Have you ever experienced a story in your head that's very graphic, but that opens the AI biographic tool and holds up for half a day to write "Sabbonke, Rain Night, Man, Neon Light, 8k"
As a result, men may be standing on neon lights, and the rain is nowhere to be found。
Today's article is the time to give you a complete farewell "keyword draw."。
I'll make it clear in the shortest possible time Gemini subscopePrompt wordIt's a core methodology, and then it's just a setCopyable paste, immediately usableThe "Night Street Man's Nine Palace Spectrum" tip。
Take it away, thanks。
Let's get one thing straight: Gemini doesn't eat keywords, he eats stories
OLD GENERATION AI DRAWINGS, ESSENTIALLY LABEL MATCHING. EVERY COMMA YOU ENTER SEPARATES, THROWS AN ELEMENT INTO THE POOL, AND THE MODEL COLLAGES。
Gemini is different。
It's a big language modelI can read sentencesI don't know. It knows that there is a difference between "rainwater creates blurred lines in the outlook" and "rainwater in the background". It understands the principal and distinguishes causality。
so,Don't write any more labels, a word。
Bad writing:
saberpunk, rain night, neon lights, man, run, 8k, fantasy engine rendering
Good writing:
On a Sabonk-style rainy night, a man in a black windsuit ran on a wet, slippery street with neon lights. Rainwater creates a blurry line in the outlook against the background of a steel building that is high in the clouds。
The second approach looks long, but models can accurately allocate the layers of each element -- who's in the outlook, who's in the background, who's moving, who's still。
This is the first principle of the Gemini-era hint:Replace the label with narratives。
It's not too bad when four things are right
It's called "Four Pillars" in the industry
1. Who are you (role-setting)
Tell Gemini it's identity. And don't underestimate this step, "You're a Hollywood spectrographer with 20 years of experience, so you can turn the output from "Ai" to "the sense of film."。
2. Doing what (specific tasks)
"Generating a map" is too vague. To write: "A 16:9 film-level spectroscopy based on the following plays, using the medium view."
3. Background (context)
This is "Why." To add, "It's a late night after the hero's first major failure, the model automatically darkens the light and turns cold。
4. Delivery format (physical specifications)
Draw scale, need text, need specific layout. Make it clear that there is no room for models。
An all-powerful formula for each spectroscopy
Compress the theory above into a line of formulas, and each time a hint is made, it's all right:
Subject + Action + Environment + Camera Language + Style atmosphere
An example:
A 25-year-old young Asian male (who is the main body) is a young Asian male, who is a young Asian male
I'm standing in front of a window in a convenience store, staring in the window (motion)
In the middle of the night in the streets, the ground reflects the warm colour of fluorescent lamps (environment)
In the middle view, the body is brightened by warm light and the body is engulfed by a cold, blue shadow (photo language)
Film-level Teal-and-Orange Tuning, Kodak Vision 3,500T granular sense (style atmosphere)。
Take note of the details:
- The subject needs an imageI don't know. It's not a man, it's an Asian youth of 25 years, short hair, dark hoodie, old rags。
- You have to use a strong verbI don't know. Not “in the street”, but “crashing” or “walking with your hands in your head”。
- Emotions don't write abstractI don't know. Sadness, “No, write”, “tear down on the shoulders, staring at the ground with undried tears on the face”。
The model does not paint "great" and "justice", but it can draw a "silent shadow that stretches the shadow under a street light."。
Use camera language to make the picture "directive."
