탑다운 평면도를 기준으로 배경 세트의 위치 일관성을 100% 유지하며 대화형 비디오를 제작할 때 사용합니다.
{
"refs": {
" @/img1 ": "LOCATION MAP. 90 degree top-down view, north is up. Treat it as a floor plan; every element keeps its position in all shots.",
" @/img2 ": "CHARACTER. Match face, freckles, hair and outfit exactly."
},
"set_layout": "Round teal pond with lily pads at center. Wooden jetty on the WEST shore. Bench with a red scarf just SOUTH of the pond, slightly west. Dirt path enters from the SOUTH and curves past the bench to the jetty. White heron in the shallows on the SOUTH-EAST shore. Reeds on the NW, NE, E and S shores. Birches and bushes ring the meadow.",
"character": "Young woman, long platinum blonde hair, freckles, blue eyes, white tee, grey denim overalls, white sneakers, gold necklaces.",
"performance": "Sassy, relaxed, talks like a friend spilling a secret. Natural pauses, breaths, a half laugh. Accurate lip sync. Eyebrow raises, eye rolls, smirks, glances off camera. Loose expressive hands matching the words.",
"look": "Photoreal cinematic, 35mm anamorphic, golden morning sun from the east, soft haze, natural skin texture, 16:9, 30s",
"shots": [
{"t": "0-4s", "cam": "Low wide from the south end of the path looking north, slow push in", "act": "Walks up the path, tucks hair behind her ear, rolls her eyes on 'total mess', flicks her hand", "line": "Okay so... AI video and keeping a location consistent? Total mess. Until now."},
{"t": "4-8s", "cam": "Wide from the east shore looking west across the pond, slow lateral track", "act": "Strolls the path, holds up one finger, draws a flat square in the air", "line": "One top-down shot of your set. Straight down, ninety degrees."},
{"t": "8-11s", "cam": "Wide from the north shore looking south, reeds in foreground", "act": "Reaches the bench, heron at frame left, points in four quick directions, tilts her head", "line": "North, east, south, west. The model reads it like a map."},
{"t": "11-16s", "cam": "Over the shoulder from behind the bench looking north at the pond", "act": "Sits, flicks the red scarf over her shoulder, gestures left to the jetty, right to the heron", "line": "Jetty stays west. Bench stays south. And that heron? Hasn't moved. Not once."},
{"t": "16-20s", "cam": "Medium close up from the pond edge looking south at her on the bench", "act": "Leans back, slow smirk, small laugh, points at camera", "line": "Cut from anywhere you want. Same pond. Same set. I'm kind of obsessed."},
{"t": "20-24s", "cam": "Extreme low angle from far at the north-east meadow edge, lens in the grass looking south-west, blurred wildflowers in front, towering birches, jetty right, heron left, she small on the bench", "act": "Spreads both arms along the bench backrest, then raises one finger high", "line": "And no, I didn't build a 3D scene. One image. That's the whole trick."},
{"t": "24-27s", "cam": "Wide profile from the end of the jetty looking east", "act": "Crosses her legs, shrugs with open palms, mock humble face", "line": "Location consistency, darling. It's not that deep."},
{"t": "27-30s", "cam": "Smooth crane rising to a 90 degree overhead view matching image_1", "act": "Looks up, waves, blows a playful kiss", "line": "You're welcome."
],
"audio": "Diegetic only: her voice (distant in the far shot), footsteps on dirt, birdsong, water lapping, reeds rustling, bench creak",
"continuity": "Pond, jetty, bench, heron and birches stay where image_1 places them. Same outfit and light throughout."
}{
"refs": {
" @/img1 ": "장소 지도. 90도 탑다운 뷰, 북쪽이 위. 평면도로 취급하며 모든 요소는 모든 샷에서 위치를 유지함.",
" @/img2 ": "캐릭터. 얼굴, 주근깨, 헤어, 의상을 정확히 일치시킬 것."
