When creating a prompt I love to have an idea of an image, a moment I'd like to capture (worthy of story), this is key to creating a great AI image. Then breaking it down further, if motion is involved, or if the image is still. If the image is a close up, that changes the dynamic of the image completely. Especially when focusing on details of a shrunken woman, at this point the surrounding elements must be accounted for to get a realistic perspective feel. So having a general idea of what type of image you'd like to capture helps tremendously when creating AI images.
Once an idea is in place, next comes the technicals of creating the prompt. Which I will list in the 5 main steps I use to create AI images.
Choosing the style of shot — down to the lens, camera model, and rendering method.
[1. Lenses: Some of my go-to lenses]
35mm Macro – This one is my go to for small-scale realism. Think: extreme clarity, shallow depth of field, subject in hyper-focus with gorgeously blurred backgrounds.
50mm Cinematic – A classic portrait lens. When I want softness, intimacy, or that “A24 indie drama” energy, I default here.
85mm Telephoto – A dreamy option when I want strong background compression. Great for isolating the shrunken figure against a “giant” object or character.
Technical Add-ons:
Add camera bodies like Canon EOS R5, Nikon Z9, or Sony Alpha 7R IV. The AI often infers sensor quality, light behavior, and even color science from these.
Use tags like DSLR photo, photorealistic, ultrarealistic, or 8K resolution to elevate rendering quality.
Prompt example: "Macro 35MM Canon EOS R5 of a tennis ball."
[2. Camera Angles]
Camera angle is where you start playing chess with perspective — especially when scale is involved. This is where it gets tricky and I've obtained many of my images combining elements here.
My Go-To Angles:
Worm’s Eye View – Makes everything above the subject feel massive. Perfect for capturing a shrunken woman dwarfed by a giantess or towering furniture.
Bird’s Eye View / Overhead Shot – Adds clarity to scene composition. Great for showing how lost or small a character is within their environment.
Dutch Angle (Tilted Frame) – Instant tension. I use this when the scene feels chaotic or emotionally intense.
Extreme Close-Up – Like, see-the-fuzz-on-the-tennis-ball close. Great for emphasizing texture and scale simultaneously.
Wide Angle / Fisheye – Distorts the world to exaggerate perspective. I use this sparingly, but when I do, it delivers.
Side Profile – Clean, emotionally grounded. Good for character-focused shots.
Over-the-Shoulder – Adds narrative tension or emotional POV, especially if a second character (like a looming giantess) is involved.
POV Shot / GoPro Selfie – For humor or immersive storytelling. Think: vlog-style, but tiny.
How I Layer These:
Say I’m creating a shot where my six-inch-tall character is facing off with a tennis ball.
I might use:
“35mm Macro lens, extreme close-up from worm’s eye view, focus on coarse yellow tennis ball fibers, Canon EOS R5 clarity, cinematic lighting, shallow depth of field. Low angled shot of a 30 year old brunette woman shrunken to 6 inches tall (true to scale) grasping the tennis ball desperately, her face pressed against the tennis ball, the grip from her hands causing the tennis ball fibers to sink in.”
[3. Background & World-Building]
Once the character is imagined, I zoom out — literally and creatively.
I try to incorporate real-world objects as visual anchors to contrast size and guide the eye.
Scaling Tools I Use:
Common objects: AA batteries, smartphones, coins, lip gloss tubes, forks — instantly recognizable.
Interior spaces: Modern kitchens, tiled floors, sleek bathrooms, and sunlit bedrooms. I lean toward monochromatic, light-toned spaces because they reflect ambient light better and create more photorealistic results.
Surface detail: Don’t forget materials. Prompt for “polished marble countertops,” “porous ceramic tiles,” “scratched plastic texture,” etc. This makes scale believable.
Environmental Enhancers:
“Subsurface scattering” for realism in skin tones.
“Ray-traced reflections” on glass or tile.
“Volumetric lighting” for sunbeams through windows.
“HDRI lighting map” to simulate a real-world photo studio or daylight environment.
[4. Character Design & Prompt Consistency – Be Obsessively Specific]
This is where you cannot be lazy. Saying “brunette woman” in your prompt almost guarantees bland results from AI image generators.
If you want character consistency, you have to describe every relevant detail:
age, build, hairstyle, clothing textures, vibe, emotion, expression, even posture.