THIS IS THE DIFFERENCE BETWEEN "AI" AND "A"AI MIRROR"The watershed。
The same scene, a different scene, different emotions。
Gemini has learned a lot of photographic metadata in training, so you just have to talk like a director, and it's gonna work。
Don't check the scenery
|
|
|
|
|---|---|---|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Angle speed check
|
|
|
|---|---|
|
|
|
|
|
|
|
|
|
|
|
|
Light speed check
|
|
|
|---|---|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Remember:Do not write "a nice light" and "Lembrandt light, flame of a lighter on the left side to form a warm orange triangle"。
THE MORE SPECIFIC, THE MORE HONEST AI IS。
Role Consistency: The ultimate pain of the lens
It's easy to make a picture, to make nine one-size-fits-all, one-man -- that's really hard。
It's just a description, even if you don't change a word, and you'll get another action model。
Solution CallThe anchor and the axis“, three steps:
Step 1: Generate role baseline maps
Use this hint below to generate repeatedly until a satisfactory "base face". Here's your whole set of nine-mong identity cards。
A photo of a young Asian man about 25 years of age in front of him。
Short hairs, wet by rain, thin body, lined on the bottom, and skin colour warms。
Dressed in a dark gray cotton hoodie (hat down), black pants
An old olive green canvas bag with a brass strap on it。
The expression is quiet fatigue and interior, looking directly at the lens。
photographer: 85mm human image focus lens, f/2.0 aperture
A soft studio light with a little cold, cyanide aid light
Clean, dark, grey background, shallow view
The facial details are very clear - skin texture, mammoths, eyebrows。
Film class sense, Kodak Vision 3,500 T-film particle sense。
1:1 draw scale。
- Key: This picture has only one purpose — to kill the face. So the background is clean, the light is even, the face is flat and the face is neutral. Don't add any fancy scenes, the more plain and stable。
Step 2: Upload as a reference map
Upload this satisfactory image to Gemini, tick "for reference." Gemini is an original multi-modular model that encodes the character of this map into mathematical restraints -- every frame that comes after it is "locked" to the face。
Step 3: The pivotal shift
The next hint is only to describe new scenes and new moves, and the role identity is locked to death。
Also remember to add an insurance policy to each frame of hint:
- “Directive red line: absolute maintenance of facial anatomy structure, hairstyle and inherently colour-based characteristics.”
If you create dozens of frames, your face starts to fall -- don't push. Change backs, shoulder lenses, panorama, shadows. The use of film language to hide the technology is sometimes more advanced。
War: Night Street Man's Nine Miyagi Mirror
The theory is over. The following is a complete reminder that can be directly reproduced。
Theme Settings
"ROVER / NIGHT WALLKER"
A 25-year-old Asian young man walks alone late in the night in the streets after the rain of neon. Emotional arc lines are drawn from confusion and depression, and by dilation, they eventually lead to light。
The Nine Palace Grid
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
♪ Bang, bang, bang, bang, bang ♪
The city skyline, the neon sign, the rain, the shadows
♪ High-end vision ♪
I'm sorry
♪ Right, right, right ♪
Smiling, smug, smug, smug, smug, smug, smug
♪ In the middle, in the near, in the low, in the middle, in the middle ♪
I'm sorry
♪ Bang, bang, bang, bang ♪
♪ The sky bridge ♪
♪ High angles ♪
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
Emotional arc design
Frame 1, city landscape, cold, vast
frame two, neon close-up
♪ Frame three, rain alley shadows, lonely, lost ♪
Frame four, side face, smoke, tumble, think, struggle
Frame five, water rewinding
In front of the convenience store, in short, warm, warm, emotional turn
The 7th frame, the bridge
The 8th frame smiles, smiles, comforts, embraces, the core turning points
Frame 9, street light cut
Description of use of role reference maps
Before you create each of the following frames, you have to do two things:
- Generates a positive picture of a satisfactory character using the role benchmark hint above
- Upload this image to Gemini dialogue, tick "For information"
Upon completion, each subsequent frame of hint begins with a line of hints [uploading of role baseline maps, ticking out” as reference — meaning that a reference map has been uploaded in your dialogue. If you're generated in the same dialogue window, you just have to upload it once。
Uniform style suffix (add each frame)
This is the “one-size-fits-all fuse” that is attached at the end of each generation。
Visual style: New black style at film level, with Sabon Neon aesthetics。
Simulation of Kodak Vision 3,500T tungsten tungsten to balance film films with a clear sense of organic particles。
Teal-and-Orange complementarities, dark black pressure low
The middle is saturated but neon focuses selectively penetrator。
It's a shallow, cream-like, hexagonal。
