부산 · 부울경 기반 전국 출강 · 공공기관 · 대학 · 직업/평생교육기관 강의강의 문의 010-4894-0871[email protected]
홈 / 자료 공유 / 공유 프롬프트 10

친구가 찍어줬나봐
연말 사진 프롬프트 5종

셀카도 삼각대도 아닌, 친구가 옆에서 찍어준 것 같은 연말 사진을 내 사진 한 장으로 만드는 프롬프트입니다. 컷별 프롬프트 5개와 얼굴만 바꾸는 편집 프롬프트, 안 닮을 때 쓰는 보정 문장 8개를 함께 담았습니다.

컷 프롬프트 5개 보정 문장 8개 ChatGPT 이미지 2026.11 기준
실제 사람 사진으로 만드는 프롬프트입니다. 본인 사진이나 동의를 받은 사진만 쓰세요. 이미지 결과는 모델·버전에 따라 다를 수 있습니다. 출처를 밝히면 자유롭게 인용·공유하셔도 됩니다.

사진 한 장만 잘 고르면 됩니다

프롬프트는 그대로 두고 첨부하는 사진만 바꿉니다. 결과는 첨부한 사진에 가장 크게 좌우됩니다.

이런 사진이 좋습니다이런 사진은 피하세요
정면이거나 15도 안쪽으로 살짝 돌린 얼굴아래에서 위로 찍은 셀카
밝고 고른 빛 (창가 낮 빛이 가장 좋음)필터·보정 앱을 강하게 쓴 사진
얼굴이 화면에 크게, 선명하게마스크·모자·선글라스·손이 얼굴을 가린 사진
평소 내 얼굴과 비슷한 사진어둡거나 한쪽만 밝은 사진
1

새 대화를 열고 내 사진 1장을 첨부합니다

2

원하는 컷의 프롬프트를 통째로 붙여넣고 보냅니다

3

2~3번 뽑아 가장 자연스러운 걸 고릅니다

4

덜 닮거나 어색하면 보정 문장을 같은 대화에 이어서 보냅니다

ChatGPT 이미지는 생성할 때마다 얼굴이 조금씩 달라집니다. 한 번에 똑같이 나오지 않아도 정상입니다. 비율은 세로 4:5로 지정했지만 다르게 나오면 잘라서 쓰세요.

"누가 찍어준 사진"은 이렇게 만들어집니다

혼자 찍은 사진과 남이 찍어준 사진은 세 가지가 다릅니다. 프롬프트는 이 세 가지를 글로 적어 넣은 것입니다.

다른 점혼자 찍은 사진누가 찍어준 사진프롬프트에서
거리팔 길이 안1.5~2m 떨어져서CAMERA & FRAMING
순간준비된 포즈웃다가, 돌아보다가MOMENT
시선렌즈 정면렌즈 밖, 또는 흘끗EXPRESSION & GAZE

프롬프트 하나는 세 덩어리입니다. LOCK은 첨부한 사진의 얼굴을 그대로 쓰라는 문단으로 모든 컷이 같습니다. SCENE부터 CAMERA까지는 장소·의상·순간·표정·일행·빛·구도 일곱 줄로, 컷마다 이 부분만 다릅니다. FINISH는 폰으로 찍은 캔디드 사진처럼 보이게 하는 문단으로 역시 모든 컷이 같습니다. 영어로 쓴 이유는 OpenAI 공식 이미지 가이드의 권장 표현을 가져왔고, 사진 용어가 영어에서 더 짧고 정확하기 때문입니다.

LOCK · 얼굴 고정

LOCK — IDENTITY (must not change):
Use the person in the attached photo as the only identity reference. Keep her exact face: face shape and width, jawline, eye shape and eyelids, eyebrow shape, nose bridge and tip, lip shape, skin tone, any moles or marks, natural facial asymmetry, and apparent age. Keep her own hairstyle, hair length and hair color from the reference. Do not beautify, slim the face, enlarge the eyes, or turn her into a generic model — she must be instantly recognizable as the same real person.
The reference photo's background, clothing, camera angle, lighting and neutral expression are for identity only: do NOT copy them. Re-light her naturally for the scene and give her the expression and pose described below.
Only the main woman uses this identity. Any companions are different people with different faces.