},
"set_layout": "중앙에 수련이 있는 청록색 원형 연못. 서쪽 기슭에 나무 선착장. 연못 남쪽 바로 옆 약간 서쪽에 빨간 스카프가 걸린 벤치. 남쪽에서 흙길이 시작되어 벤치를 지나 선착장으로 굽어짐. 남동쪽 기슭 얕은 물가에 백로. 북서, 북동, 동, 남쪽 기슭에 갈대. 자작나무와 덤불이 초원을 둘러쌈.",
"character": "젊은 여성, 긴 플래티넘 블론드 헤어, 주근깨, 파란 눈, 흰 티셔츠, 회색 데님 멜빵바지, 흰 스니커즈, 금목걸이.",
"performance": "당돌하고 여유로움, 비밀을 털어놓는 친구처럼 대화함. 자연스러운 쉼, 호흡, 가벼운 웃음. 정확한 립싱크. 눈썹 올림, 눈동자 굴림, 비웃음, 카메라 밖 주시. 말에 맞춘 자연스러운 손짓.",
"look": "실사 아나모픽 35mm 영화 느낌, 동쪽에서 비추는 따뜻한 아침 햇살, 부드러운 안개, 자연스러운 피부 질감, 16:9, 30초",
"shots": [
{"t": "0-4s", "cam": "남쪽 길 끝에서 북쪽을 바라보는 로우 와이드, 천천히 푸시인", "act": "길을 따라 걸어오며 머리를 귀 뒤로 넘기고, 'total mess' 부분에서 눈을 굴리며 손을 까딱임", "line": "자... AI 비디오에서 장소 일관성을 유지하는 거? 완전 난장판이었죠. 지금까지는요."},
{"t": "4-8s", "cam": "동쪽 기슭에서 연못 건너 서쪽을 바라보는 와이드, 천천히 측면 트래킹", "act": "길을 산책하며 손가락 하나를 들고 공중에 사각형을 그림", "line": "세트의 탑다운 샷 한 장이면 돼요. 수직 90도 직하향으로요."},
{"t": "8-11s", "cam": "북쪽 기슭에서 남쪽을 바라보는 와이드, 전경에 갈대", "act": "벤치에 도달함, 프레임 좌측에 백로, 동서남북 4방향을 빠르게 가리키며 고개를 기웃함", "line": "동서남북. 모델이 이걸 지도처럼 읽어내거든요."},
{"t": "11-16s", "cam": "벤치 뒤에서 연못 북쪽을 바라보는 오버더숄더", "act": "앉아서 빨간 스카프를 어깨 뒤로 넘기고, 좌측 선착장과 우측 백로를 가리킴", "line": "선착장은 서쪽에, 벤치는 남쪽에 그대로 있죠. 저 백로요? 한 번도 안 움직였어요."},
{"t": "16-20s", "cam": "연못가에서 벤치에 앉은 그녀를 남쪽으로 바라보는 미디엄 클로즈업", "act": "등을 기대며 미소를 지음, 가벼운 웃음 후 카메라를 가리킴", "line": "어느 각도에서 컷을 전환하든 같은 연못, 같은 세트예요. 진짜 대박이지 않나요."},
{"t": "20-24s", "cam": "북동쪽 초원 끝 먼 곳에서 풀숲 사이로 남서쪽을 바라보는 익스트림 로우 앵글, 아웃포커스된 야생화와 우뚝 선 자작나무, 우측 선착장, 좌측 백로, 벤치에 조그맣게 보이는 그녀", "act": "양팔을 벤치 등받이에 벌려 얹었다가 손가락 하나를 높이 든다", "line": "3D 씬을 만든 것도 아니에요. 이미지 단 한 장. 그게 비결전부예요."},
{"t": "24-27s", "cam": "선착장 끝에서 동쪽을 바라보는 와이드 프로필", "act": "다리를 꼬고 손바닥을 펼치며 으쓱함, 겸손한 척하는 표정", "line": "장소 일관성이요? 별거 없답니다."},
{"t": "27-30s", "cam": "image_1과 일치하는 90도 오버헤드 뷰로 상승하는 부드러운 크레인 샷", "act": "위쪽을 바라보며 손을 흔들고 장난스럽게 손키스를 날림", "line": "천만에요."
],
"audio": "동시녹음 음향만: 그녀의 목소리(원거리 샷에서는 멀게 들림), 흙길 발자국 소리, 새소리, 물결 소리, 갈대 꺾이는 소리, 벤치 삐걱임",
"continuity": "연못, 선착장, 벤치, 백로, 자작나무는 image_1이 배치한 위치에 그대로 유지. 전체 영상에서 동일한 의상과 조명 유지."
