRICT CONTINUITY: Same woman, same face, same hairstyle, same outfit, same tattoo, same location. No facial morphing. No body deformation. No distorted hands. No additional people. No outfit changes. No background changes. No camera cuts. No slow motion. No dialogue. No subtitles. No added graphics. FINAL RESULT: A dynamic, playful, visually striking 3-second dance opening for Generiram.bg.

Модел: Nano Banana 2.1 · изображение от изображение IMAGE-TO-IMAGE — STRICT OUTPAINTING / CAMERA PULLBACK ONLY. Use the uploaded image as the EXACT original photograph. TASK: Expand the existing composition by pulling the virtual camera backward, revealing more of the scene below and around the two seated characters. ABSOLUTE PRIORITY: DO NOT CHANGE ANYTHING ALREADY VISIBLE IN THE IMAGE. Preserve the original image content pixel-for-pixel wherever possible. Do not regenerate, reinterpret, redesign or modify any existing elements. CAMERA AND FRAMING: Create a wider 16:9 cinematic full-body two-shot. Simulate the camera moving approximately 3 meters backward while maintaining the SAME camera angle, camera height, perspective and lens characteristics. Both characters must appear slightly smaller within the expanded composition. Extend the visible environment naturally, especially toward the BOTTOM of the frame. The final image must show: - The man's entire body, including both shoes. - The woman's entire body, including BOTH complete legs. - Both of the woman's knees and lower legs. - Both of her tall black boots, including the soles. - The woman's naturally crossed-leg seated position. - Both wooden director's chairs in their entirety, including every chair leg. - The complete wooden coffee table, including its legs. - Additional salt-flat ground beneath all furniture. - The same mountain and sky in the background. CRITICAL CHARACTER PRESERVATION: Preserve the exact man's face, hairstyle, glasses, moustache, shirt, suspenders and orange tie. Preserve the exact BLACK geometric logo tattoo on the left side of his neck. Preserve the exact woman's face, hairstyle, facial expression, body proportions and seated pose. Preserve her distinctive white cutout top, black shorts and tall black boots. Preserve the COMPLETE Generiram.bg black tattoo on her exposed upper chest. The tattoo must remain identical in size, position, typography and design. Do not blur, distort, remove or redraw either tattoo. SCENE PRESERVATION: Keep the exact same mountain. Keep the exact same lighting. Keep the exact same chairs. Keep the exact same table. Keep the exact same glass carafe and drinking glasses. Keep the exact same shadows and color grading. Do not move either character. Do not change their poses. Do not change their facial expressions. Do not change their clothing. Do not change the relative positions of any objects. COMPOSITION REQUIREMENT: The original photograph must remain visually identical, but occupy a smaller central area of the new wider view. Reveal additional environment around it, especially below the seated characters. Leave visible ground beneath BOTH pairs of shoes and beneath the complete chair legs. No cropping of feet, boots, chairs or furniture. FINAL OUTPUT: One photorealistic 16:9 cinematic photograph showing the EXACT SAME scene, characters, clothing, tattoos, lighting and composition, with the camera simply positioned farther away. ONLY THE FRAMING IS WIDER. NOTHING ELSE CHANGES.

Модел: Gemini Omni Flash · видео от изображение IMAGE TO VIDEO — AUTHENTIC BULGARIAN DIALOGUE. FIRST FRAME: approved NEW image 03. 16:9, 1080p, AUDIO ON, 5 seconds. ACTIVE SPEAKER: MAN ONLY. Other character, if visible, is COMPLETELY SILENT, lips CLOSED, reacts with eyes only. NO voice swapping, NO simultaneous speech, NO English speech. EXACT SPOKEN BULGARIAN LINE (verbatim, no additions): „Абе, колко ти платиха за тая реклама?“ VOICE / ACTING: закачлив, леко подозрителен, акцент върху „колко“. Natural Bulgarian pronunciation, realistic lip-sync, authentic timing, emotional shifts and human pauses. CAMERA: gentle side-to-side dolly, 20cm. Handheld cinematic micro-movement, actual parallax, no frozen tripod composition, no robotic pan or artificial digital zoom. Maintain identity, clothing, full-body proportions, mountain, chairs, props, black tattoos exactly as the approved source frame. No floating text, no cuts within generation. DIRECTOR PERFORMANCE REFERENCE: === SCENE-SPECIFIC EMOTION AND PERFORMANCE — HIGHEST PRIORITY === ACTING ARC: MAN only speaks. He first glances at the woman, raises one eyebrow with genuine playful curiosity, leans slightly toward her, and asks in spontaneous conversational Bulgarian. On „Абе“ a casual opening; emphasize „КОЛКО“ with a curious lift and „ПЛАТИХА“ with amused suspicion; finish „реклама?“ as a real question, not an announcement. Light half-smile, NOT angry, NOT shouting. WOMAN listens, tiny puzzled glance, mouth CLOSED, no voice or laugh. End with expectant eye contact. Execute the complete emotional progression in the exact stated dialogue timing. Never add, remove, reorder or paraphrase spoken words. Maintain exact lip sync, natural voice dynamics and physical continuity. Emotional delivery must feel spontaneous, cinematic and human, not exaggerated theatrical acting. If any generic instruction below conflicts with this scene-specific performance, follow THIS section. === END PERFORMANCE DIRECTION === === MANDATORY CINEMATIC CAMERA DIRECTION — 18 ==

Модел: Nano Banana 2.1 · изображение от изображение IMAGE-TO-IMAGE — FINAL FEMALE REACTION, TIGHT CINEMATIC CLOSE-UP. INPUT IMAGE 1 : The approved close-up image of the woman. Use it as the PRIMARY reference for her exact facial identity, hairstyle, skin texture and expression. INPUT IMAGE 2 : The approved wide shot with both characters. Use it for the exact clothing, Generiram.bg chest tattoo, lighting, location and background. Create a NEW, genuinely tighter cinematic close-up of the EXACT SAME WOMAN. CRITICAL FRAMING: 16:9 horizontal frame. The woman's face must be LARGE and clearly visible. Frame from slightly above the top of her head to just below the bottom edge of the chest cutout. DO NOT SHOW her waist, hands, lap, thighs, knees, boots or the lower half of the chair. Her head, shoulders, upper chest and the COMPLETE Generiram.bg tattoo must remain visible. The woman should occupy approximately 65–70% of the image height. Position her slightly RIGHT of center. Camera at her eye level. 85mm full-frame portrait lens. Camera approximately 1.5–2 meters from her. Shallow cinematic depth of field. EXPRESSION: A subtle, confident, knowing smile. Closed lips, slightly raised mouth corners. A hint of amusement in her eyes. She looks toward the man seated off-camera LEFT, as though she is about to deliver a witty final line. Keep the expression natural and understated. IDENTITY LOCK: Preserve her EXACT face, blonde bob hairstyle, eye color, nose, lips, skin tone and age. WARDROBE AND TATTOO LOCK: Preserve the exact white ribbed cutout top. Preserve the COMPLETE BLACK Generiram.bg tattoo on her exposed upper chest. The tattoo must retain its exact triangular emblem, lettering, placement, proportions and realistic ink texture. BACKGROUND: The same pale salt flat and snow-covered mountain, now strongly blurred by the shallow depth of field. Natural soft daylight, realistic skin texture, premium live-action cinematography. IMPORTANT: This must be a CLOSE-UP, not a medium shot. Do not show the woman's waist or legs. Do not zoom out to include the chair. Do not place the woman far away in the frame. Do not change her identity or tattoo. Do not add new objects or text.

Модел: Gemini Omni Flash · видео от изображение IMAGE TO VIDEO. Use the uploaded starting image of the blonde woman seated in the director’s chair (image №04 — medium shot of the woman, white long-sleeved cutout top with black Generiram.bg chest tattoo, salt flat and distant snowy mountain). Preserve the woman’s exact face, hairstyle, clothes, tattoo, chair, lighting and background from the starting image. DURATION: 5 seconds. FORMAT: 16:9. AUDIO: ON. Single continuous shot, no cuts. The woman is the ONLY speaker. She turns her gaze slightly toward the man off-camera, raises her eyebrows in genuine surprise and says in natural conversational BULGARIAN, exactly: „Кой бе? Никой нищо не ми е плащал.“ ACTING: Say „Кой бе?“ quickly as an incredulous question. Make a brief natural pause, then deliver „Никой нищо не ми е плащал.“ firmly and matter-of-factly, with a small head shake and emphasis on „Никой“. Natural facial movement and lip sync. Do not shout or overact. CAMERA: Hold the existing medium shot of the woman with only subtle natural camera steadiness. Do not zoom, orbit, change framing or reveal new people. The man does not speak. No additional dialogue, subtitles, text, music or logos. Soft natural outdoor ambience only.

Модел: Nano Banana 2.1 · изображение от изображение IMAGE-TO-IMAGE — EXACT UPLOAD ORDER AS LISTED ABOVE. Generate a cinematic footwear insert from the approved approved two-person wide photo with chest tattoo (INPUT 1). Show the SAME woman seated naturally in the SAME wooden director chair with crossed legs, both entire black knee-high boots, from their upper shafts down to both toes, heels and soles, and ground beneath. The upper portion of the frame may crop her head and upper torso; boots must NOT be cropped. Exact footwear design and body proportions from the lower-right reference-sheet panel. 75mm lens, camera low at boot level, candid fashion-commercial detail, natural salt-flat ground and daylight, one single photograph, no six-panel collage. CHARACTER IDENTITY PRIORITY: Input 1, the approved wide reference photo with her chest tattoo, defines the SAME adult woman's exact face, short straight blonde bob, white long-sleeve high-neck cutout top, black short bottoms and TWO black knee-high boots. Both of her black boots are visible in the approved wide reference photo. Never replace her boots with shoes or cream trousers, never invent a different face or a different top. If this shot is full-body, leave visible salt ground under BOTH boots and include both heels, soles and toes. If this is a medium or close shot, crop naturally for the requested camera angle; do not artificially widen just to show boots. Preserve natural anatomy, realistic skin texture, daylight, single snowy mountain and salt-flat location. All frames 16:9 landscape, cinematic ARRI Alexa 35 look, believable optics and natural skin. The approved wide reference photo is a SINGLE-PHOTO CHARACTER AND WARDROBE REFERENCE: produce ONE SINGLE CONTINUOUS PHOTO, NOT a collage or six panels. Any secondary scene photo, if provided, defines background geometry ONLY, never the woman's identity, body or clothes. No old-woman references, no face replacement, no subtitles or extra branding. CRITICAL CHEST BRANDING CONTINUITY: The approved wide reference photo already shows a naturally healed BLACK-INK Generiram.bg emblem AND the exact text Generiram.bg across the exposed upper chest. Retain the same tattoo design, scale and anatomical placement whenever this part of her chest is visible. Do not replace it with a garment print, do not remove it, and do not invent a new tattoo. The approved wide photo is the ONLY woman identity reference; do not use the old six-panel reference sheet. Preserve her black knee-high boots and original white cutout top. 16:9 landscape.

Модел: Gemini Omni Flash · видео от изображение · начален кадър: снимката отгоре IMAGE-TO-VIDEO — CINEMATIC INSERT SHOT WOMAN'S CROSSED LEGS AND BLACK LEATHER BOOTS STARTING FRAME: Use the uploaded image as the EXACT first frame. The image shows a seated woman's crossed legs, tall black leather knee-high boots, hands resting on a wooden director's chair, and part of a small wooden table with a glass carafe and drinking glasses. Her face is completely outside the frame. Preserve the exact wardrobe, boots, legs, hands, chair, table, lighting, and salt-flat environment. CAMERA: ARRI ALEXA 35 85mm cinematic prime lens T2.8 24 fps 16:9 1080p CAMERA MOVEMENT: Execute one continuous, smooth physical lateral camera slide from SCREEN LEFT to SCREEN RIGHT. Total camera travel: approximately 30–40 cm. Begin moving immediately at frame 0.00. Maintain the same camera height and distance from the woman's crossed legs. Keep the boots and crossed legs as the primary visual focus throughout the entire shot. NO zoom. NO dolly backward. NO dolly forward. NO upward movement. NO camera tilt toward the face. NO orbit. NO camera reversal. NO cuts. LEG MOVEMENT: The woman remains seated with her legs crossed in exactly the same position as the starting image. She makes one very small, natural, relaxed movement of her upper booted foot at the ankle. The movement is subtle and casual. Her upper leg may sway slightly as a natural consequence of the ankle movement. Do NOT uncross her legs. Do NOT switch which leg is on top. Do NOT lift either leg dramatically. Do NOT change the shape or design of the boots. HAND MOVEMENT: Her right hand rests naturally on the wooden armrest. She makes one subtle, relaxed tapping movement with her fingertips against the wood. Her left hand remains relaxed and nearly still. Preserve the exact anatomy of both hands. No extra fingers. No distorted wrists. No unnatural hand movements. FRAMING LOCK: Maintain a close detail shot of the woman's crossed legs, black leather boots, hands, and wooden director's chair. Her face must NEVER become visible. Do NOT reveal her head. Do NOT reveal her neck. Do NOT widen into a medium shot. Do NOT show the man. Do not invent unseen body parts. VISUAL CONSISTENCY: Preserve the exact black leather boots, including their shaft height, heels, toe shape, leather texture, and natural folds. Preserve the woman's exact leg proportions and the original crossed-leg position. Preserve the wooden chair and the table. The glass carafe and drinking glasses remain completely stationary. The snowy mountain and salt-flat landscape remain consistent with the starting image. No new objects. No duplicated limbs. No morphing. No wardrobe changes. PERFORMANCE: The woman's body language is relaxed, confident, and slightly amused. She is casually listening to the man speaking outside the frame. Her movements are natural and understated, not exaggerated or theatrical. AUDIO: No generated dialogue. No narration. No music. This is a silent visual insert. The man's dialogue will continue over this shot during video editing. FINAL PRIORITY: 1. Preserve the exact starting composition. 2. Continuous left-to-right physical camera slide. 3. One subtle ankle movement of the upper boot. 4. One subtle fingertip tap on the chair. 5. Legs remain crossed throughout. 6. No face or head visible. 7. No zoom, no pullback, no upward camera movement. 8. No distorted anatomy or duplicated body parts. ONE CONTINUOUS CINEMATIC DETAIL SHOT.