얼굴형·눈·코·입·피부톤·점을 그대로 두고, 미화하지 말라는 지시입니다. 첨부 사진의 배경·옷·표정은 따라 하지 말라는 문장이 있어 증명사진을 넣어도 회색 배경이 따라 나오지 않습니다.

FINISH · 사진 질감

FINISH — PHOTO LOOK:
Photorealistic candid photo, taken by a friend on a recent smartphone's main camera (about 26mm equivalent), handheld — not a professional shoot. Honest and unposed. Real skin texture with visible pores, fine baby hairs and slight natural unevenness; no beauty filter, no airbrushing, no heavy retouching, no glamorization.
Slightly imperfect amateur framing: subject a little off-center, slight tilt, edges may crop part of the scene. Smartphone image traits: mild noise in shadows, slight motion blur on moving hands or hair, white balance from the real light sources, natural colors with no cinematic color grading, no studio lighting, no exaggerated bokeh.
No text, no readable signs, no logos, no brand marks, no watermark. Vertical 4:5 composition.

폰 기본 카메라(약 26mm), 손으로 들고 찍은 느낌, 모공이 보이는 피부, 살짝 삐뚤어진 구도를 요구합니다. 화보처럼 보이는 보정·색보정·스튜디오 조명은 빼라고 적었습니다.

컷 5종

01크리스마스 마켓 · 이름 불려서 뒤돌아본 순간02겨울바다 산책로 · 바닷바람에 머리 날린 순간03연말 홈파티 · 건배하다 크게 웃은 순간04겨울 카페 창가 · 카드 쓰다 몰래 찍힌 순간05백화점 대형 트리 · 트리 찍는 나를 찍어준 순간
CUT 01–05

연말 장면 다섯 컷

내 사진 1장을 첨부한 뒤 전체 프롬프트를 통째로 붙여넣으세요. LOCK과 FINISH가 이미 들어 있습니다.

01

크리스마스 마켓 · 이름 불려서 뒤돌아본 순간

장소야외 크리스마스 마켓, 나무 부스와 전구
시간·빛12월 초 밤, 따뜻한 전구 + 차가운 밤공기, 입김
의상크림 크롭 무스탕 + 버건디 골지 니트 + 진청 데님
순간앞서 걷다 친구가 이름을 불러 뒤돌아보며 웃음
구도2m 뒤 가슴 높이, 살짝 기울어짐, 허벅지 위부터

주의 · 뒤돌아보는 각도가 커지면 얼굴이 덜 닮습니다. 컵·간판 글자, 컵 든 손을 확인하세요.

LOCK — IDENTITY (must not change):
Use the person in the attached photo as the only identity reference. Keep her exact face: face shape and width, jawline, eye shape and eyelids, eyebrow shape, nose bridge and tip, lip shape, skin tone, any moles or marks, natural facial asymmetry, and apparent age. Keep her own hairstyle, hair length and hair color from the reference. Do not beautify, slim the face, enlarge the eyes, or turn her into a generic model — she must be instantly recognizable as the same real person.
The reference photo's background, clothing, camera angle, lighting and neutral expression are for identity only: do NOT copy them. Re-light her naturally for the scene and give her the expression and pose described below.
Only the main woman uses this identity. Any companions are different people with different faces.
SCENE: An outdoor Christmas market on a city plaza in Korea at night in early December. Rows of small wooden booths with warm string lights and pine garlands, a large lit Christmas tree far in the background, other visitors softly out of focus. All booth signs are blurred and unreadable.
OUTFIT: Cream cropped shearling (mustang-style) jacket worn open over a burgundy ribbed knit top, straight dark-wash jeans, small plain black shoulder bag. She holds a plain paper cup of hot mulled wine with no print.
MOMENT: She was walking a few steps ahead of her friend and has just turned back over her shoulder because the friend called her name — caught mid-turn, bursting into laughter.
EXPRESSION & GAZE: Genuine open laugh, eyes slightly squinted, looking back toward the camera without posing. Face in three-quarter view and fully visible — not hidden by hair, collar or the cup.
COMPANIONS: None in focus; her friend is the one taking the photo. Only background visitors.
LIGHT: Warm tungsten string lights from above and the side mixed with cooler ambient night light; faint breath vapor in the cold air.
CAMERA & FRAMING: Shot from about 2 meters behind her at chest height, slightly tilted; she is off-center to the right, framed from mid-thigh up; slight motion blur on the hand holding the cup.
FINISH — PHOTO LOOK:
Photorealistic candid photo, taken by a friend on a recent smartphone's main camera (about 26mm equivalent), handheld — not a professional shoot. Honest and unposed. Real skin texture with visible pores, fine baby hairs and slight natural unevenness; no beauty filter, no airbrushing, no heavy retouching, no glamorization.
Slightly imperfect amateur framing: subject a little off-center, slight tilt, edges may crop part of the scene. Smartphone image traits: mild noise in shadows, slight motion blur on moving hands or hair, white balance from the real light sources, natural colors with no cinematic color grading, no studio lighting, no exaggerated bokeh.
No text, no readable signs, no logos, no brand marks, no watermark. Vertical 4:5 composition.
02