}출처 · prompts.chat · Location-Locked Talking Video: Seedance 2.5 JSON Prompt That Uses a Top-Down Map to Keep Every Shot on the Same Set (30s, 8 Shots) — 원문은 출처에 게시된 그대로이며, 권리는 원작자에게 있습니다.
mopick 개선안
원문을 그대로 넣으면 AI 가 매번 다르게 답하거나 중요한 부분을 빠뜨릴 수 있습니다. 개선안은 원문의 뜻과 언어는 그대로 두고 역할 · 바꿔 넣을 입력 칸 · 결과 형식 · 지킬 조건을 보완해 다시 쓴 버전입니다. 원문과 같은 곳(ChatGPT · Claude 등)에 그대로 붙여 넣어 쓰면 됩니다.
AI 평가 점수 — 명확성 · 구체성 · 결과 형식 · 재사용성 기준(100점)
입력 칸레슨 05 배우기 ›
원문 · 없음바꿔 넣을 자리가 표시돼 있지 않아 쓸 때마다 원문을 직접 고쳐야 했습니다.
개선 · 2개[Sora / Runway Gen-3 / Luma Video Engine] · [OPTIONAL_STYLE_TWEAKS]
이 개선안에 쓰인 기법 — 프롬프트 클래스에서 배우기
원문
{
"refs": {
" @/img1 ": "LOCATION MAP. 90 degree top-down view, north is up. Treat it as a floor plan; every element keeps its position in all shots.",
" @/img2 ": "CHARACTER. Match face, freckles, hair and outfit exactly."
},
"set_layout": "Round teal pond with lily pads at center. Wooden jetty on the WEST shore. Bench with a red scarf just SOUTH of the pond, slightly west. Dirt path enters from the SOUTH and curves past the bench to the jetty. White heron in the shallows on the SOUTH-EAST shore. Reeds on the NW, NE, E and S shores. Birches and bushes ring the meadow.",
"character": "Young woman, long platinum blonde hair, freckles, blue eyes, white tee, grey denim overalls, white sneakers, gold necklaces.",
"performance": "Sassy, relaxed, talks like a friend spilling a secret. Natural pauses, breaths, a half laugh. Accurate lip sync. Eyebrow raises, eye rolls, smirks, glances off camera. Loose expressive hands matching the words.",
"look": "Photoreal cinematic, 35mm anamorphic, golden morning sun from the east, soft haze, natural skin texture, 16:9, 30s",
"shots": [
{"t": "0-4s", "cam": "Low wide from the south end of the path looking north, slow push in", "act": "Walks up the path, tucks hair behind her ear, rolls her eyes on 'total mess', flicks her hand", "line": "Okay so... AI video and keeping a location consistent? Total mess. Until now."},
{"t": "4-8s", "cam": "Wide from the east shore looking west across the pond, slow lateral track", "act": "Strolls the path, holds up one finger, draws a flat square in the air", "line": "One top-down shot of your set. Straight down, ninety degrees."},
{"t": "8-11s", "cam": "Wide from the north shore looking south, reeds in foreground", "act": "Reaches the bench, heron at frame left, points in four quick directions, tilts her head", "line": "North, east, south, west. The model reads it like a map."},
{"t": "11-16s", "cam": "Over the shoulder from behind the bench looking north at the pond", "act": "Sits, flicks the red scarf over her shoulder, gestures left to the jetty, right to the heron", "line": "Jetty stays west. Bench stays south. And that heron? Hasn't moved. Not once."},
{"t": "16-20s", "cam": "Medium close up from the pond edge looking south at her on the bench", "act": "Leans back, slow smirk, small laugh, points at camera", "line": "Cut from anywhere you want. Same pond. Same set. I'm kind of obsessed."},
{"t": "20-24s", "cam": "Extreme low angle from far at the north-east meadow edge, lens in the grass looking south-west, blurred wildflowers in front, towering birches, jetty right, heron left, she small on the bench", "act": "Spreads both arms along the bench backrest, then raises one finger high", "line": "And no, I didn't build a 3D scene. One image. That's the whole trick."},
{"t": "24-27s", "cam": "Wide profile from the end of the jetty looking east", "act": "Crosses her legs, shrugs with open palms, mock humble face", "line": "Location consistency, darling. It's not that deep."},
{"t": "27-30s", "cam": "Smooth crane rising to a 90 degree overhead view matching image_1", "act": "Looks up, waves, blows a playful kiss", "line": "You're welcome."