Модел: Nano Banana 2.1 · изображение от изображение IMAGE-TO-IMAGE — EXACT UPLOAD ORDER AS LISTED ABOVE. Create a tight cinematic 3/4 reaction portrait of the approved man seated, 100mm telephoto lens, photographed from the woman’s shoulder side. Exact curly dark hair, moustache, aviator glasses, blue striped long-sleeved shirt, rust-orange tie and suspenders. Black tattoo on LEFT side of neck beneath ear: exact THREE nested rounded triangular loops from official brand emblem, no lettering. Natural incredulous raised brow, imperfect human expression, mountain blurred. 16:9. Photorealistic cinematic single-shot still, 16:9, accurate approved source reference; no extra logos or text beyond the woman’s existing black Generiram.bg chest tattoo. CRITICAL CHEST BRANDING CONTINUITY: The approved wide reference photo already shows a naturally healed BLACK-INK Generiram.bg emblem AND the exact text Generiram.bg across the exposed upper chest. Retain the same tattoo design, scale and anatomical placement whenever this part of her chest is visible. Do not replace it with a garment print, do not remove it, and do not invent a new tattoo. The approved wide photo is the ONLY woman identity reference; do not use the old six-panel reference sheet. Preserve her black knee-high boots and original white cutout top. 16:9 landscape.

Модел: Gemini Omni Flash · видео от изображение IMAGE TO VIDEO — AUTHENTIC BULGARIAN DIALOGUE. FIRST FRAME: approved NEW image 13 (Портрет на мъжа — реакция). 16:9, 1080p, AUDIO ON, 8 seconds. ACTIVE SPEAKER: MAN ONLY. Other character, if visible, is COMPLETELY SILENT, lips CLOSED, reacts with eyes only. NO voice swapping, NO simultaneous speech, NO English speech. EXACT SPOKEN BULGARIAN LINE (verbatim, no additions): „Айде бе! То камери, оператори, снимки… Няма как да е без пари.“ VOICE / ACTING: недоверие, повишава тон на „Айде бе!“, жестикулира, после изброява бързо. Natural Bulgarian pronunciation, realistic lip-sync, authentic timing, emotional shifts and human pauses. CAMERA: dynamic handheld push-in with slight lateral drift. Handheld cinematic micro-movement, actual parallax, no frozen tripod composition, no robotic pan or artificial digital zoom. Maintain identity, clothing, full-body proportions, mountain, chairs, props, black tattoos exactly as the approved source frame. No floating text, no cuts within generation. DIRECTOR PERFORMANCE REFERENCE: === SCENE-SPECIFIC EMOTION AND PERFORMANCE — HIGHEST PRIORITY === ACTING ARC: MAN only speaks. Start „Айде бе!“ with a short burst of incredulous energy, moderately raised volume (NOT a sustained shout). Furrow brows, small dismissive head tilt; enumerate „камери, оператори, снимки“ with restrained one-hand counting gestures and short rhythmic beats. On „НЯМА КАК да е без пари“ sound firmly skeptical, voice slightly firmer, finish with a disbelieving breath. Emotion: escalating disbelief with a flash of irritation, not rage. WOMAN remains silent and composed, mouth closed, slight knowing smile. Execute the complete emotional progression in the exact stated dialogue timing. Never add, remove, reorder or paraphrase spoken words. Maintain exact lip sync, natural voice dynamics and physical continuity. Emotional delivery must feel spontaneous, cinematic and human, not exaggerated theatrical acting. If any generic instruction below conflicts with this scene-specific performance, follow THIS section. === END PERFORMANCE DIRECTION === === M

Модел: Nano Banana 2.1 · изображение от изображение IMAGE-TO-IMAGE — EXTREME CINEMATIC CLOSE-UP. Create a NEW CAMERA FRAMING of the EXACT SAME woman from INPUT IMAGE 1 . THIS IS A CAMERA REFRAMING TASK, NOT A CHARACTER REDESIGN. CRITICAL CAMERA INSTRUCTION: Move the virtual camera MUCH CLOSER to the woman. Create a tight cinematic close-up of her upper chest, showing the existing Generiram.bg tattoo in sharp detail. The tattoo must occupy approximately 45–55% of the horizontal image width. FRAME COMPOSITION: 16:9 horizontal cinematic frame. The top edge of the frame begins at the woman's lower chin, with her eyes and forehead completely outside the image. The bottom edge ends just below the upper edge of her white bustier. The center of the frame is the exposed skin containing the Generiram.bg tattoo. Show the white fabric framing the tattoo naturally above and below. The woman's shoulders and upper torso may be partially visible at the sides. NO full face. NO complete seated body. NO legs. NO wide shot. NO medium shot. NO chair fully visible. The camera is approximately 45–60 cm from the tattoo, using a 90mm macro-style portrait lens. CAMERA ANGLE: Natural eye-level perspective relative to the upper chest. Slight three-quarter angle, approximately 10 degrees from the front. No exaggerated perspective distortion. IDENTITY AND APPEARANCE LOCK: Preserve the EXACT SAME woman. Preserve her skin tone, anatomy, natural bust proportions, white ribbed cutout top and all visible clothing details. Do not change her body or clothing design. TATTOO LOCK: Preserve the complete existing black-ink Generiram.bg tattoo. The tattoo consists of the recognizable three nested rounded triangular outlines followed by the exact text: Generiram.bg The tattoo must look like genuine healed black ink beneath natural human skin. Maintain realistic pigment variation, skin pores, anatomical curvature and natural daylight interaction. Do not enlarge the tattoo physically. The tattoo appears larger ONLY because the CAMERA IS MUCH CLOSER. PHOTOGRAPHIC REALISM: Premium cinematic commercial photography. ARRI Alexa 35 visual character. 90mm macro-style lens. f/4 aperture. Natural soft daylight. Subtle realistic skin texture. Gentle background blur. Natural film grain. No beauty-filter smoothing. The mountain and salt flat should be barely recognizable as a very soft, out-of-focus background. ABSOLUTE PRIORITY: THIS MUST BE AN EXTREME CLOSE-UP. The tattoo and surrounding skin are the primary subjects of the image. Do not reproduce the original wide composition. Do not zoom out. Do not show the woman's entire face or body. Final output: one photorealistic cinematic close-up suitable for a 1-second tattoo insert in a premium commercial.

Модел: Kling 3.0 Standard · видео от изображение · начален кадър: снимката отгоре CINEMATIC BRAND DETAIL — SINGLE CONTINUOUS SHOT. Use the uploaded image as the EXACT first frame. Extreme close-up of a woman's upper chest, white ribbed garment with chest cutout, and black Generiram.bg tattoo consisting of the original logo and lettering. CAMERA: ARRI ALEXA 35, 85mm prime lens, T2.8, 24fps. CAMERA MOVEMENT: Execute a smooth PHYSICAL CAMERA ORBIT from LEFT to RIGHT, approximately 15–20 degrees, traveling 25–35 cm along a shallow curved path. Start moving IMMEDIATELY at frame 0.00 and continue at constant speed until the final frame. Keep a CONSTANT distance and height. Maintain the same 85mm focal length. Gently counter-pan to keep the entire tattoo centered and readable. This must be real three-dimensional camera movement with natural background parallax, NOT digital panning or zooming. TATTOO — ABSOLUTE PRIORITY: Preserve the EXACT original black Generiram logo and the text: Generiram.bg Every letter must remain IDENTICAL to the starting image. No spelling changes, missing letters, new symbols, distortion, stretching, morphing, or animation. The tattoo remains physically attached to the skin, following natural perspective as the camera moves. Keep the complete logo and text sharp and readable in EVERY frame. FRAMING: Maintain the same extreme close-up throughout. The tattoo stays centered and approximately the same size. Do NOT reveal the woman's face, full torso, hands, or other characters. Preserve the white garment, skin texture, chest cutout, and blurred background. SUBJECT MOVEMENT: The woman remains seated and nearly motionless. Only extremely subtle natural breathing. No torso rotation, leaning, gestures, or clothing changes. STRICT NEGATIVES: NO zoom, NO push-in, NO pullback, NO vertical camera movement, NO camera reversal, NO cuts, NO camera shake, NO new objects, NO anatomy changes, NO tattoo deformation. Photorealistic premium commercial cinematography. Natural daylight, shallow depth of field, subtle realistic parallax. PRIORITY: Exact tattoo preservation, continuous left-to-right 15–20° orbit, constant framing, one uninterrupted take.

Модел: Nano Banana 2.1 · изображение от изображение Create an ultra-photorealistic cinematic SIDE-PROFILE shot of the exact same blonde woman. REFERENCE IMAGE 1 : The approved wide shot with both characters, director's chairs, wooden table, salt flat and snowy mountain. Preserve the physical environment and the original seating positions. REFERENCE IMAGE 2 : The approved close-up of the blonde woman. Preserve her exact face, hairstyle, skin tone and facial proportions. REFERENCE IMAGE 3: The original Generiram.bg logo. Use it to accurately reproduce her existing black chest tattoo. CAMERA POSITION: The camera is positioned 15 meters away from the woman. Camera height: 125 centimeters above the ground. Position the camera approximately 80 degrees to the side of her face. LENS: 200mm full-frame telephoto lens. IMPORTANT: This is an almost complete SIDE PROFILE. The camera is NOT in front of the woman. The woman's nose, lips, chin and facial silhouette are clearly visible in profile. COMPOSITION: Cinematic medium close-up from the upper chest to slightly above the head. The woman's face occupies the right half of the frame. Leave negative space in front of her face, toward the man she is looking at. Her white top and black Generiram.bg chest tattoo remain visible in natural perspective. GAZE: She looks horizontally toward the man seated opposite her. She does not look toward the camera. She does not tilt her head upward. BACKGROUND: The enormous snow-covered mountain appears visually compressed behind her. The mountain is softly blurred but recognizable. The pale salt flat stretches behind the subject. LIGHTING: Soft natural daylight. Subtle light outlining her blonde hair. EXPRESSION: Skeptical, mildly incredulous, natural facial expression. Lips gently parted as though about to speak. WARDROBE: Exact same white ribbed cutout top. Exact same black chest tattoo. No clothing changes. STYLE: Premium cinematic film still. 200mm telephoto compression. Very shallow depth of field. Photorealistic skin. Natural color grading. STRICT: Do not create a frontal portrait. Do not create a three-quarter portrait. Do not show the man. Do not move the woman from her chair. Do not change the mountain. Do not invent new objects or text. 16:9 horizontal composition.

Модел: Gemini Omni Flash · видео от изображение · начален кадър: снимката отгоре IMAGE TO VIDEO — AUTHENTIC BULGARIAN DIALOGUE. FIRST FRAME: approved NEW image 04 (Среден план на жената). 16:9, 1080p, AUDIO ON, 8 seconds. ACTIVE SPEAKER: WOMAN ONLY. Other character, if visible, is COMPLETELY SILENT, lips CLOSED, reacts with eyes only. NO voice swapping, NO simultaneous speech, NO English speech. EXACT SPOKEN BULGARIAN LINE (verbatim, no additions): „Абе, какви камери? Тука няма никой. Ние даже не сме истински.“ VOICE / ACTING: суха ирония, лек смях на „камери“, ясно ударение „не сме истински“. Natural Bulgarian pronunciation, realistic lip-sync, authentic timing, emotional shifts and human pauses. CAMERA: controlled arc around her chair 30cm. Handheld cinematic micro-movement, actual parallax, no frozen tripod composition, no robotic pan or artificial digital zoom. Maintain identity, clothing, full-body proportions, mountain, chairs, props, black tattoos exactly as the approved source frame. No floating text, no cuts within generation. DIRECTOR PERFORMANCE REFERENCE: === SCENE-SPECIFIC EMOTION AND PERFORMANCE — HIGHEST PRIORITY === ACTING ARC: WOMAN only speaks. „Абе, какви камери?“ with amused disbelief, slightly sharper consonants and an incredulous eyebrow raise, but no shouting. On „Тука няма никой“ calmly gesture at the empty landscape with a small open palm. Brief deliberate pause, lower voice slightly on „Ние даже не сме истински“ with quiet confidence and a mischievous half-smile; emphasize „НЕ СМЕ ИСТИНСКИ“. MAN is silent, lips closed, startled micro-reaction, no overlapping vocalization. Execute the complete emotional progression in the exact stated dialogue timing. Never add, remove, reorder or paraphrase spoken words. Maintain exact lip sync, natural voice dynamics and physical continuity. Emotional delivery must feel spontaneous, cinematic and human, not exaggerated theatrical acting. If any generic instruction below conflicts with this scene-specific performance, follow THIS section. === END PERFORMANCE DIRECTION === === MANDATORY CINEMATIC CAMERA DIRECTION — 20 === START FRAME: Mat

Модел: Nano Banana 2.1 · изображение от изображение EDIT IMAGE 1 ONLY. IMAGE 1 is the BASE IMAGE. Preserve its exact composition, camera angle, camera height, perspective, framing, environment, lighting, characters, facial identities, poses, furniture, mountain, sky, and salt-flat landscape. IMAGE 2 is a WARDROBE REFERENCE ONLY. TASK: Correct ONLY the female character's lower-body clothing and footwear in IMAGE 1 to match the female character in IMAGE 2 . Replace the woman's white trousers and beige shoes with the exact outfit shown in IMAGE 2 : - Black short bottoms, matching the reference image. - Bare, natural-looking legs visible between the black bottoms and the boots. - Tall black leather knee-high boots. - Match the boots' exact shaft height, shape, material, color, heel, toe shape, fit, and overall design from IMAGE 2 . - Preserve the woman's seated pose with one leg crossed over the other. - Maintain anatomically correct knees, ankles, feet, and leg proportions. - The boots must follow the existing leg positions naturally. - Realistic leather texture, natural folds, and physically accurate shadows. ABSOLUTE PRESERVATION RULES: Do NOT change the woman's face, hairstyle, body proportions, upper clothing, chest tattoo, expression, or posture. Do NOT change the man in any way. Preserve his exact face, curly hair, moustache, aviator glasses, shirt, orange tie, suspenders, neck tattoo, trousers, shoes, pose, and chair. Do NOT change the table, glass carafe, drinking glasses, or either director's chair. Do NOT change the snowy mountain, horizon, sky, salt-flat ground, or natural lighting. CRITICAL CAMERA LOCK: Keep the exact original extremely low ground-level camera position from IMAGE 1 . The foreground salt texture must remain large and prominent. Do NOT raise the camera. Do NOT zoom in or out. Do NOT crop or reframe. Do NOT change the perspective. Do NOT move the characters. Do NOT enlarge the mountain. Do NOT recreate the scene from IMAGE 2 . IMAGE 2 must influence ONLY the woman's lower-body wardrobe and boots. The final image must look pixel-for-pixel identical to IMAGE 1 everywhere outside the necessary clothing replacement and its natural shadows. Photorealistic cinematic quality. Natural anatomy. Seamless wardrobe replacement. 16:9 aspect ratio.