Rainwater magnifies the reflection of all surfaces。
It's a melancholy, atmospheric, deep urban, emotional resonance. 16:9 Paint scale。
Nine Frames of Independence
Each of the following frames can be used independently and will be created and manually spelled into nine palaces。
Frame 1: Urban skyline (extreme vision — building a world view)
a very long-term view of the city shows a big asian urban skyline in the night after the rain. numerous skyscrapers pierced low clouds, and windows gave small amber colours and blue light. the wet roof reflected a fragmentary neon light, creating electric blue and red strips. there's no one on the scene. using 16mm wide-angle shot, deep view, capture endless city details。
[ Paste style suffixing ]
- This frame doesn't show up. Its role is to tell the audience that the story happens in a big, cold, bright city。
Frame 2: Neon
An old neon sign is very close, and the Chinese characters on it are inverted. Raindrops slowly slide along the surface of hot glass lamps. The light depths the background as a pure abstract swath — red, electric blue and warm amber. Photographed using micro-species, with a strong focus on the broken glass sense and warm tungsten light permeated through the beads。
[ Paste style suffixing ]
- Pure air frame. Draws the physical quality of neon lamps to the extreme with a micro-species lens, creating a colour base for the entire spectroscope: Magenta, electric blue, warm amber。
Frame 3: The Back of Rain Lane (Prospect & Lonely Visual Definition)
[upload role baseline, ticking' as reference]
The profile of the face, hair and colour are kept in strict conformity with the uploading role reference。
The vision, in the shadow of the role, he walked alone in a narrow late-night alley after the rain. On both sides of the alley are high-rised buildings, full of entangled pink and blue neon signs. The wet placard road maps the long strips of neon above its head. He pulls up his hoodie hat, with a slight tweak and closes his shoulders inside. The diagram uses a symmetrical perspective and places him in the centre of the disappearance point of the alley。
[ Paste style suffixing ]
- Classic spectroscopy: a little perspective + person center + back. Solitude does not require language, and the drawings themselves tell stories。
Frame 4: Smoke on the side
[upload role baseline, ticking' as reference]
The profile of the face, hair and colour are kept in strict conformity with the uploading role reference。
Mid-near view, on the left side of the game when you smoke. The flames of the lighter projected a sharp Rembrandt-type warm orange triangle on his cheeks and noses, while the other side of the face fell into an impermeable dark shadow. A line of smoke is curled up and captured by cold, blue light under the invisible neon sign above. His eyes were closed, his eyes were low, and he was clearly exhausted. Side 45 angles, classic contrast light。
[ Paste style suffixing ]
- this frame is the headline of the light shadow design. the lighter's hot oranges vs neon blues split up in the same face -- the cool conflict is the externalization of the inner conflict。
Frame 5: Rewinding water (very low angle . self-fragmentation)
[upload role baseline, ticking' as reference]
The profile of the face, hair and colour are kept in strict conformity with the uploading role reference。
Very low angles were taken, cameras were almost attached to the moist ground and they were shot up. The outlook is dominated by a shallow water-filled lid, which is like a mirror, reflecting a confusing reflection of the neon sign and the architectural surface — a reflection of the red, electro-blue and gold woven into a vivid and distorted watercolor. The role's dark leather boots and black pants were standing on the far edge of the water, and his reflection was distorted by a soft tatter。
[ Paste style suffixing ]
- Put the camera on the ground and use it as a mirror. The reflection is distorted by waterlines — the most intuitive visual metaphor of self-lost. No face, just a pair of shoes and a broken reflection。
Frame 6: Facilitator shop stay
[upload role baseline, ticking' as reference]
The profile of the face, hair and colour are kept in strict conformity with the uploading role reference。
In the middle view, the role was silently standing in the streets of the night in front of a light-lighted Japanese convenience store. Warm fluorescent light is leaked from within the glass curtain wall, spilled on the wet sidewalk, forming a soft amber-colored, warm-backed rectangle on the ground. He stood on the borderline of the light — half of him warmed by cozy tungsten and half of him swallowed by the cold blue and blue shadows of the streets. He looked inside the window and looked in his eyes with a quiet desire。
[ Paste style suffixing ]
- The warmest one in the nine frames. The light at the convenience store is the only warm-faced source of light in the entire cold city. Man stands on the line of warm and cold — trying to get in, not。
Frame 7: Skybridge Subduction
[upload role baseline, ticking' as reference]