What I Actually Write:
“23-year-old curly brunette woman, light olive skin tone, soft glam makeup with winged eyeliner and peach-toned blush, silky brunette hair in a slightly messy bun, wearing a low-cut (busty), worn-in grey tank top with faded pink graphics that say ‘Shrink Me’ and a cherry logo, subtle muscle tone visible in shoulders, standing confidently, her toned legs glistening in the sun.” - something like this.
The more you feed the AI, the more your character feels real — and can be recalled consistently across multiple prompts or frames.
[5. Lighting – Realism Lives Here]
I almost always prompt for white diffuse ambient light or soft natural sunlight. Sometimes I will blend the two elements, white ambient light blends with natural sunshine rays from partially open blinds (window is off screen). I tend to use white ambient light as it livens up the scene and realism.
My Lighting Favorites:
Soft window lighting – Makes skin glow, adds realism to shadows.
Golden hour lighting – Dreamy warmth, cinematic effect.
Ray-traced shadows – Natural falloff, helps with ground contact shadows (super important for realism).
Volumetric light beams – Good for fantasy or “sunbeams cutting through dust” vibes.
Backlighting or Rim Lighting – Adds glow and dimensionality.
Prompt example:
“white ambient daylight from large window behind camera, soft shadows on floor tiles, subtle backlight highlights edges of 30 year old brunette's silhouette, Canon-style dynamic range.”
Avoid harsh LED or fluorescent lighting unless you're going stylized or surreal. Harsh colored lights tend to flatten the image and reduce believability.
Bonus: Visual Storytelling Through Composition
You can use all these techniques to tell a bigger story within a single frame. Here are some ideas:
Foreground Focus, Background Story: A detailed close-up of your shrunken woman reaching for a piece of popcorn, with a blurred giantess watching in the background. It’s a story in a frame.
Environmental Tension: A low-angle, side-lit shot of the character hiding behind a dropped pencil, with giant footsteps incoming — light and shadows doing most of the storytelling.
Dynamic Triangles: Use the “rule of thirds” or triangular composition to guide the viewer’s eye through the frame — from the small subject, to the oversized object, to the emotional cue (like an expression or interaction).
[Final Thoughts]
Great prompting is thinking like a cinematographer, set designer, and storyteller all at once.
When you know what shot you want, you can engineer it into existence. Break down the scene into:
- Lens and style
- Camera angle
- Background environment
- Character design
- Lighting
===========================================
Below is an example prompt I used in Maggie and Jane (25th Anniversary Remaster) - Please feel free to analyze and use any elements!
"A photorealistic cinematic still, captured with a 50mm full-frame lens, blended with macro focal depth for shallow focus and atmospheric intimacy. The scene is set in a dimly lit, cluttered college dorm room, rendered in moody amber-orange tones. Lighting is soft and dramatic — coming from a flickering desk lamp positioned in the far background, casting gentle rim light and creating deep shadows across the space. The ambient glow spreads unevenly, giving the room a slightly eerie but cozy tension.
In the foreground, kneeling on the hardwood floor, is a cute half-Asian woman in her early 20s, with short, slightly tousled black hair, and a playful, fascinated expression. She wears a loose black computer-graphic T-shirt and black gym shorts, sitting comfortably on her heels with her bare knees pressing into a sweatshirt on the floor. Her bare feet are visible, casually posed behind her.
She's holding a steaming cup of ramen very close to her face with both hands. The cup is slightly worn, printed with classic red-and-white graphics, and a white plastic fork is seen inside, partly submerged in broth. Steam rises from the cup in visible swirls, fogging her round glasses slightly and catching the backlight from the desk lamp in ethereal tendrils.
Her face is partially illuminated, eyes squinting in delight and mischief as she leans in and speaks directly into the cup, lips only inches away. Her expression is a blend of playful curiosity and surreal fascination, as though she’s expecting the ramen to respond. The steam and fork reflections play across her cheeks and nose, enhancing the mystical undertone of the moment.
The floor around her is messy and realistic — scattered items like crumpled chip bags, a half-empty water bottle, a phone charger, sticky notes, and a pair of earbuds create natural texture. A tangled bedsheet and an open laptop glowing faintly can be seen blurred in the deep background."
RESULT