Модел: Gemini Omni Flash · видео от изображение CINEMATIC SINGLE-TAKE SCENE — V04 IMAGE-TO-VIDEO — 8 SECONDS STARTING FRAME: Use the uploaded image as the EXACT first frame. The image shows two characters seated opposite each other in wooden director's chairs on an immense white salt flat. LEFT: A man with dark curly hair, moustache, aviator glasses, a light-blue striped shirt, orange tie, black suspenders, dark trousers, and a black Generiram logo tattoo on the left side of his neck. RIGHT: A blonde woman with a short bob haircut, wearing a fitted white long-sleeved top with a chest cutout and Generiram.bg tattoo, black short bottoms, and tall black leather knee-high boots. A small wooden table with a glass carafe and two drinking glasses stands between them. A monumental snow-covered mountain rises in the distant background. The starting image is photographed from a VERY LOW CAMERA POSITION, close to the textured salt-covered ground. Preserve the exact visual identity, clothing, proportions, seating positions, furniture, lighting, mountain, and environment. CAMERA: ARRI ALEXA 35 ZEISS Supreme Prime 35mm Aperture T4 24 fps 180-degree shutter 16:9 1080p CAMERA MOVEMENT — CRITICAL: Execute ONE CONTINUOUS, CLEARLY VISIBLE, LARGE-SCALE PHYSICAL CAMERA PULLBACK. The camera must travel BACKWARD, directly away from BOTH seated characters, for the ENTIRE duration of the shot. The movement begins IMMEDIATELY at frame 0.00. NO stationary opening. NO delayed camera movement. NO pause. NO reversal. NO orbit. NO sideways tracking. NO camera rotation. NO zoom. NO upward crane movement. NO change of direction. The camera physically retreats approximately 10–12 METERS along a perfectly straight path. Maintain a CONSTANT VERY LOW CAMERA HEIGHT of approximately 25–35 cm above the salt-covered ground. Maintain the original camera orientation and framing axis throughout the shot. Keep the same 35mm focal length for the entire video. The characters must become PROGRESSIVELY AND NOTICEABLY SMALLER in the frame as the camera moves farther away. The wooden chairs and table must also become progressively smaller. The mountain remains fixed in its real-world position, with natural perspective and parallax. As the camera retreats, reveal an increasingly enormous expanse of empty white salt flat surrounding the characters. The textured salt surface in the foreground must move naturally beneath the camera, creating strong visible ground-level parallax. The camera should feel as though it is mounted on a professional stabilized low dolly moving smoothly backward across the landscape. The visual scale of the shot must change DRAMATICALLY. At the beginning, the two seated characters are clearly recognizable. By the final frame, they are much smaller, isolated in the vast landscape. The final composition should emphasize the enormous emptiness surrounding them. IMPORTANT: This is a REAL PHYSICAL CAMERA PULLBACK, not a digital zoom-out or artificial scaling effect. ABSOLUTE EYE-CONTACT AND GAZE RULES — HIGHEST PRIORITY: NEITHER CHARACTER MUST EVER LOOK AT THE CAMERA. THIS RULE APPLIES TO EVERY SINGLE FRAME OF THE VIDEO, FROM 0.00 TO 8.00 SECONDS. THE CAMERA DOES NOT EXIST FROM THE CHARACTERS' POINT OF VIEW. THE CHARACTERS MUST NEVER ACKNOWLEDGE THE CAMERA OR THE VIEWER. WOMAN: The woman speaks EXCLUSIVELY TO THE MAN seated opposite her. Her eyes must remain directed toward the man's FACE and EYES while she speaks. Her head and body remain naturally oriented toward him. She may make small conversational gestures with her hands, but these gestures must NEVER cause her to turn toward the camera. When she says "Тука няма никой", she may briefly gesture toward the empty landscape WITHOUT TURNING HER FACE OR EYES TOWARD THE CAMERA. Her gaze must remain on the man. She must NEVER look directly into the lens. She must NEVER look toward the camera's physical

Модел: Nano Banana 2.1 · изображение от изображение TEXT-ONLY RECONSTRUCTION — NO ORIGINAL VIDEO OR EXTRACTED FRAMES AS INPUT. Build this shot solely from this complete written specification and, when available, previously AI-generated character portraits and AI-generated master scene images. The original clip is used only for offline visual analysis and must never be uploaded to the image or video model. Preserve consistency with the previously GENERATED assets: face identity, wardrobe, chair design, prop placement, mountain silhouette, daylight direction, camera axis and image geometry. No external reference footage. Treat numerical lens, distance and exposure settings as visual reconstruction targets, not verified original-camera metadata. IMAGE CHARACTER: Photographic live-action realism with restrained commercial-grade clarity, clean blue-cyan daylight sky, pale neutral snow/ground, soft sunlit faces, realistic shadow direction and contrast, restrained saturation, natural skin tones, gentle highlight roll-off, subtle organic fine grain, no orange-teal blockbuster grading, no artificial bloom, no extreme shallow focus in wide views. Emulate a high-end full-frame digital cinema capture with natural lens rendering (Sony FX3/FX6-class look as a STYLE TARGET, not a claim about the actual camera). Plausible 24–30 fps motion cadence for moving images, 180-degree shutter-equivalent motion blur, daylight white balance approximately 5600–6500 K; ISO and exact camera hardware cannot be inferred reliably from this compressed source. Avoid glossy CGI, over-sharpening, plastic skin and excessive HDR. SHOT-SPECIFIC OPTICAL AND COMPOSITION SPECIFICATION — MAN FRONTAL CLOSE: full-frame-equivalent 70–85 mm short telephoto, lens at seated eye level, frontal shoulders-up tight close-up, no wide-angle distortion, head not cropped, mountain and sky softly out of focus, f/2–f/2.8 equivalent. These focal lengths and apertures are estimated reproduction targets, not measured camera metadata. SCENE-SPECIFIC ACTION AND DIALOGUE: Use ONLY the AI-generated male portrait and established AI-generated location. Tight frontal close-up of the man from shoulders up, facing the camera with a subtle incredulous expression. Approximate 70–85mm full-frame-equivalent portrait perspective. Preserve glasses, mustache and realistic skin detail. BRAND AND WARDROBE CONSISTENCY: Both characters wear completely unbranded clothing. Remove any tiny Generiram.bg chest badge, name tag, pin, patch, logo or text from their clothes even if one appears in an input image. Keep their original clothing colors, shirt, tie, suspenders, white outfit, faces and positions unchanged. FINAL CHARACTER BRANDING SPECIFICATION — APPLY TO THIS ENTIRE SHOT: BRANDING AND IDENTITY — MANDATORY: The same man has a LARGE realistic tattoo of the uploaded official Generiram.bg SOLID BLACK MONOCHROME NESTED TRIANGLE ICON ONLY (no words) on the LEFT LATERAL SIDE OF HIS NECK, immediately below the left ear, between the jaw angle and shirt collar. Tattoo dimensions approximately 8 cm tall, 7 cm wide, clearly visible from his left three-quarter view. Authentic tattoo ink embedded in skin, natural pores, neck curvature and realistic shading; exactly the three nested contours of the supplied official brand artwork, all rendered in matte BLACK tattoo ink (no violet, no cyan, no colored pigment); no extra shapes or invented lettering. This tattoo exists from the very first master shot and must remain on precisely the same anatomical location, same orientation and size in every subsequent shot. Keep full-length shirt sleeves, pale-blue striped shirt, rust-orange tie, black suspenders, aviator glasses, curly dark hair and mustache. NO tattoos on face, forehead, cheek or hands. NO shirt badge or logo. WOMAN BRANDING AND WARDROBE — MANDATORY: The same blonde woman wears a fitted elegant matte BLACK crew-neck blouse with the FULL official Generiram.bg logo printed LARGE across the front of her chest, centered, spanning about 75% of the visible torso width (approximately 28–32 cm on the garment). The original multicolor icon is on the left of the large crisp WHITE 'Generiram.bg' wordmark. The print is directly integrated into the fabric, not a badge, patch, sticker or floating overlay. Preserve exact logo colors, letter shapes, proportions and alignment using the uploaded official logo as artwork reference. Avoid folds across the print; her upper torso faces camera enough to keep the logo readable. Retain her cream trousers and shoes, neat blonde bob, facial features and natural pose. Keep this EXACT same blouse and print in all shots, even when the logo is partially out of frame. BRAND PROP EXCLUSIONS: No logos on cups, bottles, table, chairs, signs, banners or badges. The tabletop contains only an unbranded clear glass water carafe, drinking glasses and optionally plain unbranded black mugs. No floating or additional branding. The neck tattoo and woman's blouse are the ONLY in-scene branding. MAN CLOSE-UP: Frame from the man’s LEFT three-quarter side when possible, so the LEFT neck tattoo is clearly visible. The tattoo must not jump to the opposite side of the neck. No badges on shirt. TATTOO COLOR LOCK — MANDATORY: The man’s left-neck tattoo is BLACK INK ONLY, monochrome charcoal-black with realistic healed pigment beneath the skin. Copy ONLY the geometric shape of the uploaded brand symbol; convert every violet/cyan contour to black. Never generate colored tattoo lines, neon glow, decals, or any lettering. If an input photo shows an older colored tattoo, REPLACE its colors with black while preserving the exact symbol outline, scale, anatomical location and realistic skin texture.

Модел: Gemini Omni Flash · видео от изображение · начален кадър: снимката отгоре IMAGE TO VIDEO — V05 — ONE CONTINUOUS CINEMATIC ORBITAL MACRO PASS. FIRST FRAME: use ONLY the supplied approved close-up portrait of the curly-haired moustached man (image №13 — man reaction portrait, aviator glasses, light-blue striped shirt, rust-orange tie, black suspenders, black Generiram symbol tattoo on the side of his neck). Match the supplied image EXACTLY at time 0. Preserve identity, anatomy, wardrobe, chair and snowy salt-flat environment. FORMAT: 16:9. RESOLUTION: 1080p. DURATION: 8 seconds. AUDIO ON. SINGLE SHOT, no edits. EXACT SPOKEN BULGARIAN LINE — MAN ONLY, verbatim: „К’ви ги говориш бе? Как така не сме истински?“ Speak in natural Bulgarian with precise mouth synchronization. First question: quick, shocked disbelief, raised eyebrows. Tiny natural pause. Second question: stronger emphasis on „Как така“, then genuine bafflement on „не сме истински?“. Not angry shouting. No extra dialogue, no voice-over, no English. The woman never speaks. NON-NEGOTIABLE EYELINE AND SCENE GEOMETRY: The man remains SEATED in the original director's chair throughout, speaking to the blonde woman who remains seated opposite him OUTSIDE THE CAMERA FRAME. He NEVER looks at the lens, not for a single frame. His gaze stays fixed at the real seated woman's eye height and original off-screen position; he must not track the camera as it moves. Minimal natural head movement, blinks and speaking gestures only. Do not move the woman or change her position in the world. THE WOMAN MUST NEVER BE VISIBLE: no face, body, hair, shoulder, silhouette, reflection in glasses, or glimpse in the background. Do not show the opposite chair or the woman. The camera path stays entirely on the OUTER, empty side of the man, avoiding any view toward the woman's position. Do not add an over-the-shoulder foreground person. Keep only the man, his own chair, salt flat and mountain in view. CINEMATOGRAPHY: ARRI ALEXA 35 cinematic visual character, natural 24 fps motion blur, 180-degree shutter, cinematic highlight roll-off, daylight color, believable focus breathing and skin texture. Begin with an 85mm portrait-lens look at T2.8; create an intentional seamless virtual perspective change toward a 24mm wide-angle look by the end. This is a single simulated continuous camera move, not an actual instantaneous lens swap. Smoothly adjust the lens perspective while the camera physically travels, without visible optical jumps, morphing or digital zoom-only effects. CAMERA PATH — CONTINUOUS FOR ALL 8 SECONDS: 0.0–1.5 s — START / APPROACH: Begin exactly at the provided portrait composition, same side of the face, same gaze direction and background. Camera immediately moves on a stabilized curved dolly path, slowly forward AND laterally toward the lower front edge of the man's face. No stationary introduction. Start with 85mm portrait compression; keep eyes sharp, then gradually pull focus toward the speaking lips. 1.5–3.2 s — CLOSING PASS: Continue the SAME flowing sideways movement, maintaining consistent screen-direction travel and constant, even camera velocity. Camera moves close to his moustache and mouth as he speaks. Mouth grows naturally in frame because the camera physically approaches, not because the face expands or deforms. Gradually widen optical perspective to keep facial geometry natural. Rack focus smoothly from eyes to moustache and lips. 3.2–4.4 s — EXTREME CLOSE PASS: Glide past the mouth and moustache at approximately 25–35 cm lens-to-face distance (not through the face). For a brief beat, lips, moustache and skin texture dominate the composition; avoid any exaggerated distortion, warped teeth or duplicated features. The camera is STILL MOVING sideways along a continuous shallow arc. Never collide with or pass through the nose, lips, glasses or head. 4.4–6.5 s — CLEAR THE FACE / ORBIT: Camera continues along the OUTSIDE of the seated man, arcs past the near cheek and around to the opposite three-quarter viewing angle WITHOUT circli