겨울바다 산책로 · 바닷바람에 머리 날린 순간

장소부산 바닷가 데크 산책로
시간·빛맑은 겨울 오후 3~4시, 옆에서 드는 낮은 해
의상아이보리 숏패딩 + 회색 니트 헤어밴드 + 차콜 머플러
순간바람에 머리가 날려 웃으며 머리카락을 걷어냄
구도눈높이 1.5m, 허리 위, 수평선 기울어짐, 난간이 아래를 자름

주의 · 머리카락이 눈·코를 가리면 다른 사람처럼 보입니다. 그럴 땐 다시 뽑으세요.

LOCK — IDENTITY (must not change):
Use the person in the attached photo as the only identity reference. Keep her exact face: face shape and width, jawline, eye shape and eyelids, eyebrow shape, nose bridge and tip, lip shape, skin tone, any moles or marks, natural facial asymmetry, and apparent age. Keep her own hairstyle, hair length and hair color from the reference. Do not beautify, slim the face, enlarge the eyes, or turn her into a generic model — she must be instantly recognizable as the same real person.
The reference photo's background, clothing, camera angle, lighting and neutral expression are for identity only: do NOT copy them. Re-light her naturally for the scene and give her the expression and pose described below.
Only the main woman uses this identity. Any companions are different people with different faces.
SCENE: A seaside promenade in Busan, Korea, on a cold, clear winter afternoon. Wooden boardwalk railing, a wide sandy beach and dark blue winter sea behind her, a few people walking far away, high-rise buildings faint along the shoreline. No signs.
OUTFIT: Ivory short puffer jacket zipped halfway, a wide grey chunky knit headband covering her ears, a charcoal knit scarf loosely wrapped, light-wash wide jeans.
MOMENT: A strong gust of sea wind has just blown her hair across; she is laughing while pulling strands away from her face with one hand, the other hand holding her scarf.
EXPRESSION & GAZE: Laughing with her mouth slightly open, glancing toward the camera mid-laugh. Her eyes, nose and mouth stay clearly visible — hair moves around the face but does not cover it.
COMPANIONS: None in frame; her friend is taking the photo.
LIGHT: Low winter sun around 3–4 pm from the side: crisp, slightly cool seaside daylight with warm sunlight on one cheek.
CAMERA & FRAMING: Eye level from about 1.5 meters, framed from the waist up, horizon slightly tilted, she is left of center; part of the railing cuts into the bottom corner of the frame.
FINISH — PHOTO LOOK:
Photorealistic candid photo, taken by a friend on a recent smartphone's main camera (about 26mm equivalent), handheld — not a professional shoot. Honest and unposed. Real skin texture with visible pores, fine baby hairs and slight natural unevenness; no beauty filter, no airbrushing, no heavy retouching, no glamorization.
Slightly imperfect amateur framing: subject a little off-center, slight tilt, edges may crop part of the scene. Smartphone image traits: mild noise in shadows, slight motion blur on moving hands or hair, white balance from the real light sources, natural colors with no cinematic color grading, no studio lighting, no exaggerated bokeh.
No text, no readable signs, no logos, no brand marks, no watermark. Vertical 4:5 composition.
03