The profile of the face, hair and colour are kept in strict conformity with the uploading role reference。
High angles are projected and the camera looks down vertically from above the footbridge. The role is that of a tiny solitude beneath the distance, and is passing alone through a wide, open and rainy crossroads. White dotted lines of road signs and zebra lines form a distinct geometric pattern on the dark, humid asphalt surface. His tiny shadow was pulled long under the only sodium steam lamp。
[ Paste style suffixing ]
- The doppelgangers seem small. Geometric roadlines + open spaces + a little human shadow = a sense of weakness crushed by a huge city machine。
Frame 8: Smiling back
[upload role baseline, ticking' as reference]
The profile of the face, hair and colour are kept in strict conformity with the uploading role reference。
intimate facial features, the role turns around in the middle of the night, looking directly at the audience across the shoulder. a small, but sincere, tired and peaceful smile now floats into his mouth. his dark eyes reflected a tiny, warm amber light. a drop of rain drew a trace along his temple. the background is rendered as a cream-like dreamy neon vortex vortex - pink, gold and blue soft and light balls. use 85mm human image focus lens, f/1.4 large arcade, to achieve maximum body separation。
[ Paste style suffixing ]
- It's a set of nine frames of soul。 the first seven frames put pressure on the mood for so long, and here they are released with a "tired but calm smile". 85mm f/1.4 turning the background into a mirage, all focus on that smile。
Frame 9: Roadlight Crunch
[upload role baseline, ticking' as reference]
The profile of the face, hair and colour are kept in strict conformity with the uploading role reference。
Vision, role back-to-view viewers go a long way along a straight and empty path of growing up. On both sides of the road are high road lamps arranged at equal ranges, each of which produces a warm sodium orange cone light column down. He is in the middle of the picture and is being presented as a clean, dark-coloured clipping, which is lined to the golden reverse light of the roadlights from the farthest. The wet surface brings back a rhythmic pattern that extends to unlimited distances. A layer of mist is low and low, which softens the distant light into the radiance of the emptiness。
[ Paste style suffixing ]
- The last frame doesn't need to see the face. A cut, an endless road, a repeat light — he continues to walk with his own loneliness and relief. The story does not end, but the picture itself is the best ending。
One-time generation of a complete message of the whole nine-gauge
If the model is used to support the generation of a one-time nine-gauge collage (rather than a frame-by-frame) you can use the complete hint below. This is a whole phrase. Don't open it:
A carefully crafted film-level collage containing nine separate square photographs organized in a 3x3 grid in a dark carbon gray, fine texture background, with each photograph separated by a dark-coloured border with a small gleaming light. The image depicts a young Asian male walking alone on the streets of the late-night city of Neon after the rain, in an all-embracing atmosphere, depression and strong film-class new black aesthetics。
the upper left map is a very long-term view of the city, showing a vast urban skyline in asia on the night after the rain. numerous skyscrapers pierced low clouds, and windows gave small amber colours and blue light. the wet roof reflected a fragmentary neon light, creating electric blue and red strips. there's no one on the scene. using 16mm wide-angle shot, deep view, capture endless city details。
The graph above is a very close-up of an old Neon sign, and the Chinese characters are insinuated. Raindrops slowly slide along the surface of hot glass lamps. The light depths the background as a pure abstract swath — red, electric blue and warm amber. Photographed using micro-species, with a strong focus on the broken glass sense and warm tungsten light permeated through the beads。
The top right picture shows the back of a young Asian man who walked alone in a narrow alley after the rain. On both sides of the alley are high-rised buildings, full of entangled pink and blue neon signs. The moist surface of the tablets is like a mirror, and the long strips of neon lamps above their heads are scintillated. He was wearing a dark coal gray hoodie, a hat pulled up, black straight pants and an old canvas bag slashed. Whisper in the shadows and close in on your shoulders. The diagram uses a symmetrical perspective and places him in the centre of the disappearance point of the alley。
The mid-left picture is a medium-nose view on the left side of a young man's smoke. The flames of the lighter projected a sharp Rembrandt-type warm orange triangle on his cheeks and noses, while the other side of the face fell into an impermeable dark shadow. A line of smoke is curled up and captured by cold, blue light under the invisible neon sign above. His eyes were closed, his eyes were low, and he was clearly exhausted. The luminous contrast of the Book。