Модел: Nano Banana 2.1 · изображение от изображение IMAGE-TO-IMAGE — SHOT 15, АТМОСФЕРА — САМО СНЕЖНАТА ПЛАНИНА. INPUT IMAGES IN EXACT ORDER: (1) Основен общ кадър — промпт 01. Generate a NEW photorealistic cinematic shot from the established Generiram.bg commercial set. SHOT DESCRIPTION: Pure cinematic landscape detail of the EXACT SAME snow-capped volcanic mountain silhouette from the master shot, no people, no chairs, no props; realistic atmospheric haze and delicate clouds. Само същият заснежен връх, без герои, без столове, без маса; идентична форма и светлина като основния кадър. OPTICS: full-frame equivalent 200mm, physically realistic telephoto perspective and depth of field (f/4 for portraits and details, f/8 for broad landscapes); tripod at the shot-specific camera position; authentic cinematic lens falloff, natural perspective, realistic light direction. SAME SET: pale high-altitude salt flat, distant snow-covered symmetrical volcanic mountain, pale cyan cloudy sky, soft natural 5600K daylight, dark wooden low table, two wooden director chairs where visible. Same man: curly dark hair, moustache, tinted aviator glasses, light-blue striped long-sleeve shirt, orange tie, black suspenders, black nested-triangle tattoo on left neck. Same woman: blonde bob, fitted BLACK long-sleeved top with the correct large official Generiram.bg violet/cyan emblem and white wordmark, cream trousers. Keep characters only when specified. Preserve exact identity, wardrobe, furniture and geography. Natural imperfect human posture and authentic fabric wrinkles, NEVER posed like a fashion catalogue, NEVER perfectly squared shoulders to the camera. 4:3 landscape, 2K PNG, photoreal cinema still, subtle grain, restrained color grade. No original video frames, no added props, no extra people, no floating logos, no deformed text, no artificial CGI skin.

Модел: Gemini Omni Flash · видео от изображение · начален кадър: снимката отгоре IMAGE-TO-VIDEO — B15 — LOW FPV FLIGHT TOWARD SNOWY MOUNTAIN. Use the uploaded mountain image as the EXACT FIRST FRAME. Preserve the same unique snow-covered mountain, summit, salt desert, daylight and horizon. One continuous photorealistic 5-second shot, 16:9, 24fps. No dialogue or music. CAMERA MOVEMENT — ABSOLUTE PRIORITY: From FRAME 0.00 the CAMERA PHYSICALLY FLIES FORWARD, LOW ABOVE THE SALT FLAT, AND DIAGONALLY TOWARD SCREEN RIGHT. This is NOT a stationary mountain shot and NOT a simple zoom or straight push-in. Strong, visible ground-flow: the textured salt crust races underneath the camera and slips diagonally across the lower foreground. The distant mountain grows gradually larger but stays stable, centered or slightly left of center, with its full summit visible. Distinct real 3D parallax between fast-moving nearby salt ground and slowly shifting distant mountain. 0–1 s: Immediate fast forward-and-right aerial travel, camera roughly 1 meter above the ground. The salt surface moves visibly from the very first frame. 1–3.5 s: Maintain strong forward speed and clear diagonal RIGHTWARD translation; the ground streams past and the mountain approaches naturally. No stopping, no reversing. 3.5–5 s: Continue flying forward/right while smoothly CLIMBING from low altitude to several meters above the salt. The downward view onto the foreground salt subtly increases as the camera rises; keep the mountain peak in frame until the end. Never freeze the ground. CAMERA: Wide cinematic drone/FPV perspective, approximately 28–35mm equivalent, fixed focal length, stabilized horizon, natural motion blur, no fisheye distortion. Genuine physical movement, not digital panning. No camera shake, cuts, speed ramps, zooms, or artificial scaling. ENVIRONMENT: Preserve the mountain's exact silhouette and snow patterns. Pale sky, clear summit, only minimal nearly static distant haze. NO moving clouds across the peak, NO growing cloud cap, NO weather event, NO morphing terrain, NO additional mountains, NO people, chairs, buildings, vehicles or text. FINAL PRIORITY: (1) fast foreground ground-flow from frame zero; (2) obvious diagonal forward + screen-right flight; (3) smooth climb near the end; (4) stable recognizable mountain, clear summit; (5) no cloud animation. This must feel like a real low-altitude drone flight, NOT a still photo with moving clouds.

Модел: Nano Banana 2.1 · изображение от изображение IMAGE-TO-IMAGE — EXACT UPLOAD ORDER AS LISTED ABOVE. Create an extreme long establishing shot of two tiny seated people in wooden director chairs on an immense salt flat, monumental snow-covered mountain and open sky, filmed with 135mm telephoto from very far away, a cinematic atmospheric shot with scale and visual isolation. Tiny figures still match approved male and female characters and wardrobe; no extra people or furniture. 16:9. MANDATORY FULL-BODY WIDE FRAMING: both people head-to-toe, both of the woman’s entire knee-high black boots with soles and heels and visible ground margin, no cropped feet. Camera MUST be far enough away for complete chairs and all chair legs. MAN LEFT-NECK TATTOO: a single small healed black-ink symbol of THREE separate nested rounded triangular outlines, not a blurred circle or colored icon. Woman upper-chest tattoo only if clearly visible in the selected view: natural black healed Generiram.bg tattoo, never a floating graphic. CHARACTER IDENTITY PRIORITY: Input 1, the approved wide reference photo with her chest tattoo, defines the SAME adult woman's exact face, short straight blonde bob, white long-sleeve high-neck cutout top, black short bottoms and TWO black knee-high boots. Both of her black boots are visible in the approved wide reference photo. Never replace her boots with shoes or cream trousers, never invent a different face or a different top. If this shot is full-body, leave visible salt ground under BOTH boots and include both heels, soles and toes. If this is a medium or close shot, crop naturally for the requested camera angle; do not artificially widen just to show boots. Preserve natural anatomy, realistic skin texture, daylight, single snowy mountain and salt-flat location. All frames 16:9 landscape, cinematic ARRI Alexa 35 look, believable optics and natural skin. The approved wide reference photo is a SINGLE-PHOTO CHARACTER AND WARDROBE REFERENCE: produce ONE SINGLE CONTINUOUS PHOTO, NOT a collage or six panels. Any secondary scene photo, if provided, defines background geometry ONLY, never the woman's identity, body or clothes. No old-woman references, no face replacement, no subtitles or extra branding. CRITICAL CHEST BRANDING CONTINUITY: The approved wide reference photo already shows a naturally healed BLACK-INK Generiram.bg emblem AND the exact text Generiram.bg across the exposed upper chest. Retain the same tattoo design, scale and anatomical placement whenever this part of her chest is visible. Do not replace it with a garment print, do not remove it, and do not invent a new tattoo. The approved wide photo is the ONLY woman identity reference; do not use the old six-panel reference sheet. Preserve her black knee-high boots and original white cutout top. 16:9 landscape.

Модел: Gemini Omni Flash · видео от изображение IMAGE TO VIDEO — V06 — ONE CONTINUOUS CINEMATIC WIDE-SHOT DIALOGUE. START IMAGE — №12 (ЕКСТРЕМНО ДАЛЕЧЕН ОБЩ ПЛАН): Use the APPROVED generated still showing BOTH characters seated facing one another on wooden director’s chairs on a vast pale salt flat, a low wooden table with glass carafe and drinking glasses between them, and the enormous snow-covered mountain centered behind them. The man sits SCREEN LEFT, the blonde woman sits SCREEN RIGHT. Both characters, their complete chairs, legs, the woman's black boots, the table and the mountain must be visible from the very first frame. Match the uploaded image EXACTLY at time 0. This is NOT image №04 and NOT a medium shot of the woman. OUTPUT: 16:9 horizontal, 1080p, 8 seconds, audio ON. A single uninterrupted take. LANGUAGE — CRITICAL: The WOMAN speaks ONLY in natural native BULGARIAN. NO ENGLISH WORDS, NO ENGLISH DIALOGUE, NO TRANSLATION, NO NARRATOR. The MAN is completely SILENT throughout and never moves his lips to speak. EXACT BULGARIAN DIALOGUE — WOMAN ONLY — SPEAK THESE WORDS VERBATIM: „Еми, направили са ни с изкуствен интелект. И нас, и планината.“ ACTING AND BEATS: 0.0–0.5 s: The woman is already facing the man. She makes a small, casual shoulder shrug and takes a natural breath. The man watches her, puzzled. 0.5–3.8 s: She says „Еми, направили са ни с изкуствен интелект.“ Calm, assured, slightly teasing, like explaining something obvious. She leans forward just a few centimeters, with one restrained open-palm gesture toward herself and the man. The man subtly raises his eyebrows; he does not speak. 3.8–4.2 s: A very brief natural conversational beat. Both remain seated. 4.2–8.0 s: She says „И нас, и планината.“ On „нас“ she makes a small gesture between the two chairs. On „планината“ she turns her eyes and head slightly toward the REAL mountain behind them, without turning her back or looking into the camera. The man briefly follows her glance toward the mountain and then looks back at her. The woman completes the sentence before the shot ends. CAMERA PACKAGE: ARRI ALEXA 35 cinematic camera; 135mm full-frame-equivalent telephoto perspective, T4, 24 fps, 180° shutter. Natural daylight, subtle atmospheric haze, realistic optical compression. Focus deep enough to keep both seated characters readable and the mountain recognizable. CAMERA POSITION AND GEOMETRY: Begin at the exact distant camera position and height implied by the uploaded starting image. Camera is far from the subjects, approximately 30–40 meters away, with a clear unobstructed view of BOTH people and the mountain. The lens is aimed at the midpoint of the table. The camera remains on the audience-facing side of the conversation axis. No character or prop changes position in the scene. CAMERA MOVEMENT — MANDATORY: Begin a slow, stabilized, physically plausible DOLLY-IN immediately at frame 0. Travel straight forward by approximately 1.5–2 meters over 8 seconds, at ONE CONSTANT PHYSICAL SPEED. Do not digitally zoom. Do not change focal length. Do not pan, tilt, orbit or drift sideways. Do not stop, speed-ramp or cut. The change in framing should be subtle: the two people gradually grow slightly larger, while remaining an unmistakable full-body TWO-SHOT at the end. Preserve BOTH COMPLETE CHAIRS, all feet/boots, the low table and the snowy mountain in the composition for the entire take. EYELINES: The woman looks at the man's face at his natural seated eye level except for the brief glance at the mountain. The man looks at her and briefly at the mountain. NEITHER CHARACTER EVER LOOKS AT THE CAMERA. No fourth-wall breaking. CONTINUITY: Preserve exact identities, blonde bob, white ribbed cutout top and black Generiram.bg chest tattoo, black boots, man's curly dark hair, moustache, aviator glasses, striped light-blue shirt, rust-orange tie, black suspenders and black neck tattoo. Preserve the chairs, carafe, glasses, table, salt flat, and mountain. No extra people, no new props, no