연말 홈파티 · 건배하다 크게 웃은 순간

장소아파트 거실, 케이크와 작은 트리
시간·빛12월 중순 밤, 폰 플래시 + 뒤쪽 트리 전구
의상크림·네이비 페어아일 니트, 작은 실버 귀걸이
순간건배하려 몸을 숙이며 친구 말에 크게 웃음
구도테이블 맞은편에서 살짝 위, 잔이 위쪽에 걸리게 꽉 차게

주의 · 친구 얼굴이 내 얼굴을 닮거나, 잔과 손이 겹쳐 깨질 수 있습니다.

LOCK — IDENTITY (must not change):
Use the person in the attached photo as the only identity reference. Keep her exact face: face shape and width, jawline, eye shape and eyelids, eyebrow shape, nose bridge and tip, lip shape, skin tone, any moles or marks, natural facial asymmetry, and apparent age. Keep her own hairstyle, hair length and hair color from the reference. Do not beautify, slim the face, enlarge the eyes, or turn her into a generic model — she must be instantly recognizable as the same real person.
The reference photo's background, clothing, camera angle, lighting and neutral expression are for identity only: do NOT copy them. Re-light her naturally for the scene and give her the expression and pose described below.
Only the main woman uses this identity. Any companions are different people with different faces.
SCENE: A small year-end house party in a Korean apartment living room in mid-December. A low table with a homemade- looking cake, finger food on plates and glasses of sparkling drink; a small tabletop Christmas tree with fairy lights in the background; floor cushions.
OUTFIT: Cream and navy Fair Isle (Nordic pattern) crew-neck sweater, small silver hoop earrings.
MOMENT: Everyone is clinking glasses for a toast; she leans forward to clink and is laughing hard at something a friend just said.
EXPRESSION & GAZE: Big natural laugh, eyes on the glasses and her friend, not on the camera. Face in three-quarter view and fully visible.
COMPANIONS: Two or three friends only partly in frame — their hands and glasses entering from the edges, one friend's shoulder blurred in the foreground. Friends' faces are cropped out of frame or turned away.
LIGHT: Direct on-camera smartphone flash at night: brightly lit subject, slightly harsh shadow behind her, warm fairy lights glowing in the darker background.
CAMERA & FRAMING: Taken from across the table at seated height, slightly high angle, tight crop with the clinking glasses in the upper part of the frame; slight motion blur on the glasses.
FINISH — PHOTO LOOK:
Photorealistic candid photo, taken by a friend on a recent smartphone's main camera (about 26mm equivalent), handheld — not a professional shoot. Honest and unposed. Real skin texture with visible pores, fine baby hairs and slight natural unevenness; no beauty filter, no airbrushing, no heavy retouching, no glamorization.
Slightly imperfect amateur framing: subject a little off-center, slight tilt, edges may crop part of the scene. Smartphone image traits: mild noise in shadows, slight motion blur on moving hands or hair, white balance from the real light sources, natural colors with no cinematic color grading, no studio lighting, no exaggerated bokeh.
No text, no readable signs, no logos, no brand marks, no watermark. Vertical 4:5 composition.
04

겨울 카페 창가 · 카드 쓰다 몰래 찍힌 순간

장소카페 창가 자리, 코코아와 딸기 케이크
시간·빛평일 오후, 창가 자연광 + 따뜻한 펜던트 조명
의상버건디 V넥 니트 카디건 + 흰 티 + 얇은 금목걸이
순간크리스마스 카드 쓰기에 몰두, 옅은 미소
구도맞은편 앉은 눈높이, 친구 머그가 앞에 흐리게

주의 · 시선이 아래라 얼굴이 가장 쉽게 바뀌는 컷입니다. 안 닮으면 보정 01부터 쓰세요.