],
"audio": "Diegetic only: her voice (distant in the far shot), footsteps on dirt, birdsong, water lapping, reeds rustling, bench creak",
"continuity": "Pond, jetty, bench, heron and birches stay where image_1 places them. Same outfit and light throughout."
}개선안 · 색칠한 [칸]에 내 내용을 넣어 쓰세요
{ "project_meta": { "target_engine": "[Sora / Runway Gen-3 / Luma Video Engine]", "aspect_ratio": "16:9", "fps": 30, "duration": "30s" }, "refs": { "@/img1": "LOCATION MAP. 90 degree top-down view, north is up. Treat it as a floor plan; every element keeps its exact position in all shots.", "@/img2": "CHARACTER. Match face, freckles, hair and outfit exactly." }, "set_layout": "Round teal pond with lily pads at center. Wooden jetty on the WEST shore. Bench with a red scarf just SOUTH of the pond, slightly west. Dirt path enters from the SOUTH and curves past the bench to the jetty. White heron in the shallows on the SOUTH-EAST shore. Reeds on the NW, NE, E and S shores. Birches and bushes ring the meadow.", "character": "Young woman, long platinum blonde hair, freckles, blue eyes, white tee, grey denim overalls, white sneakers, gold necklaces. [OPTIONAL_STYLE_TWEAKS]", "performance": "Sassy, relaxed, talks like a friend spilling a secret. Natural pauses, breaths, a half laugh. Precise 1:1 lip sync. Eyebrow raises, eye rolls, smirks, glances off camera. Loose expressive hands matching the words.", "look": "Photoreal cinematic, 35mm anamorphic, golden morning sun from the east, soft haze, natural skin texture, 16:9, 30s", "shots": [ {"t": "0-4s", "cam": "Low wide from the south end of the path looking north, slow push in", "act": "Walks up the path, tucks hair behind her ear, rolls her eyes on 'total mess', flicks her hand", "line": "Okay so... AI video and keeping a location consistent? Total mess. Until now."}, {"t": "4-8s", "cam": "Wide from the east shore looking west across the pond, slow lateral track", "act": "Strolls the path, holds up one finger, draws a flat square in the air", "line": "One top-down shot of your set. Straight down, ninety degrees."}, {"t": "8-11s", "cam": "Wide from the north shore looking south, reeds in foreground", "act": "Reaches the bench, heron at frame left, points in four quick directions, tilts her head", "line": "North, east, south, west. The model reads it like a map."}, {"t": "11-16s", "cam": "Over the shoulder from behind the bench looking north at the pond", "act": "Sits, flicks the red scarf over her shoulder, gestures left to the jetty, right to the heron", "line": "Jetty stays west. Bench stays south. And that heron? Hasn't moved. Not once."}, {"t": "16-20s", "cam": "Medium close up from the pond edge looking south at her on the bench", "act": "Leans back, slow smirk, small laugh, points at camera", "line": "Cut from anywhere you want. Same pond. Same set. I'm kind of obsessed."}, {"t": "20-24s", "cam": "Extreme low angle from far at the north-east meadow edge, lens in the grass looking south-west, blurred wildflowers in front, towering birches, jetty right, heron left, she small on the bench", "act": "Spreads both arms along the bench backrest, then raises one finger high", "line": "And no, I didn't build a 3D scene. One image. That's the whole trick."}, {"t": "24-27s", "cam": "Wide profile from the end of the jetty looking east", "act": "Crosses her legs, shrugs with open palms, mock humble face", "line": "Location consistency, darling. It's not that deep."}, {"t": "27-30s", "cam": "Smooth crane rising to a 90 degree overhead view matching image_1", "act": "Looks up, waves, blows a playful kiss", "line": "You're welcome." ], "audio": { "dialogue": "Diegetic only: her voice (spatially distant in far shot)", "ambiance": "Footsteps on dirt, birdsong, water lapping, reeds rustling, bench creak" }, "continuity": "Strict spatial lock based on @/img1. Pond, jetty, bench, heron, and birches remain fixed. Same outfit and golden lighting throughout." }