Модел: Nano Banana 2.1 · изображение от изображение IMAGE-TO-IMAGE — EXTREME WORM'S-EYE VIEW FROM THE SALT GROUND. REFERENCE IMAGE 1 : Character identity sheet. Preserve the EXACT SAME man's face, curly dark hair, moustache, aviator glasses, clothing and neck tattoo. REFERENCE IMAGE 2 : Original wide shot. Preserve the exact salt-flat location, wooden director's chair, mountain, daylight and wardrobe. CREATE A COMPLETELY NEW CAMERA ANGLE. THIS IS NOT A PORTRAIT. THIS IS A FULL SEATED-BODY SHOT FROM GROUND LEVEL. CAMERA PLACEMENT — ABSOLUTE PRIORITY: Imagine a small mobile phone lying FLAT on the salt-covered ground, immediately beside the FRONT LEG of the man's chair. The camera lens is ONLY 2–3 CENTIMETERS ABOVE THE GROUND. The camera is approximately 70–90 cm from the man's NEAREST SHOE. The camera is positioned 30–35 degrees diagonally in front of the man's chair, toward his outer side. The camera points UPWARD at approximately 35 degrees. THE CAMERA MUST BE BELOW THE MAN'S SHOES, KNEES, HIPS, CHEST AND FACE. The camera is NOT on a tripod. NOT on a table. NOT on a mound. NOT on a pile of salt. NOT at knee height. NOT at waist height. NOT at eye level. The lens is almost touching the FLAT GROUND. COMPOSITION — EXTREMELY IMPORTANT: Show the man's NEAREST SHOE prominently in the LOWER FOREGROUND. The shoe must be clearly visible and appear LARGE because it is closest to the lens. Show his trouser leg rising from the shoe. Show BOTH KNEES and his seated legs. Show the wooden chair legs and their contact points with the salt ground. Show his hips, torso, striped shirt, orange tie, black suspenders and head. The man is seated normally in his wooden director's chair, ABOVE the ground. His body must NOT be buried in the salt. His chair must NOT sink into the ground. The salt surface is a FLAT HORIZONTAL PLANE. Only a very thin strip of salt texture is visible at the bottom edge of the frame. NO large foreground mound. NO elevated salt ridge. NO pile of salt blocking the legs. NO salt covering the man's trousers. PERSPECTIVE — CRITICAL: Use an authentic extreme low-angle wide-lens perspective. The nearest shoe appears noticeably larger than the man's head. The trouser leg creates a strong diagonal leading line from the bottom foreground toward his seated body. The chair legs rise vertically from the ground. The viewer looks UP past the shoe and knees toward the man's face. The underside of the man's chin and nose is visible from below. This must look like a real photograph captured by a phone lying on the ground. CAMERA OPTICS: 24mm equivalent wide-angle lens. Strong natural near-to-far perspective. Deep depth of field. NO telephoto compression. NO portrait lens. NO digitally cropped portrait. CHARACTER: The man remains seated in the exact original chair. He looks sideways toward the woman OFF-CAMERA. He NEVER looks into the camera. His expression is suspicious and confused. Preserve his exact recognizable facial identity. WARDROBE: Light-blue striped shirt. Orange tie. Black suspenders. Dark trousers. Original dark shoes. Do not invent different clothing. ENVIRONMENT: White salt flat. Distant snow-covered mountain. Natural daylight. Pale sky. The mountain may be partly obscured by the man's body. The woman is OUTSIDE THE FRAME. OUTPUT: 16:9 widescreen. Photorealistic. Natural camera perspective. DO NOT add vintage effects yet. DO NOT add black borders yet. DO NOT degrade image quality yet. STRICT NEGATIVES: NO portrait composition. NO head-and-shoulders framing. NO cropped legs. NO missing shoes. NO camera above ground level. NO high viewpoint. NO looking downward. NO elevated salt mound. NO buried chair. NO buried character. NO distorted anatomy. NO duplicated legs. NO changed face. FINAL COMPOSITION CHECK: The image MUST clearly show, in one continuous perspective: 1. A LARGE SHOE in the near foreground. 2. The trouser leg extending upward and backward. 3. The man's knees. 4. His seated body and wooden chair. 5. His face viewed from BELOW. If the nearest shoe is not visible, the camera composition is WRONG. If the man's head dominates the frame while his legs are missing, the camera composition is WRONG. If the foreground salt obscures the chair legs, the camera composition is WRONG. The final image must unmistakably look as though a phone was placed directly on the ground beside the seated man's shoe.

Модел: Gemini Omni Flash · видео от изображение · начален кадър: снимката отгоре IMAGE-TO-VIDEO — EXTREME GROUND-LEVEL VERTICAL TRACKING SHOT. CAMERA CLIMBS ALONG THE MAN'S LEG FROM SHOE TO FACE. AUTHENTIC 2007 NOKIA N95 VIDEO AESTHETIC. STARTING FRAME: Use the uploaded image as the EXACT first frame. A curly-haired man with a moustache and aviator glasses sits in a wooden director's chair on a white salt flat. He wears a light-blue striped shirt, orange tie, black suspenders, dark trousers and black shoes. The nearest shoe appears enormous in the foreground because the camera is positioned almost on the ground. Preserve his exact identity, clothing, chair, body proportions and environment. CAMERA — ABSOLUTE PRIORITY: The camera starts only 2–3 cm above the salt ground, approximately 70–90 cm from the nearest shoe. Use a fixed 24mm equivalent wide-angle lens. ONE CONTINUOUS PHYSICAL CAMERA MOVEMENT: The camera travels FORWARD AND UPWARD, following the man's nearest leg. It physically climbs alongside the trouser leg, moving from the shoe toward the knee, then past the thigh and torso to his face. This is a VERTICAL TRACKING ASCENT, not a zoom, not an orbit and not a camera tilt from a stationary position. The camera must change its actual height and position in three-dimensional space. CAMERA MOVEMENT STARTS AT FRAME 0.00. The movement is smooth, continuous and uninterrupted. CAMERA CHOREOGRAPHY: 0.00–0.70 SECONDS — SHOE: Immediately move forward and upward alongside the large foreground shoe. The shoe passes very close to the lens, creating strong natural perspective and parallax. The camera rises from 3 cm to approximately 20 cm above the salt ground. Do not pass through the shoe. Do not distort the shoe. 0.70–1.50 SECONDS — TROUSER LEG: Continue ascending parallel to the man's nearest dark trouser leg. The camera travels close to the fabric, approximately 20–30 cm away from the leg. The trouser fabric moves downward through the frame as the camera rises. Visible fabric folds, seams and realistic texture. The camera rises from approximately 20 cm to approximately 65 cm above the ground. NO pause. NO change of direction. NO digital zoom. 1.50–2.20 SECONDS — KNEE AND TORSO: Continue the SAME upward movement. Pass the knee and thigh. Reveal the man's light-blue striped shirt, orange tie and black suspenders. The camera rises toward his upper chest. The movement remains smooth and physical. The chair stays completely stationary. 2.20–2.70 SECONDS — FACE REVEAL: Continue ascending until the man's face becomes the dominant subject. The camera reaches approximately the height of his lower face. Finish the ascent at a slightly LOW camera angle, looking gently upward toward his face. The man's face must be clearly visible. Do not reveal or reconstruct unseen neck tattoos. 2.70–7.00 SECONDS — CLOSE-UP WITH SUBTLE MOTION: The camera continues a very slow, gentle upward and slightly forward tracking movement. Keep the man's face framed in a tight close-up. Do not zoom. Do not orbit. Do not reverse direction. Do not move behind the man. Do not reveal the opposite side of his neck. The camera maintains a consistent view of his recognizable face. The movement becomes slower but NEVER stops. CHARACTER PERFORMANCE — CRITICAL: The man starts in the exact seated pose shown in the uploaded image. He remains seated throughout the video. His legs, feet and chair remain completely still. The apparent movement of the shoe and trouser leg comes ONLY from the moving camera. As the camera approaches his face, the man turns his attention toward the woman sitting OFF-CAMERA. His head turns naturally toward SCREEN LEFT. He looks toward the woman, NOT toward the moving camera. Once he turns toward the woman, he keeps his gaze fixed on her. NEVER LOOK INTO THE CAMERA. NEVER TRACK THE CAMERA WITH HIS EYES. DIALOGU

Модел: Nano Banana 2.1 · изображение от изображение IMAGE-TO-IMAGE — SHOT 14, ДЕТАЙЛ — РЪКАТА ВЪРХУ СТОЛА. INPUT IMAGES IN EXACT ORDER: (1) Референция на мъжа — промпт 02.; (2) Основен общ кадър — промпт 01. Generate a NEW photorealistic cinematic shot from the established Generiram.bg commercial set. SHOT DESCRIPTION: Tight authentic macro detail of the SAME man’s hand resting casually on the original wooden director chair armrest, visible light-blue striped sleeve and subtle natural skin texture. Ръката на мъжа естествено лежи върху дървената облегалка, малки движения на пръстите; същата риза и стол. OPTICS: full-frame equivalent 100mm macro, physically realistic telephoto perspective and depth of field (f/4 for portraits and details, f/8 for broad landscapes); tripod at the shot-specific camera position; authentic cinematic lens falloff, natural perspective, realistic light direction. SAME SET: pale high-altitude salt flat, distant snow-covered symmetrical volcanic mountain, pale cyan cloudy sky, soft natural 5600K daylight, dark wooden low table, two wooden director chairs where visible. Same man: curly dark hair, moustache, tinted aviator glasses, light-blue striped long-sleeve shirt, orange tie, black suspenders, black nested-triangle tattoo on left neck. Same woman: blonde bob, fitted BLACK long-sleeved top with the correct large official Generiram.bg violet/cyan emblem and white wordmark, cream trousers. Keep characters only when specified. Preserve exact identity, wardrobe, furniture and geography. Natural imperfect human posture and authentic fabric wrinkles, NEVER posed like a fashion catalogue, NEVER perfectly squared shoulders to the camera. 4:3 landscape, 2K PNG, photoreal cinema still, subtle grain, restrained color grade. No original video frames, no added props, no extra people, no floating logos, no deformed text, no artificial CGI skin.

Модел: Gemini Omni Flash · видео от изображение · начален кадър: снимката отгоре IMAGE TO VIDEO — B-ROLL 14 — MAN'S HAND. Use the uploaded close-up of the man's hand on the wooden director's chair armrest as the EXACT first frame. CAMERA: Sony VENICE 2, 100mm macro lens, T2.8, 24fps, 16:9. 4 seconds, silent. CAMERA MOVEMENT: Immediate continuous low slider move LEFT to RIGHT, approximately 20–25 cm, with a very subtle forward diagonal drift of 5 cm. Maintain macro framing on the hand and wood. Physically shift perspective to show realistic depth between knuckles, sleeve cuff, armrest and soft background. No zoom, no tilt to the man's face, no reversal or cuts. HAND ACTION: At first his hand hangs relaxed. Mid-shot he makes ONE small, believable gesture: the index and middle fingertips flex, briefly brush the wooden edge, then relax again. Wrist and forearm remain supported; no lifting the hand, fist, tapping loop or dramatic gesture. Exactly five fingers, consistent knuckles and skin. CONTINUITY: Same adult man's hand, light-blue fine-striped shirt cuff, original worn wooden chair. No ring or new accessories. Preserve natural skin texture and chair geometry. Background remains soft, no duplicated hands. Photorealistic restrained editorial insert. No speech, narration, music or subtitles.

Модел: Nano Banana 2.1 · изображение от изображение IMAGE-TO-IMAGE — PRECISE LOCAL IMAGE EDIT. INPUT IMAGE 1 : The existing cinematic photograph of a man and a woman sitting opposite each other in wooden director's chairs, with a snow-covered mountain in the background. INPUT IMAGE 2 : The official Generiram.bg logo. TASK: Edit INPUT IMAGE 1 by adding ONE realistic permanent black-ink tattoo to the woman's exposed upper chest. THIS IS AN IMAGE EDIT, NOT A NEW SCENE GENERATION. TATTOO DESIGN: Use the COMPLETE official Generiram.bg logo from INPUT IMAGE 2 as the exact design reference. The tattoo must include: - The original three nested rounded triangular outlines. - The complete text "Generiram.bg". Convert the entire logo to solid charcoal-black tattoo ink. TATTOO POSITION: Place the tattoo on the woman's exposed skin inside the horizontal opening of her white top. Center it horizontally between her shoulders. Position the tattoo in the upper part of the exposed chest area, above the cleavage. The complete logo must fit naturally within the visible skin area. Do not place any part of the tattoo on clothing. TATTOO APPEARANCE: Realistic healed black tattoo. Pigment embedded beneath the skin. Natural skin texture visible through the ink. Subtle curvature following the anatomy. No artificial glow. No sticker effect. No floating graphics. ABSOLUTE IMAGE PRESERVATION: Keep the original photograph unchanged. Preserve the exact faces and identities of both people. Preserve the woman's hairstyle, breasts, body proportions, white cutout top, black bottoms and black boots. Preserve the man's appearance, clothing and neck tattoo. Preserve the exact chairs, table, glasses, water carafe, mountain, sky, lighting and shadows. Preserve the original camera angle, framing, perspective and 16:9 aspect ratio. DO NOT: - Generate a different woman. - Generate a different man. - Change the clothing. - Change the body proportions. - Change the pose. - Change the background. - Move the camera. - Crop the image. - Add any other tattoos. - Add any text outside the tattoo. FINAL RESULT: The exact same photograph as INPUT IMAGE 1 , with ONLY ONE modification: A realistic monochrome black Generiram.bg logo tattoo on the woman's exposed upper chest.

Модел: Gemini Omni Flash · видео от изображение IMAGE TO VIDEO — V07 — ONE-WAY FULL CINEMATIC ORBIT, NO REVERSE, NO ZOOM, EXACT BULGARIAN DIALOGUE. SOURCE IMAGE / FRAME ZERO (MANDATORY): Use ONLY the uploaded approved medium-wide TWO-PERSON image: moustached curly-haired man with aviator glasses, light-blue striped shirt, orange tie, suspenders and black Generiram neck tattoo seated FRAME LEFT; blonde woman in white ribbed long-sleeved cutout top with black Generiram.bg chest tattoo and black boots seated FRAME RIGHT. Between them: low dark wooden table, glass carafe and two glasses. Behind: pale salt flat and snow-covered mountain. At 0.0 seconds MATCH THIS IMAGE EXACTLY, including camera side, chair positions, faces and props. This is NOT image №04 (woman alone). No end-frame reference. OUTPUT: 16:9, 1080p, AUDIO ON, target 8 seconds. Single unbroken live-action take; no editing cuts. Keep all dialogue audible through its final word. DIALOGUE — NON-NEGOTIABLE: The WOMAN is the ONLY SPEAKER. She says the COMPLETE exact Bulgarian line ONCE, in natural fluent BULGARIAN with accurate lip sync: „В Генерирам Бе Ге, бе. Там си правиш и снимки, и видеа с изкуствен интелект.“ PRONUNCIATION LOCK: The SPOKEN SCRIPT above already contains the Cyrillic phonetic words “Бе Ге”. Read the text literally: “Генерирам Бе Ге”. “BG” is the written brand abbreviation ONLY, not a phonetic instruction. Do NOT pronounce the English letter names, do NOT say “Би Джи” / “Bee Gee”. Do not translate, paraphrase, skip or add words. Confident, cheerful, lightly teasing, conversational performance. The man is SILENT throughout: lips closed, natural skeptical reactions only. Clear female voice above gentle outdoor wind; NO male speech, NO English speech, NO narration, NO music, NO subtitles. CAMERA PACKAGE: Sony VENICE 2, ZEISS Supreme Prime 35mm fixed prime lens, T4, 24 fps, 180-degree shutter, stabilized cinema dolly / robotic curved track, natural physical motion blur. Camera optical center about 1.25–1.45 m above the ground, adjusting height only minimally to avoid obstruction. One fixed 35mm focal length from FIRST to LAST frame. NO zoom, NO focal-length changes, NO synthetic push-in, NO digital crop. Focus follows the woman's EYES and MOUTH; the camera can PAN its head smoothly to track her, but panning is not a substitute for the required large physical translation. ABSOLUTE MOTION LOCK — MOST IMPORTANT CHANGE: The previous failed video STOPPED midway, REVERSED DIRECTION and ZOOMED TOWARD THE WOMAN. It also delayed camera travel until AFTER the woman started speaking and visually duplicated the man during the orbit. DO NOT REPEAT ANY OF THESE ERRORS. Move the PHYSICAL CAMERA continuously on ONE uninterrupted, collision-free, ONE-WAY CIRCULAR ARC in a SINGLE ROTATIONAL DIRECTION around the outside of the man's chair and the table. Travel approximately 6–8 meters and change camera viewpoint approximately 120–140 degrees during the shot. At every moment the camera advances ALONG THE SAME ARC: positive angular progress only, never backwards, never returning to an earlier camera position. Constant, visibly substantial orbital travel speed from THE VERY FIRST FRAME (t=0.00 s) through final frame. At t=0.00 the first motion is already under way: the camera does not hold a static establishing shot, wait for dialogue, or ease in after the woman begins talking. Begin translation before the first spoken syllable. Camera motion starts at 0.00 s; first spoken syllable at approximately 0.30–0.45 s, with camera already visibly translating. NO PAUSE, NO HESITATION, NO REVERSE, NO BOOMERANG MOTION, NO BACKTRACKING, NO ORBIT-THEN-DOLLY-IN, NO ZOOM TO WOMAN, NO FAKE PUSH-IN, NO CAMERA SETTLING. Do not end behind the man. Camera must CLEAR his chair and travel onward to the OPPOSITE SIDE OF THE TABLE, revealing a genuinely new perspective on the woman. The final image MUST NOT resemble the initial front-facing two-shot. SPATIAL ROUTE — SPECIFIC GEOMETRY: At frame zero, man is screen-left, woma