LOCK — IDENTITY (must not change):
Use the person in the attached photo as the only identity reference. Keep her exact face: face shape and width, jawline, eye shape and eyelids, eyebrow shape, nose bridge and tip, lip shape, skin tone, any moles or marks, natural facial asymmetry, and apparent age. Keep her own hairstyle, hair length and hair color from the reference. Do not beautify, slim the face, enlarge the eyes, or turn her into a generic model — she must be instantly recognizable as the same real person.
The reference photo's background, clothing, camera angle, lighting and neutral expression are for identity only: do NOT copy them. Re-light her naturally for the scene and give her the expression and pose described below.
Only the main woman uses this identity. Any companions are different people with different faces.
SCENE: A quiet winter café in Korea, window seat on a weekday afternoon. Wooden table, large window with soft daylight, a ceramic mug of hot cocoa with light steam, a small plate with a half-eaten slice of strawberry cake, a few blank Christmas cards and a pen. The street outside is softly blurred; any signs outside are unreadable.
OUTFIT: Deep burgundy V-neck knit cardigan over a plain white T-shirt, thin gold necklace.
MOMENT: She is absorbed in writing a Christmas card, pen on paper, unaware of being photographed, smiling a little as she thinks of the person she is writing to.
EXPRESSION & GAZE: Eyes looking down at the card, soft small smile. Her head is only slightly lowered so her whole face — eyes, nose and mouth — is clearly visible in three-quarter view.
COMPANIONS: A friend sits across the table; only the friend's hand holding a mug enters the bottom-left foreground, out of focus.
LIGHT: Soft natural window light from the side mixed with warm café pendant light; gentle shadows.
CAMERA & FRAMING: Shot from across the table at seated eye level, slightly off-center; the friend's mug blurred in the foreground; the handwriting on the card is illegible.
FINISH — PHOTO LOOK:
Photorealistic candid photo, taken by a friend on a recent smartphone's main camera (about 26mm equivalent), handheld — not a professional shoot. Honest and unposed. Real skin texture with visible pores, fine baby hairs and slight natural unevenness; no beauty filter, no airbrushing, no heavy retouching, no glamorization.
Slightly imperfect amateur framing: subject a little off-center, slight tilt, edges may crop part of the scene. Smartphone image traits: mild noise in shadows, slight motion blur on moving hands or hair, white balance from the real light sources, natural colors with no cinematic color grading, no studio lighting, no exaggerated bokeh.
No text, no readable signs, no logos, no brand marks, no watermark. Vertical 4:5 composition.
05

백화점 대형 트리 · 트리 찍는 나를 찍어준 순간

장소백화점 아트리움의 대형 트리
시간·빛트리의 금빛 반사 + 차가운 천장 조명
의상브라운 스웨이드 하이넥 재킷 + 크림 비니 + 블랙 와이드 팬츠
순간폰으로 트리를 찍고 결과를 보며 감탄의 미소
구도옆에서 살짝 낮게, 트리가 위 배경, 인물은 왼쪽 아래

주의 · 위를 보는 각도 + 비니라 얼굴 단서가 가장 적습니다. 폰 화면·쇼핑백 글자도 확인하세요.

