low-angle fashion editorial, blonde woman with long braided pigtails, crouching on cracked concrete, one knee grounded, other leg bent with foot forward, adjusting black rectangular sunglasses, bold matte red lipstick, vibrant yellow strapless crop top with black botanical line illustrations, matching yellow high-waisted shorts with black botanical patterns, oversized detached sleeves with graffiti graphics in teal orange and pink, multiple chunky gold chain necklaces, bright yellow platform combat boots with black laces and black rubber sole, black doodle illustrations on boot leather, towering graffiti-covered concrete walls, teal and orange stylized lettering, abstract bubble shapes, barbed wire strung between industrial buildings, vivid cyan sky with scattered cumulus clouds, harsh midday sunlight 5500K from directly above, hard shadows with sharp edges, saturated colors with high contrast, fish-eye lens 8mm distortion, barrel curvature at frame edges, urban street photography aesthetic, hyper-realistic skin texture with visible pores, glossy editorial finish, 8K detail --ar 2:3 --style raw
Click the prompt box above to select it, then press Ctrl+C (Cmd+C on Mac) to copy. Keep fish-eye lens 8mm distortion — it is the line that makes the image read as street photography rather than a lookbook shot.
Ask for "street fashion" and you will usually get a catalogue image: a person standing on an empty pavement, evenly lit, wearing clothes that look like clothing samples. The genre is not hard because it is complicated. It is hard because the interesting part of street photography is what happens between the clothes and the city, and a prompt that only names garments never asks for it.
This prompt works because it does four things that a generic fashion prompt skips: it fixes a specific camera and lens, it names a real light source with a real temperature, it describes the pose as an action rather than a stance, and it uses distortion deliberately instead of hoping for it.
The lens is doing the work, not the clothes
fish-eye lens 8mm distortion, barrel curvature at frame edges is the single most important line in the prompt. At 8mm the perspective is extreme: the near leg and the boots swell outward, the graffiti wall bends, and verticals curve. That curvature is the visual signature of street photography shot wide, and it is the thing that separates the image from a flat lookbook shot.
Writing it out explicitly matters, because the default is the opposite. Left to itself the model composes conservatively, around 50mm, which produces the compressed and tidy look you are trying to avoid. A guide on writing editorial-grade image prompts makes the general case: pairing a lighting description with a specific camera and lens reference is the biggest single jump from generic to photographic, because lens choice changes how the scene is compressed and how much of the background crowds in.
The cost of an 8mm look is distortion you cannot undo. If the face stretches at the edges of the frame, either move the camera description closer to the subject or switch to a wide but less extreme 24mm. Both are legitimate; what is not legitimate is asking for a fisheye and then complaining the result looks distorted.
Harsh midday light is a real lighting setup
The prompt specifies harsh midday sunlight 5500K from directly above, hard shadows with sharp edges. That is a real condition with real behaviour, and describing it accurately is what makes the shadows come out right.
Overhead sun casts short, hard-edged shadows directly beneath the subject and under the chin. It also lights the tops of surfaces and leaves vertical faces comparatively dark. In this crouching pose the effect is useful: the yellow fabric and the boot tops catch the light while the face sits in partial shade, which gives the image depth without any studio equipment.
Two related details worth copying. The temperature figure matters — daylight around 5500K reads as neutral, and naming it keeps the image from drifting orange. And "hard shadows with sharp edges" is doing more work than it looks: shadow softness is a direct function of source size, so describing the shadows tells the model what kind of light it is modelling, without you having to explain the physics.
That guide is worth reading for the underlying principle: it recommends naming a recognised photographic setup — hard monolight, north-facing window, golden hour backlight — and specifying angle, size and hardness, because generic words like "good lighting" leave the model with no decision to make and it defaults to something flat.
Describe how the fabric behaves, not what it is made of
Note the garment descriptions in this prompt: oversized detached sleeves, chunky gold chain necklaces, bright yellow platform combat boots with black rubber sole. These are physical descriptions rather than fabric names, and that is the standard to aim for.
A fashion prompt guide written from a photographer's point of view makes the argument directly: name the physics, not the fibre. Silk falls. Wool crepe holds. Taffeta stands away from the body. Jersey clings and pools. Its example is that "structured wool holding its shape" beats "wool coat" every time, because the first tells the model what to draw and the second only tells it a category.
The same guide suggests naming the silhouette first, before any detail, because the shape the garment cuts against the background is what the eye reads before it reads texture. That is why the crouch is described so precisely here — one knee grounded, the other leg bent with the foot forward. A crouch is a silhouette decision.
Street specifics are what make it read as street
The background is doing real work: towering graffiti-covered concrete walls, barbed wire strung between industrial buildings, vivid cyan sky with scattered cumulus clouds. Generic "urban background" produces a grey blur. Naming the vertical scale and the sky is what tells the eye the model is looking up from ground level.
For variations, several fashion prompt collections recommend keeping the light, the lens, the framing and the colour grade fixed and changing only the outfit and the pose. That is worth doing: it is what produces a series that reads as one shoot rather than four unrelated images.
One honest warning about the aspect ratio. This prompt asks for --ar 2:3, and at that ratio a full-body vertical crop with 8mm distortion forces a real trade-off: you either lose the boots or lose the head. A product-photography prompt guide recommends 4:5 for editorial work and reserves 9:16 for vertical social crops. 2:3 sits between them and suits a full figure better than a square would, but if the framing keeps failing, moving to 4:5 is usually the cheaper fix than fighting the composition.
Fix the parameter that fights your instruction
The prompt ends with --style raw and no stylize value, which means stylize stays at its default of 100. That is deliberate for this kind of image. Midjourney's stylize documentation describes the parameter as how much of the model's own aesthetic is applied over your prompt, on a scale from 0 to 1000, where low values stick to the facts of the prompt. With a prompt this specific, that is the behaviour you want.
When you are ready to iterate, lock the seed. Once a composition works, fixing the --seed and changing a single phrase is the only way to know that a change in the image came from that phrase.
Where this stops working
Two limits worth stating plainly. First, 8mm at this framing can distort the face badly if the subject drifts toward the edge, and the model is not reliable about keeping a face out of the distortion zone. If that happens, ask for the subject more centrally rather than adding more lens language.
Second, the graffiti is generated, not photographed. The lettering will be convincing as texture and unreadable as words. That is fine for an editorial image and would be a problem if the design depended on the text being legible.
The useful framing is that this prompt is not asking for "a fashion model". It is describing a specific camera, a specific light and a specific ground-level viewpoint, and letting the clothes follow from them.
Label: Cinematic Fashion Poster
Key Principle: Name the camera and lens before the clothes. An 8mm fisheye with overhead hard light and a ground-level viewpoint will make the garment read correctly on its own — the outfit is the last thing you should be describing.
Comments
Post a Comment