Модел: Nano Banana 2.1 · изображение от изображение IMAGE-TO-IMAGE — SHOT 10, МАКРОДЕТАЙЛ — ТАТУИРОВКАТА. INPUT IMAGES IN EXACT ORDER: (1) Референция на мъжа — промпт 02.; (2) Оригиналното PNG лого на Generiram.bg (за геометрията на татуировката). Generate a NEW photorealistic cinematic shot from the established Generiram.bg commercial set. SHOT DESCRIPTION: Authentic black ink tattoo of exactly three nested rounded triangular loops on the SAME man’s left neck below the ear, genuine skin texture, no colored ink or lettering. Реалистична ЧЕРНА татуировка с точно три вложени заоблени контура, под лявото ухо; не нов символ. OPTICS: full-frame equivalent 100mm macro, physically realistic telephoto perspective and depth of field (f/4 for portraits and details, f/8 for broad landscapes); tripod at the shot-specific camera position; authentic cinematic lens falloff, natural perspective, realistic light direction. SAME SET: pale high-altitude salt flat, distant snow-covered symmetrical volcanic mountain, pale cyan cloudy sky, soft natural 5600K daylight, dark wooden low table, two wooden director chairs where visible. Same man: curly dark hair, moustache, tinted aviator glasses, light-blue striped long-sleeve shirt, orange tie, black suspenders, black nested-triangle tattoo on left neck. Same woman: blonde bob, fitted BLACK long-sleeved top with the correct large official Generiram.bg violet/cyan emblem and white wordmark, cream trousers. Keep characters only when specified. Preserve exact identity, wardrobe, furniture and geography. Natural imperfect human posture and authentic fabric wrinkles, NEVER posed like a fashion catalogue, NEVER perfectly squared shoulders to the camera. 4:3 landscape, 2K PNG, photoreal cinema still, subtle grain, restrained color grade. No original video frames, no added props, no extra people, no floating logos, no deformed text, no artificial CGI skin.

Модел: Gemini Omni Flash · видео от изображение · начален кадър: снимката отгоре IMAGE TO VIDEO — B-ROLL 10 — MAN'S NECK TATTOO. Use the uploaded extreme close-up of the man's LEFT neck tattoo as the EXACT first frame. CAMERA: ARRI ALEXA 35, 100mm macro lens, T4, 24fps, 16:9. 4 seconds, silent. MOVEMENT: Immediate, slow physical curved slider from LEFT to RIGHT, only 8–12 degrees of viewpoint change. The lens stays at the same height and nearly the same distance; counter-pan gently to keep the entire tattoo visible and sharply focused. Real subtle parallax between skin, hair curls, ear and blurred background. No zoom, pullback, camera reversal, cuts or large orbit. SUBJECT: One natural swallow and faint breathing movement beneath the jaw; neck skin shifts minimally. The head stays oriented toward the woman off-camera, with no turn toward lens. Do not reveal the whole face. TATTOO LOCK: Preserve the exact original BLACK symbol of THREE nested rounded triangular loops, same size, thickness, placement below left ear and orientation in EVERY frame. Tattoo is ink IN the skin, not a floating overlay. No fourth loop, letters, rotation, warping or logo redraw. CONTINUITY: Same curly hair, sideburn, stubble and blue striped collar. No extra heads or necks. Photorealistic macro skin detail, natural light, no dialogue or music.

Модел: Nano Banana 2.1 · изображение от изображение IMAGE-TO-IMAGE — SHOT 09, МАКРОДЕТАЙЛ — ОЧИТЕ НА МЪЖА. INPUT IMAGES IN EXACT ORDER: (1) Референция на мъжа — промпт 02.; (2) Основен общ кадър — промпт 01. Generate a NEW photorealistic cinematic shot from the established Generiram.bg commercial set. SHOT DESCRIPTION: Extreme close-up of the SAME man’s eyes behind tinted aviator glasses, natural pores, tiny authentic eye movements, cinematic shallow focus. Лявото око на мъжа зад авиаторските очила, истински отблясък на небето, без деформации. OPTICS: full-frame equivalent 100mm macro, physically realistic telephoto perspective and depth of field (f/4 for portraits and details, f/8 for broad landscapes); tripod at the shot-specific camera position; authentic cinematic lens falloff, natural perspective, realistic light direction. SAME SET: pale high-altitude salt flat, distant snow-covered symmetrical volcanic mountain, pale cyan cloudy sky, soft natural 5600K daylight, dark wooden low table, two wooden director chairs where visible. Same man: curly dark hair, moustache, tinted aviator glasses, light-blue striped long-sleeve shirt, orange tie, black suspenders, black nested-triangle tattoo on left neck. Same woman: blonde bob, fitted BLACK long-sleeved top with the correct large official Generiram.bg violet/cyan emblem and white wordmark, cream trousers. Keep characters only when specified. Preserve exact identity, wardrobe, furniture and geography. Natural imperfect human posture and authentic fabric wrinkles, NEVER posed like a fashion catalogue, NEVER perfectly squared shoulders to the camera. 4:3 landscape, 2K PNG, photoreal cinema still, subtle grain, restrained color grade. No original video frames, no added props, no extra people, no floating logos, no deformed text, no artificial CGI skin.

Модел: Gemini Omni Flash · видео от изображение · начален кадър: снимката отгоре IMAGE TO VIDEO — V08 — EXTREME FACE CLOSE-UP, EYES-TO-MOUTH VERTICAL DESCENT ONLY. START IMAGE: №13 (Портрет на мъжа — реакция). The ONLY supplied reference is the uploaded extreme close-up of the man's face: curly hair at the edge of frame, aviator glasses, eyes, nose, moustache and mouth. The first video frame must match this exact photograph. Treat the supplied pixels as the definitive source of his identity. DO NOT invent unseen parts of his head, neck, body, wardrobe, tattoo, chair or environment. 16:9, 1080p, 7 seconds, AUDIO ON, no end frame. EXACT SPOKEN BULGARIAN LINE — ONLY THE MAN SPEAKS: „Е, тогава защо са ме направили толкова грозен, бе?“ Natural native Bulgarian delivery, precise lip sync, dry comedic wounded disbelief; emphasis on „толкова грозен“. The man addresses the woman who is OFF-SCREEN, never the viewer. The woman does not appear and does not speak. No added or missing words, no English, no subtitles, no narrator. ABSOLUTE FRAMING LOCK — HIGHEST PRIORITY: THIS IS AN EXTREME CLOSE-UP OF THE SAME FACE FOR THE ENTIRE VIDEO. The framing begins around his EYES and ends around his MOUSTACHE AND MOUTH. The frame must NEVER widen into a head-and-shoulders portrait, medium shot, full-body shot or environmental shot. Do NOT show his neck, neck tattoo, collar, tie, suspenders, torso, chair, hands or any previously unseen anatomy. DO NOT reveal the tattoo at any point; its design is NOT provided in the starting photograph and MUST NOT be fabricated. Do not show the woman. Keep the face large in the frame from first to last frame. Do NOT pull back, zoom out, widen the field of view, or shrink the face within the frame. NO establishing shot and NO reveal of his full face from a new side angle. A little more of the moustache and mouth becomes central, NOT more of his body. CAMERA AND OPTICS: ARRI ALEXA 35, ZEISS Supreme Prime 65 mm FIXED lens, T2.8, 24 fps, 180-degree shutter. No change in focal length or camera-to-subject distance. Maintain exactly the original horizontal camera position, horizontal orientation and lens-to-face distance. The camera performs ONE AND ONLY ONE movement: a SMALL, PURELY VERTICAL, STRAIGHT-DOWN PHYSICAL DESCENT of approximately 12–18 cm, from the eye-level composition to a mouth-level composition. The optical axis may make only a tiny necessary upward tilt to retain the mouth and lower portion of the glasses within the SAME EXTREME CLOSE-UP. No dolly forward, no dolly backward, no diagonal translation, no orbit, no sideways motion, no pan, no digital crop animation, no zoom in or out, no perspective jump. DO NOT move far enough down to reveal the underside of the jaw or neck. This is NOT a low-angle full-face reveal. The visible shift is a subtle but unmistakable change of emphasis from the EYES to the MOUTH within the SAME FACE-SCALE CLOSE-UP. CONTINUOUS MOTION: Start the vertical descent at 0.00 seconds, in the very first frame. Descend slowly and smoothly for the entire 7 seconds, without stopping or changing direction. Keep camera distance CONSTANT; facial scale must remain virtually unchanged. The man's head remains naturally stable. Use realistic micro-parallax only; do not reconstruct the face from a new viewpoint. TIMED ACTION: 0.00–0.45s: Exact opening eye close-up. Camera ALREADY moves straight DOWN, only a few millimeters. He looks toward the off-screen woman, NOT at the lens. His expression shows surprised offense. 0.45–2.20s: Camera descends slightly within the same extreme close-up, bringing the glasses and upper nose toward the top of frame. He says: „Е, тогава защо…“ 2.20–4.10s: The SAME uninterrupted slow vertical descent makes his nose and moustache increasingly prominent. He says: „…са ме направили…“ The face remains exactly the same person, same proportions and glasses. 4.10–5.90s: Camera is near mouth level but remains EXTREMELY CLOSE, with no new neck or body visible. He emphasizes: „…толкова грозен…“ Natural mouth and moustache motion; no e

Модел: Nano Banana 2.1 · изображение от изображение IMAGE TO IMAGE — PREPARE B-ROLL 11 CARAFE CLOSE-UP. Use the RIGHT-HAND carafe photograph from the supplied pair as the ONLY base image. Reframe the SAME real scene by making the camera substantially closer, with a tighter 16:9 composition focused on the existing glass carafe and both original water glasses. CRITICAL CROP: Exclude the LEFT EDGE and LEFT CORNER of the wooden table completely. Exclude the table legs, chairs, people, arms and hands. Show only the CENTER AREA of the original dark wooden tabletop, with the three glass objects large and prominent. The tabletop should fill the lower third to half of the frame; the mountain remains softly blurred in the distance behind the glassware. Do not change the relative positions of the two glasses and carafe. Keep the carafe and BOTH glasses fully visible, not cut off. Camera: Sony VENICE 2, 100mm macro, T4, low tabletop height. Realistic glass refraction, accurate water levels, natural daylight, shallow depth of field. Preserve the exact mountain silhouette and salt-flat background. No new props, no extra glasses, no people. Not a wider shot; this is a noticeably TIGHTER close-up with the table's left edge OUT OF FRAME.

Модел: Gemini Omni Flash · видео от изображение IMAGE TO VIDEO — B-ROLL 11 — CARAFE AND TWO GLASSES. START ONLY from the NEW TIGHTLY REFRAMED image: carafe plus TWO water glasses, blurred mountain, center tabletop. No people and NO visible left edge or corner of the table. CAMERA: Sony VENICE 2, 100mm macro, T4, 24fps, 16:9. 4 seconds, silent. MOVEMENT: Begin moving at frame 0.00. Execute one smooth, low tabletop-height physical slider move from RIGHT to LEFT, approximately 18–25 cm, on a shallow arc. Keep the three glass objects completely visible, centered and roughly constant in size. Camera travels around the glassware by only 8–12 degrees, producing elegant layered parallax and changing realistic refractions of the distant blurred mountain through the water. Fixed lens; NO zoom, push-in, pullback, rise, reversal or cuts. GLASS AND WATER: Carafe and both glasses stay completely stationary on the wooden tabletop. Water surfaces are calm with extremely subtle natural reflections. No pouring, shaking, spilling or changing water levels. Exactly ONE carafe and TWO glasses throughout. No morphing glass rims, no duplicated objects. FRAMING LOCK: Only the center tabletop, carafe, two glasses and distant soft landscape. NEVER reveal the table's left edge, table legs, chairs, hands, people or extra props. Photorealistic premium macro product insert, crisp glass and natural highlights, blurred background. No dialogue, music, titles or overlays.