LOCK — IDENTITY (must not change):
Use the person in the attached photo as the only identity reference. Keep her exact face: face shape and width, jawline, eye shape and eyelids, eyebrow shape, nose bridge and tip, lip shape, skin tone, any moles or marks, natural facial asymmetry, and apparent age. Keep her own hairstyle, hair length and hair color from the reference. Do not beautify, slim the face, enlarge the eyes, or turn her into a generic model — she must be instantly recognizable as the same real person.
The reference photo's background, clothing, camera angle, lighting and neutral expression are for identity only: do NOT copy them. Re-light her naturally for the scene and give her the expression and pose described below.
Only the main woman uses this identity. Any companions are different people with different faces.
SCENE: The atrium of a large department store in Korea decorated for Christmas: a giant decorated Christmas tree, hanging ornaments, shoppers on the upper floors. No store names, no logos, no readable signage anywhere.
OUTFIT: Brown suede jacket with a high neck, cream knit beanie with her hair visible below it, black wide trousers, a plain paper shopping bag with no print.
MOMENT: She has stopped to take a photo of the giant tree with her phone held up at chest height and is caught smiling in delight while checking the shot.
EXPRESSION & GAZE: Looking up between the tree and her phone, lips parted in a delighted smile. Her face is turned about 45 degrees toward the camera so her features are clearly visible. The phone screen is dark or unreadable.
COMPANIONS: Her friend is shooting from the side; other shoppers blurred in the background.
LIGHT: Warm golden light from the tree reflecting on her face, mixed with cooler overhead mall lighting.
CAMERA & FRAMING: Taken from her side at a slightly low angle; the tree fills the upper background; she occupies the lower-left third of the frame.
FINISH — PHOTO LOOK:
Photorealistic candid photo, taken by a friend on a recent smartphone's main camera (about 26mm equivalent), handheld — not a professional shoot. Honest and unposed. Real skin texture with visible pores, fine baby hairs and slight natural unevenness; no beauty filter, no airbrushing, no heavy retouching, no glamorization.
Slightly imperfect amateur framing: subject a little off-center, slight tilt, edges may crop part of the scene. Smartphone image traits: mild noise in shadows, slight motion blur on moving hands or hair, white balance from the real light sources, natural colors with no cinematic color grading, no studio lighting, no exaggerated bokeh.
No text, no readable signs, no logos, no brand marks, no watermark. Vertical 4:5 composition.

마음에 드는 사진에 얼굴만 바꿔 넣기

새로 뽑을 때마다 구도가 달라지는 게 싫다면, 이미 만든 사진을 그대로 두고 얼굴만 바꿀 수 있습니다.

1

첫 번째 첨부: 바꾸고 싶은 완성 사진 (Image 1)

2

두 번째 첨부: 내 얼굴 사진 (Image 2)

3

아래 문장을 붙여넣고 [EXPR] 자리에 컷별 표정 문구를 넣어 보냅니다

첨부 순서가 바뀌면 반대로 바뀝니다. 꼭 완성 사진을 먼저, 얼굴 사진을 두 번째로 넣으세요.

Image 1: the finished photo to preserve. Image 2: identity reference only.
CHANGE ONLY: replace the woman's face and identity in Image 1 with the person from Image 2 — same face shape, eyes and eyelids, eyebrows, nose, lips, jawline, skin tone, moles and natural asymmetry as Image 2. Use Image 2's hairstyle and hair color, arranged naturally for the scene in Image 1 (wind, beanie, headband etc. stay as in Image 1). PRESERVE EXACTLY FROM IMAGE 1: pose, body, hands, outfit, the same kind of expression (keep the [EXPR]), gaze direction, head angle, companions, background, lighting direction and color, camera angle, framing, crop, grain and motion blur. Match the new face to Image 1's lighting and angle. Do NOT copy Image 2's gray background, black top, straight-on pose, neutral expression or flat studio light. Keep real skin texture; no beautification, no smoothing. No text, no logos, no watermark.
컷[EXPR] 자리에 넣을 표정 문구
CUT 01 크리스마스 마켓genuine open laugh
CUT 02 겨울바다 산책로laugh with mouth slightly open
CUT 03 연말 홈파티big natural laugh
CUT 04 겨울 카페 창가soft small smile
CUT 05 백화점 대형 트리delighted smile
CONTROL · 01–08

안 닮거나 어색할 때 보정 문장 8

새 대화를 열지 말고 같은 대화에 이어서 보내세요. 한 번에 한 가지만 고칩니다. [ ] 안은 상황에 맞게 바꿔 쓰세요.

01

얼굴이 안 닮을 때 (기준 사진 다시 첨부 후)

Edit the last image. Change only her face so it matches the attached reference photo more closely: same eye shape and eyelid type, eye spacing, nose bridge and tip width, lip shape and thickness, jawline and face width, and moles. Keep the same expression type, head angle, hair, outfit, lighting, background and framing exactly as in the last image.
02

피부가 플라스틱처럼 매끈할 때

Edit the last image. Change only the skin rendering: restore real skin texture — visible pores, slight redness around the nose and cheeks, natural under-eye texture, fine baby hairs. Remove any airbrushed, waxy or glowing look. Do not change any facial features, expression, lighting, outfit, background or framing.
03