The medium graph is taken from a very low angle and the camera is almost attached to the wet ground. The outlook is dominated by a shallow water-filled lid, which is like a mirror, reflecting a confusing reflection of the neon sign and the architectural surface — a reflection of the red, electro-blue and gold woven into a vivid and distorted watercolor. The young man's dark-skinned boots and black pants were standing on the far edge of the water, and his reflection was distorted by a soft tatter。
The middle right is a medium view, with young men standing still in front of a bright Japanese convenience store. Warm fluorescent light is leaking through the glass curtain wall, spilling over the wet sidewalk and forming a soft amber-coloured rectangle on the ground. He stood on the borderline of this light — half of him warmed by comfortable tungsten and half of him swallowed up by the cold blue and blue shadows of the streets — and he looked inside the window, with a quiet desire in his eyes。
The lower left map is taken from a high angle and the camera is seen from the vertically down above the manwalk. The young man is a small solitude beneath the distance and is passing alone through a wide open and rain-maped crossroads. White dotted lines of road signs and zebra lines form a distinct geometric pattern on the dark, humid asphalt surface. His tiny shadow was pulled long under the only sodium steam lamp, creating a deep sense of isolation and vulnerability。
the middle graph is a close facial feature, with young men turning their heads and looking directly at the audience across their shoulders. a small, but sincere, tired and peaceful smile now floats into his mouth. his dark eyes reflected a tiny, warm amber light. a drop of rain drew a trace along his temple. the background is rendered as a cream-like dreamy neon vortex vortex - pink, gold and blue soft and light balls. use 85mm human image focus lens, f/1.4 large arcade, to achieve maximum body separation。
The bottom right picture is a vision, with young men walking far behind the viewers along a straight and empty path of growth. On both sides of the road are high road lamps arranged at equal ranges, each of which produces a warm sodium orange cone light column down. He is in the middle of the picture and is being presented as a clean, dark-coloured clipping, which is lined to the golden reverse light of the roadlights from the farthest. The wet surface brings back a rhythmic pattern that extends to unlimited distances. A layer of mist is low and low, which softens the distant light into the radiance of the emptiness。
The overall picture reflects a real film-class street photography style: depression, atmosphere, deep city feeling, emotional resonance, and new black aesthetics. Each photograph maintains an accurate 1:1 scale of the painting, and the overall collage is roughly square. Mainly urban artificial night lighting: cold electric blue, red and cyanid neon flashes, interwoven with warm sodium orange street lights and fluorescent light at convenience stores. A close-up image uses a clear and dark comparison with Rembrandt light. Simulates the apparent organic sense of Kodak Vision 3,500T film. A shallow-to-medium-deep, with a six-sided, cream-like dispersion. Teal-and-Orange Complementary Colour Modified, Dark Black Underpressed, Saturated in the Middle, and Neon's Focus Color Selective Passage。
- Remember to upload the reference
The full flow of the nine-legged lens
Finally, a summary of the complete steps from zero to finished products:
Step 1: Role-setting
The role anchor text is written, and a satisfactory "base face" is generated repeatedly。
Step two: style-setting
Fix the style suffix, lock the color system and the texture。
Step three: Dismantling the scene
Split the story by emotional arc into nine visual beats, one emotion, one action。
Step 4: Matching the camera
Each scene is allocated (distant/medium/near/special) + angle (horizontal/top/top) + light type。
Step five: frame by frame generation
Puts the role anchor + scenario description + style suffix together and creates a frame. Each frame is confirmed and the next frame is generated。
Step 6: Collapse into a map
Spelling with any puzzle tool (Canva, Figma, or even PPT) into 3x3 of the Nine Palaces, with title and signature。
Or, if your model supports, you can do it in one step, with a complete hint。
The direction of the variant that can be replaced directly
In the same framework, the new theme is a new set of nine palaces:
|
|
|
|
|
|---|---|---|---|
| The Rain Day Café |
|
|
|
| End of the night |
|
|
|
| The morning roof |
|
|
|
| Walking down the street |
|
|
|
Final Thoughts
THE ESSENCE OF THE AI SPECTROSCOPE IS NOT TO LET THE MODEL PAINT YOU, BUT TO USE THE TEXT AS A GUIDE, TO MOVE THE FILM IN YOUR HEAD WITH A HINT。
The model is just your photographer. You're the director。
And a good director always knows what he wants。