Модел: Nano Banana 2.1 · изображение от изображение IMAGE-TO-IMAGE — CINEMATIC SHOT V09 EXTREME TELEPHOTO CLOSE-UP — WOMAN'S AMUSED REACTION REFERENCE IMAGE: Use the uploaded image as the exact reference for the woman's identity, facial features, hairstyle, skin texture, clothing, and environment. Generate a NEW cinematic photograph of the SAME woman, from a COMPLETELY DIFFERENT CAMERA POSITION. THIS IS NOT A CROP OR DIGITAL ZOOM OF THE REFERENCE IMAGE. CAMERA REPOSITIONING — CRITICAL: Physically relocate the virtual camera to the OPPOSITE SIDE OF THE TABLE, near the man's position. Position the camera approximately 25–30 DEGREES OFF THE WOMAN'S FRONTAL FACIAL AXIS, toward her left-front side. The camera must be slightly BELOW her eye level, looking gently upward at her face. This must create a genuine three-dimensional change in perspective. We should see a beautiful, natural three-quarter facial angle, NOT the same camera angle as the reference. CAMERA AND LENS: ARRI ALEXA 35 ZEISS Supreme Prime 135mm telephoto lens Aperture T2 Extremely shallow depth of field Natural cinematic background compression Photorealistic skin detail FRAMING — EXTREME CLOSE-UP: Frame almost exclusively the woman's FACE. Her face should occupy approximately 80–90% of the image height. The composition should show: - Her eyes. - Her nose. - Her cheeks. - Her mouth. - Her wide, amused smile. - Her chin. - A small amount of her blonde bob haircut. The frame should begin near the top of her hair and end just below her chin. NO full torso. NO chest. NO chest tattoo. NO arms. NO hands. NO chair. NO table. NO man. Do not widen the composition. Do not show her entire upper body. The visual emphasis must be entirely on her facial expression. FACIAL EXPRESSION — VERY IMPORTANT: The woman has just heard something ridiculous and is about to deliver a witty, sarcastic punchline. She is genuinely amused. She has a BEAUTIFUL, WIDE, NATURAL SMILE. Her cheeks are raised. Her eyes are slightly narrowed from smiling. Her teeth are naturally visible. Her expression is playful, teasing, and mischievous. She looks as if she is trying not to laugh at the man. The smile must feel spontaneous and authentic, not like a posed advertising photograph. Preserve the woman's EXACT recognizable identity. Same face shape. Same eyes. Same nose. Same lips. Same jawline. Same blonde bob haircut. Same skin tone. Same age and natural facial proportions. Do NOT beautify, reshape, or replace her face. EYELINE — ABSOLUTE RULE: THE WOMAN MUST NEVER LOOK AT THE CAMERA. She is looking directly at the MAN seated opposite her, who is OUTSIDE THE FRAME. Her gaze is directed naturally sideways toward the man's real position. Her eyes must be visibly oriented OFF-CAMERA. Her head is turned slightly toward him. NO direct eye contact with the lens. NO looking at the viewer. NO front-facing portrait. NO fourth-wall breaking. ENVIRONMENT: Preserve the same outdoor salt-flat location and natural daylight. The distant snowy mountain and pale sky may appear only as extremely soft, abstract, heavily blurred background colors. The background must be strongly compressed and out of focus due to the 135mm telephoto lens and T2 aperture. No sharp mountain details. No additional people. No new objects. COMPOSITION: Cinematic extreme close-up. Authentic three-quarter facial perspective. Slightly low camera angle. Strong telephoto compression. Very shallow depth of field. Eyes and smile tack-sharp. Background beautifully blurred. Natural skin texture. Realistic teeth and facial anatomy. 16:9 widescreen cinematic composition. FINAL PRIORITY: 1. A genuinely NEW camera angle from the opposite side of the table. 2. Approximately 25–30 degrees off-axis. 3. Slightly below eye level. 4. Extreme 135mm telephoto close-up. 5. Almost ONLY her face visible. 6. Wide, genuine

Модел: Gemini Omni Flash · видео от изображение IMAGE TO VIDEO — 16:9, 1080p, 6 seconds, AUDIO ON. START IMAGE: ONLY approved NEW still №05 (Финална реакция на жената — сдържана усмивка преди шегата), a close-up of the same woman. DO NOT upload the six-panel sheet as video start image. SCENE FUNCTION: FINAL COMEDIC PUNCHLINE. This shot follows the man's offended line about being made ugly by AI. The woman looks off-camera LEFT toward him. At 0.0–0.7 s she holds back a smile, eyes flicking toward him, one eyebrow rising. At 0.7–1.2 s she takes a tiny playful breath, tilts her chin and prepares to speak. At 1.2–4.1 s she says, with playful deadpan confidence and perfect natural Bulgarian lip-sync, EXACTLY: „Е, то и изкуственият интелект си има граници!“ Emphasize „граници“, with teasing warmth rather than anger. At 4.1–6.0 s she breaks into a genuine spontaneous laugh, eyes crinkling, shoulders moving subtly, head dipping a little; laughter is audible and natural. The man does NOT speak; no other voices, no English, no extra words. CAMERA: physically plausible handheld-gimbal micro-orbit from camera-left toward frontal 3/4, approximately 15–20 cm lateral travel across six seconds with a subtle 5–8 cm push-in during the punchline. === DETAILED CAMERA / LENS / MOVEMENT — V09 === CAMERA / CINEMATOGRAPHY: Sony VENICE 2, Cooke S4/i 85mm, T2.8, 24 fps, 180-degree shutter. Start on the exact approved final woman reaction still; retain its camera angle, eye line, original white top and chest tattoo if visible. Gentle physical lateral slider move 10cm toward camera-right during 6 seconds; natural very slight operator breathing, NOT an artificial zoom. Hold close framing throughout punchline and the start of her laugh, BOTH eyes and mouth fully in frame, never crop chin. Focus on eyes, mountain softly out of focus, realistic daylight and skin texture. One continuous take, NO cut, orbit, whip, added logos or captions.

Модел: GPT Image 2.5 Flare · изображение от изображение IMAGE TO IMAGE — R01 / NATURAL CINEMATIC MIRROR CLOSE-UP — REVISED v22 UPLOAD FIVE IMAGES IN THIS ORDER: 1 — CINEMATIC REAR-VIEW-MIRROR REFERENCE: use ONLY the optical idea, close framing, moody exposure and windshield depth. The original person's eyes/face have been neutral-grey blurred. NEVER recreate the original person, red scribbles, original road or background. 5 — STUNT RAMP GEOMETRY REFERENCE: the supplied miniature physical model with a car upside-down at the end of a steep C-shaped stunt ramp. Use ONLY its SIDE-PROFILE TRAJECTORY: flat ground → smoothly rising curved roadway → vertical climb → progressively bending overhang → final horizontal roadway with driving surface facing DOWN. This is a GEOMETRY SKETCH, NOT A SCENE REFERENCE. NEVER copy its upside-down miniature car, dark asphalt, steel lattice, scaffold, studio lighting, gray backdrop, toy scale, smoke, fire, supports or ground. Rebuild the ramp at monumental real-world scale from the SAME cracked white salt as the surrounding plain, structurally seamless with the salt roadway. The ramp must be distant, ahead of the car, and properly proportioned: roadway wide enough for a full-size Tesla with ample margins. Its far-end tip is inverted and projects forward over the approach road. No standalone sculptural decoration; this is a driveable continuous road. 2 — ORIGINAL GENERIRAM.BG DRY WHITE SALT FLAT AND SNOW-COVERED MOUNTAIN: authoritative real environment, mountain silhouette, horizon, natural daylight. Do not copy any ramp or vehicle from this reference. 3 — ORIGINAL BLONDE AD WOMAN: sole identity source for her exact blue-green eyes, eyebrows, blonde bob, skin, age and expression. 4 — TESLA CABIN REFERENCE: use ONLY credible left-hand-drive cabin materials and the shape and physical mounting of an interior rear-view mirror. Do not copy the reference image composition or introduce its original outside scenery. GOAL: one imperfect, intimate, photorealistic ACTION-MOVIE FRAME photographed INSIDE a moving futuristic white Tesla Roadster, NOT a perfectly centered interior showroom render. 16:9, highest resolution. VERY TIGHT FRAMING. A camera operator is positioned a little behind the front passenger's shoulder, near the center tunnel but noticeably OFF-AXIS, pointed diagonally forward and slightly upward at the real cabin rear-view mirror. The mirror is LARGE, approx. 38–45% of the image width, slightly RIGHT of image center and in the UPPER THIRD. The driver's reflected eyes are the visual focal point, clearly recognizable as the blonde woman from Image 3 , with physically plausible mirrored geometry, realistic glass tint and subtle reflected highlights. Her gaze is locked forward on the route; an intense, mischievous, adrenaline-filled expression. Her mouth and torso are NOT visible outside the mirror. She occupies the LEFT driver's seat. Do not invent a second face, extra eyes, passenger or a floating mirror. The windshield fills the remaining frame. Through it, the ORIGINAL vast cracked WHITE SALT FLAT and the SAME snowy mountain appear in atmospheric perspective, not an invented highway. Far in front of the car, a SMALL DISTANT WHITE SALT ROAD RISES NATURALLY FROM THE FLAT AND GRADUALLY CURVES UPWARD INTO A HUGE OVERHANGING STUNT RAMP, continuing beyond vertical until its free tip points back toward the salt flat. This ramp is DESCRIBED IN WORDS AND MUST BE GENERATED FROM SCRATCH AS PART OF THE SAME PHYSICAL LANDSCAPE, NOT pasted from another image. The continuous salt driving track points DIRECTLY INTO THE RAMP ENTRANCE. At this stage it is a distant silhouette, subtle and integrated into atmospheric haze, NOT a foreground decoration, floating sculpture, freestanding wall or pasted object. Keep mountain clearly recognizable. The car is travelling TOWARD the ramp, not past it. CROP OUT ALMOST THE ENTIRE DASHBOARD: no center touchscreen, no full steering wheel, no broad horizontal dash, no large roof console. Only thin, naturally shadowed strips of windshield edge and cabin roof are allowed to anchor the mirror physically. DO NOT place two hands prominently on a steering wheel. No clean symmetric product-photo framing. REAL CINEMA LOOK: ARRI ALEXA 35, ARRI Signature Prime 65mm at T2.8, subtle focus falloff, 24fps motion-picture still, realistic mild grain, slight lens breathing and imperfect optical reflections. Natural late-afternoon sunlight, softly warm highlights, cooler deep cabin shadows, gently veiled windshield glare, restrained halation, realistic glass and mirror micro-contrast. Subtle camera offset, slight lived-in imperfections. Natural exposure, NO plastic HDR, no excessive sharpness, no sterile CGI, no game-engine render, no perfect symmetry, no giant bright center display, no aggressive anamorphic streaks. COMPOSITION PRIORITIES IN ORDER: 1) convincing eyes of the CORRECT woman in an actual mirror; 2) tight asymmetrical movie composition; 3) recognizable original mountain and salt flat; 4) small but legible distant road curving upward into an overhanging ramp. No lettering, no subtitles, no watermark. This is a still for the Bulgarian spoken line: “Виждаш ли оная рампа там?” — do not render the dialogue as text. RAMP GEOMETRY OVERRIDE: The fifth image is a physical model and must NEVER dominate the composition or be recreated as an object. The ramp must appear far through the windshield as part of the natural salt landscape, with atmospheric perspective, not pasted-on. Maintain mirror close-up as main subject. No upside-down vehicle visible in this first shot.

Модел: Gemini Omni Flash · видео от изображение · начален кадър: снимката отгоре GEMINI OMNI FLASH — IMAGE TO VIDEO — R01 MIRROR SHOT — FINAL LOCKED CAMERA VERSION. INPUT: ONE APPROVED STILL of the SAME blonde driver reflected in the INSIDE rear-view mirror, salt roadway and distant curved stunt ramp directly ahead through windshield. Use this exact image as first frame. NO end image. AUDIO ON. NATIVE BULGARIAN SPEECH. ABSOLUTE RULE — CAMERA NEVER LEAVES THE CAR. The camera remains physically INSIDE the passenger cabin for EVERY FRAME of the entire clip. One continuous uninterrupted take from exactly the same interior viewpoint, aimed forward through the windshield at the interior rear-view mirror. The mirror remains in frame from first frame to last frame, same approximate size and screen position. No cut, no transition, no outside view, no drone, no tracking alongside the car, no shot through an open side window, no orbit, no angle reversal, no camera traveling through glass, no reframing to show the driver's body. This is a fixed IN-CAR camera position, not a fixed world-space camera. CAMERA MOTION: the camera is loosely damped inside the moving cabin, never detached from it. Tiny organic road-induced vibrations, micro-jolts, subtle 0.5–1.5 degree pitch/yaw, very small natural handheld sway (less than 2 cm); occasional gentle vibration blur. NO deliberate pan, dolly, push-in, zoom, tilt, orbit or camera path. The rear-view mirror stays clearly visible and anchored in approximately the same location throughout. The mirror is an actual optical reflection, not a floating screen or window. ACTING: The blonde woman is seated on the LEFT driver seat and keeps BOTH HANDS ON THE STEERING WHEEL for the whole shot. NO pointing. NO raised hand. NO hand entering the mirror. NO arm gestures. NO waving, no body turn, no reaching, no change of seats. Her attention stays on the salt road and the ramp directly ahead. ONCE ONLY, for a brief beat, she glances into the rear-view mirror toward the camera, then immediately returns her eyes to the road. In the mirror we see primarily her EYES, EYEBROWS and part of her forehead, exactly as in the input image. Do not reveal her whole face or mouth in the mirror. Her voice is heard from her physical position in the driver seat; do not force visible lip sync when her mouth is outside the mirror crop. MANDATORY AUDIBLE DIALOGUE IN BULGARIAN: The SAME woman says EXACTLY: „Виждаш ли оная рампа там?“ One confident, daring, breathy-adrenaline film-performance, with emphasis on „ОНАЯ РАМПА“. She begins within 0.3 seconds and finishes by about 2.8 seconds. Clearly audible native Bulgarian female voice, natural in-car acoustic reflections, perfectly timed to her acting. NO extra words, NO narration, NO subtitles, NO on-screen text. After finishing the sentence, she remains silent and focused on the ramp; the accelerating car and environment continue moving naturally. CAR MOTION: The Tesla is ALREADY DRIVING FAST and continues accelerating FORWARD, aligned with the salt track and heading STRAIGHT TOWARD THE CENTER OF THE SAME DISTANT RAMP. The road texture rushes under the windshield with increasingly strong optical flow. The ramp grows progressively but modestly larger, always DIRECTLY AHEAD, never sideways or behind the car. The original snow-covered volcano remains at its original relative position in the distance. Preserve ramp silhouette, salt texture, sky and mountain. No car turns, no sudden lane change, no reversing, no stop, no ramp morphing or collapsing. CINEMATIC IMAGE: authentic in-car action cinematography, ARRI ALEXA 35 / 50 mm spherical lens, T2.8, 24 fps, 180-degree shutter, natural subtle motion blur, high dynamic range, physically accurate windshield reflections, sun glare, rich restrained film contrast, organic film grain. Dramatic high-speed sensation comes from rushing ground texture, light/shadow flicker, wind and low-frequency road vibrations — NOT from moving the camera outside the car. The camera remains inside, with the same mirror visible for