장면·구도가 바뀌어 버렸을 때

Go back to the previous version. Change only [the one thing you wanted]. Keep everything else identical: face, identity, pose, hands, outfit, expression, background, companions, lighting, camera angle, framing and crop.
04

증명사진 배경·옷·무표정이 따라 나올 때

Edit the last image. Keep her identity, but remove anything copied from the reference photo: no gray studio background, no plain black top, no straight-on passport pose, no neutral passport expression. Restore the scene, outfit, candid pose and [laugh/smile] described in the original prompt.
05

친구 얼굴이 내 얼굴을 닮을 때

Edit the last image. Only the main woman has the reference identity. Make every companion a clearly different person, and keep companions' faces cropped out of frame or turned away. Change nothing else.
06

손·잔·폰이 깨질 때

Edit the last image. Fix only her hands: natural anatomy with five fingers on each hand, a relaxed realistic grip on the [cup/glass/phone/pen]. Keep everything else identical.
07

글자·로고가 생겼을 때

Edit the last image. Remove all letters, words, numbers, logos and brand marks from signs, cups, bags, screens and clothing; replace them with plain surfaces or blurred, unreadable signage. Change nothing else.
08

너무 화보처럼 깔끔할 때

Edit the last image. Make it look like a casual phone snapshot by a friend: slightly tilted and less perfectly centered framing, flatter phone lighting, a bit of shadow noise, no color grading, no studio glow. Keep her face, expression, outfit and the scene exactly the same.

나만의 장면으로 바꾸는 법

LOCK과 FINISH는 그대로 두고, 가운데 장면 일곱 줄만 바꾸면 어떤 장소든 같은 느낌으로 만들 수 있습니다.

줄무엇을 적나잘 되는 요령
SCENE장소·시기간판은 "unreadable"로. 실제 상호를 쓰면 글자가 생깁니다.
OUTFIT옷·소품로고·글자 없는 무지. 들고 있는 물건은 하나만.
MOMENT찍힌 순간"~하다가"로 쓰기. 포즈가 아니라 행동을 적습니다.
EXPRESSION & GAZE표정·시선"face fully visible"을 꼭 넣기. 얼굴이 가려지면 닮지 않습니다.
COMPANIONS일행친구는 손·어깨만. 얼굴이 나오면 내 얼굴을 닮기 쉽습니다.
LIGHT빛실제 그 장소에 있는 빛 두 가지를 섞어 적기.
CAMERA & FRAMING찍는 위치거리(m)·높이·기울기·어디서 잘리는지까지.

얼굴을 가장 잘 지키는 건 프롬프트보다 첨부 사진입니다. 아래를 보는 컷, 위를 보는 컷은 얼굴이 쉽게 바뀝니다. 얼굴이 카메라 쪽으로 45도 안쪽이면 안전합니다. 마음에 드는 장면이 나오면 새로 뽑지 말고 얼굴만 바꾸는 방법을 쓰세요.

자주 막히는 곳

어떤 AI에서 쓰나요?ChatGPT 이미지 기준으로 작성했습니다. 내 사진 1장을 첨부하고 컷별 프롬프트를 통째로 붙여넣으면 됩니다. 버전에 따라 결과가 다를 수 있습니다.
얼굴이 매번 조금씩 달라집니다정상입니다. ChatGPT 이미지는 생성할 때마다 얼굴이 조금씩 달라집니다. 2~3번 뽑아 가장 자연스러운 것을 고르고, 덜 닮으면 같은 대화에 보정 문장 01을 이어서 보내세요.
프롬프트가 왜 영어인가요?OpenAI 공식 이미지 가이드의 권장 표현을 그대로 가져왔고, 렌즈·플래시·흔들림 같은 사진 용어가 영어에서 더 짧고 정확하기 때문입니다. 복사해서 붙여넣기만 하면 되니 영어를 몰라도 됩니다.
어떤 사진을 첨부해야 하나요?정면이거나 15도 안쪽으로 살짝 돌린 얼굴, 밝고 고른 빛, 얼굴이 크고 선명한 사진이 좋습니다. 아래에서 찍은 셀카, 필터를 강하게 쓴 사진, 얼굴이 가려진 사진은 피하세요.