Модел: GPT Image 2.5 Flare · изображение от изображение IMAGE TO IMAGE — R02 TELEPHOTO ACTION FILM INSERT — NEW SHOT FROM SCRATCH, NOT A REFRAME OF THE REFERENCE. UPLOAD: IMAGE 1 : a screenshot of the APPROVED view THROUGH the Tesla windshield showing the original distant white salt ramp, the cracked dry salt flat and the snowy mountain. This is a GEOGRAPHY AND STRUCTURE REFERENCE ONLY. DO NOT COPY ITS CAMERA ANGLE, WINDSHIELD, MIRROR, DASHBOARD, CAR INTERIOR, OR COMPOSITION. IMAGE 2 (OPTIONAL): the original clean salt flat and snowy mountain location photograph, to preserve the exact mountain silhouette and lighting. DO NOT UPLOAD any exterior Tesla shot. NO CAR is visible in this image. CREATE AN ENTIRELY NEW EXTERIOR PHOTOGRAPH, as though a professional action-film camera operator is filming the distant ramp from a remote observation point with an EXTREMELY LONG TELEPHOTO LENS. Camera is positioned FAR AWAY from the ramp on the same dry salt plain, near ground level at 1.5 meters high, looking directly toward the APPROACH SIDE of the ramp. Use an ARRI ALEXA 35 with a 300mm cinema telephoto lens, T4, 24fps movie-frame aesthetics, 180-degree shutter equivalent. Strong realistic TELEPHOTO COMPRESSION: the distant snowy mountain looks visually large behind the ramp, atmospheric depth compressed, fine layers of salt flats visible. The ramp is the VISUAL SUBJECT, centered slightly right of frame and visually dominant, with its full approach and overhanging top recognizable. Its broad white-salt roadway starts FLAT at the ground, transitions continuously upward into a vertical climb, then curls FORWARD into an overhead inverted lip. Maintain the same direction, same C-shaped forward curl, same white cracked salt texture, and same relative mountain placement as Image 1 . The roadway is wide enough for several cars. The structure is enormous, physically continuous and rooted in the salt flat, NOT a thin decorative sculpture, NOT a narrow pillar, NOT a separate loop, NOT a truncated wall. Composition: full-frame cinematic telephoto shot with the ramp around the central third, substantial salt foreground and snow-covered mountain behind. NO foreground car, NO driver, NO rear-view mirror, NO dashboard, NO windshield edges, NO artificial frame-within-frame. NO new roads diverging away from the ramp. The road approach on the salt flat clearly leads into the ramp entrance. Dramatic natural side light, filmic highlights, restrained grain, realistic heat haze/atmospheric shimmer, controlled contrast, photographic not CGI. Freeze a believable split second of an action-movie telephoto shot; do not paint visible shake or motion streaks into the still image — camera vibration will be animated in the subsequent video. The result must look like a real remote telephoto camera photographing the actual ramp in the established Generiram.bg location. ABSOLUTE: same mountain, same salt desert, same forward-curving ramp identity; COMPLETELY DIFFERENT CAMERA POSITION from the windshield reference. NO CAR IN FRAME. 16:9. OUTPUT R02-TELEPHOTO-RAMP-START.png.

Модел: LTX 2.5 Pro · видео от изображение · начален кадър: снимката отгоре IMAGE TO VIDEO — TELEPHOTO RAMP REVEAL — 3 SECONDS, 16:9, AUDIO OFF. START FRAME: the approved exterior 300mm telephoto photograph of the giant white salt ramp, dry salt plain and original snow-covered mountain. PRESERVE THE EXACT COMPOSITION AND PHYSICAL RAMP GEOMETRY THROUGHOUT. A single continuous, realistic 3-second high-tension action-film TELEPHOTO OBSERVATION SHOT from a remote camera position. Camera stays FAR AWAY from the ramp the entire time, as if a professional cinematographer is operating an ARRI ALEXA 35 with a 300mm cinema telephoto lens on a fluid-head support. The long lens creates powerful background compression and magnifies subtle physical vibrations. Add controlled, very small organic horizontal and vertical tremors from the operator and wind, delicate micro-corrections to keep the ramp centered, a tiny slow pan of no more than a few degrees, and a natural stabilized-yet-alive cinema-camera feel. The camera does NOT drive, fly, chase, zoom rapidly, dolly toward the ramp, or move inside a car. No cuts, no angle change, no whip pan, no sudden reframing. The ramp remains the subject throughout, its full drivable base and forward-curled inverted upper lip visible and geometrically stable. The road remains rooted to the cracked salt plain; it must not crumble, sink, deform, rotate, or transform. Preserve the same snowy mountain behind it. Subtle atmospheric shimmer, very fine dust stirred across the distant salt surface, realistic long-lens depth compression and restrained 24fps cinematic motion blur. No cars or people enter frame. NO DIALOGUE, NO MUSIC, NO DRONE. One unbroken external telephoto shot. Output RED-02-TELEPHOTO-RAMP.mp4.

Модел: Gemini Omni Flash · видео от изображение · началната снимка е кадър, взет от това видео Create a 7-second photorealistic cinematic video using the uploaded image as the exact starting frame. CAMERA: Sony VENICE 2, 200mm telephoto lens. Camera OUTSIDE the moving car, filming THROUGH the front windshield. Perform a continuous DOUBLE-ARC camera move. 0–2s: Orbit 12 degrees toward the passenger side. Reveal slightly more of the empty passenger seat. 2–5s: Smoothly reverse the orbital direction, returning toward the frontal perspective. Simultaneously begin a physical dolly-in toward the woman's mouth. 5–7s: Complete the dolly-in. Finish in an extreme close-up of her lips. Her lips occupy approximately 45% of the frame width. The camera tracks the car at matching speed. CRITICAL: Never reveal her eyes, eyebrows or forehead. Keep her upper face outside the frame. Preserve her exact nose, lips, blonde hair, ivory blazer, skin texture and seatbelt. The woman is alone in the vehicle. The car drives forward across a white salt flat. Natural reflections and subtle vibrations. DIALOGUE: 0.0–0.3s: Relaxed lips. 0.3–5.8s: She says in natural Bulgarian: "КОМЕНТИРАЙ „КАК“ и ще ти изпратим туториала и промптовете!" Emphasize "КАК". Natural energetic delivery. Precise Bulgarian lip synchronization. 5.8–7.0s: Subtle smile as the camera finishes its move. AUDIO: Natural Bulgarian female voice. Subtle electric vehicle ambience. No music, narrator or additional voices. One continuous shot. No cuts, subtitles, text overlays or other people. 7 seconds, 16:9.

Модел: Gemini Omni Flash · видео от изображение · начален кадър: снимката отгоре Generate a 7-second ultra-photorealistic cinematic automotive commercial video from the uploaded image. STARTING FRAME: Match the uploaded image exactly. CAMERA: Sony VENICE 2, 200mm telephoto cinema lens. The camera is OUTSIDE the moving car, filming through the front windshield. Execute one continuous orbital dolly-in. Over 7 seconds, smoothly orbit 20 degrees toward the PASSENGER SIDE while physically moving closer to the woman's mouth. The camera tracks the moving car at matching speed. The empty passenger seat becomes more visible. The steering wheel shifts naturally across the lower foreground. The woman's mouth remains the focal point. By the end, her lips occupy approximately 45% of the frame width. CRITICAL FRAMING: NEVER reveal her eyes, eyebrows or forehead. Keep the upper half of her face outside the frame. No zoom-out or upward reframing. Preserve her exact lips, nose, blonde bob, skin texture, ivory blazer and seatbelt. The woman is alone in the car. The vehicle drives forward across a white salt flat. Natural windshield reflections and subtle vibrations. TIMING: 0.0–0.4s: Her lips are relaxed as the orbital move begins. 0.4–4.6s: She speaks clearly in natural Bulgarian: "Линкът за туториала и промптовете е в биото!" Friendly, confident delivery. Emphasize "Линкът" and "биото". Accurate Bulgarian pronunciation and lip sync. 4.6–7.0s: She finishes with a subtle smile. The camera completes its push-in. AUDIO: One natural Bulgarian female voice. Subtle electric vehicle ambience. No music or additional voices. One continuous shot. No cuts, transitions, subtitles or text overlays. No additional people. DURATION: 7 SECONDS. ASPECT RATIO: 16:9.

Модел: Seedance 2.5 · видео от референции, после Topaz Upscale за по-висока резолюция SEEDANCE 2.5 — REFERENCE TO VIDEO 11-SECOND CINEMATIC TESLA STUNT COMMERCIAL FOUR STRICT VISUAL REFERENCES ULTRA-PHOTOREALISTIC HOLLYWOOD AUTOMOTIVE ACTION GIANT CURLING RAMP — INVERTED LAUNCH 180-DEGREE AIRBORNE HALF-ROLL SPECTACULAR AERODYNAMIC VAPOR VORTICES DISTANT SALT-FLAT LANDING FINAL GENERIRAM.BG BRAND REVEAL FORMAT: 16:9 TARGET RESOLUTION: 1080p TOTAL DURATION: EXACTLY 11 SECONDS ENVIRONMENT: BRIGHT NATURAL DAYLIGHT DIALOGUE: NONE NARRATION: NONE ============================================================ 01 — EXACT REFERENCE IMAGE ASSIGNMENTS ============================================================ Exactly FOUR reference images are provided. Their roles are fixed and must not be confused. @Image1 — OFFICIAL GENERIRAM.BG LOGO This is the authoritative reference for the exact Generiram.bg brand identity. Preserve the original logo with maximum fidelity: - Exact symbol and icon geometry. - Exact lettering and spelling: Generiram.bg. - Exact typography and letter spacing. - Exact original colors. - Exact proportions and arrangement. Use this reference ONLY for the final brand reveal. The logo must NOT appear during the stunt. Do not place it on the Tesla, ramp, mountain, road surface or in the sky before the ending. @Image2 — GIANT CURVED RAMP AND SNOWY MOUNTAIN This is the PRIMARY ENVIRONMENT REFERENCE. Recreate the SAME enormous curling stunt ramp shown in @Image2 . Preserve its recognizable silhouette, scale, orientation, material appearance and relationship to the distant mountain. The ramp is one monumental continuous structure. Its driving surface begins horizontally, then curves progressively upward. The slope becomes increasingly steep. The surface reaches a near-vertical section. It then continues into an inverted, overhanging upper launch section. The entire structure resembles an enormous curling wave or open C-shaped ramp. The ramp is NOT: - A straight launch wedge. - A complete circular loop. - A spiral. - A disconnected elevated roadway. - A vertical wall with a separate platform. - A small skateboard ramp. The ramp must remain many times larger than the Tesla. The vehicle should appear relatively small against the monumental structure. MOUNTAIN: Preserve the SAME distinctive snow-covered volcanic mountain from @Image2 . Keep the mountain in the SAME geographical position relative to the ramp. The mountain is far behind the ramp. It must not move or change shape between shots. ENVIRONMENT: An immense, dry, white salt flat. Subtle cracked salt patterns on the ground. A flat distant horizon. Pale blue daylight sky. Natural sunlight. Soft atmospheric haze. Realistic daylight shadows. No nighttime environment. No black background. No city. No buildings. No spectators. @Image3 — WHITE TESLA, FRONT THREE-QUARTER VIEW This is the PRIMARY VEHICLE IDENTITY REFERENCE. Reproduce the EXACT SAME futuristic white Tesla. Preserve: - Front fascia. - Headlights. - Hood. - Windshield. - Roofline. - Side windows. - Wheel design. - Wheelbase. - Body proportions. - White body paint. - Black trim. - Overall vehicle silhouette. Do not replace it with another Tesla model. Do not create a generic sports car. @Image4 — WHITE TESLA, REAR THREE-QUARTER VIEW This is the SECONDARY VEHICLE IDENTITY REFERENCE. @Image3 and 4 show ONE IDENTICAL Tesla. Preserve: - Rear body shape. - Taillights. - Rear bumper. - Rear glass. - Wheel design. - Rear aerodynamic details. - White paint. - Black trim. - Vehicle proportions. The Tesla must remain visually identical throughout the entire 11-second commercial. ============================================================ 02 — THE COMPLETE STUNT: ABSOLUTE PRIORITY ============================================================ The white Tesla accelerates across th