A white grand piano stretches from left to right across the frame. Focus falls on a few keys in the middle—a team of tiny people in blue overalls are scrubbing the key surfaces with brushes and buckets. Keys in the foreground and background melt into creamy blur.
This is the tilt-shift miniature photography effect: making real-sized objects look like toy models.
If you've never written a prompt like this before, don't worry. This tutorial starts with three fundamental concepts, walks you through building the complete prompt step by step, and ends with the 5 mistakes beginners make most often.
Final Result Preview
This is what we're building toward. Notice three things:
- Blurry edges, sharp center — the signature of tilt-shift
- Crisp figure details — you can see the wrinkles in their overalls and the metallic sheen on the buckets
- Ivory-like key texture — not flat black-and-white rectangles
By the end of this guide, you'll know exactly which words control each of these effects.
3 Core Concepts You Need First
Before writing any prompt, understand these three ideas. No photography background needed—30 seconds each.
Concept 1: Tilt-Shift Effect
Tilt-shift is a lens technique that tilts the lens plane so only a narrow horizontal band is in focus while everything above and below goes blurry. This blur pattern tricks your brain into thinking you're looking at a tiny model—because when you look at small objects up close with your eyes, the depth of field is naturally this shallow.
In a prompt, tilt-shift photo (four words) activates this effect.
Concept 2: Shallow Depth of Field
Depth of field is the range of distance that appears sharp in an image. The shallower it is, the narrower the sharp zone and the stronger the blur.
shallow depth of field→ very narrow sharp band, strong blur (maximum miniature feel)moderate depth of field→ medium, more context visibledeep depth of field→ almost everything sharp (miniature feel disappears)
For tilt-shift miniature, always use shallow.
Concept 3: Tiny Human Figures
The "soul" of a miniature scene isn't the scene itself—it's the tiny people in it. Without them, a miniature scene is just a close-up photo. The figures provide scale reference: the keys look massive because the people look small.
In your prompt, you need to:
- Establish their existence:
tiny human figures - Specify their action:
cleaning the piano keys with brushes and buckets - Keep it simple—3-5 figures is plenty
Step 1 → Step 2 → Step 3: Building the Full Prompt
Step 1: Set the Composition and Viewpoint
Decide where your "camera" goes. For a piano scene, the most effective angle is a side view from above—looking along the keys from the bass end to the treble end.
As prompt text:
side view from the left looking along the keys to the right
Why not a front view? Because you'd only see the edges of a few keys, missing the low-to-high sweep that makes tilt-shift so effective.
Step 2: Layer in Tilt-Shift and Depth of Field
Add the technical parameters on top of composition:
Tilt-shift photo of a grand piano keyboard, side view from the
left looking along the keys to the right, shallow depth of field
At this point, you have a tilt-shift piano close-up. But the "tiny world" feeling is still missing.
Step 3: Add Figures and Atmosphere
Complete the scene with character actions, mood, and quality flags:
Tilt-shift photo of a grand piano keyboard, side view from the
left looking along the keys to the right, shallow depth of field,
tiny human figures cleaning the piano keys with brushes, cloths,
and buckets, whimsical and surreal scene, soft lighting,
hyper-detailed, high realism.
Done. That's the complete prompt. Three layers, each building on the last.
Here's what each new addition controls:
| Addition | Purpose |
|---|---|
tiny human figures |
Creates miniature people for scale contrast |
cleaning with brushes, cloths, buckets |
Specifies character actions and props |
whimsical and surreal |
Sets the mood: playful + surreal |
soft lighting |
Gentle illumination, avoids harsh shadows |
hyper-detailed |
Ensures fine detail resolution |
high realism |
Photographic-grade realism |
Secrets for First-Try Success
Three tips that push your success rate from 50% to 90%:
1. More specific action descriptions = better results
| Vague | Specific | Difference |
|---|---|---|
people on piano |
figures cleaning keys with brushes and buckets |
Specific version gives figures clear poses and props |
tiny workers |
tiny figures in blue overalls carrying mops |
Adding clothing details creates narrative |
2. Always specify lighting type
soft lighting is the safest default. For specific moods:
warm golden hour light from the right→ warm side light, nostalgic feelcool blue ambient light→ cool tones, modern feeldramatic top-down spotlight→ overhead light, theatrical feel
Without lighting specification, the AI picks randomly and results become unpredictable.
3. Don't overcrowd the scene
3-5 figures cleaning keys is enough. Writing 20 figures doing different activities including painting, dancing, playing music, cooking... overwhelms the AI and produces chaotic compositions.
For a different approach to miniature scenes, check out the miniature 3D world in container guide—it achieves a similar effect through an entirely different method.
Upgrade Challenge: 3 Variations on the Base
Once you've mastered the basic version, try these three variants. Each requires changing only a few words:
Variation 1: Season swap — Winter piano
...tiny human figures skiing and ice skating on the piano keys
covered with miniature snow, winter holiday atmosphere...
Keys are dusted with micro-snow, figures ski and skate across them.
Variation 2: Instrument swap — Cello repair
Tilt-shift photo of a cello, tiny human figures climbing the
strings and repairing the bridge with miniature tools...
Figures climb cello strings and fix the bridge with tiny tools.
Variation 3: Mood swap — Midnight piano
...dim candlelight illuminating the scene, mysterious midnight
atmosphere, tiny figures in nightgowns exploring the keys...
Candlelit midnight piano exploration—mood shifts from "playful" to "mysterious."
5 Common Beginner Mistakes
| Mistake | Why It Happens | Fix (append to prompt) |
|---|---|---|
| Figures are blurry | They landed outside the depth-of-field zone | with the tiny figures positioned in the sharp focus zone |
| No tilt-shift feel | Wrote shallow DOF but forgot tilt-shift | Make sure tilt-shift photo appears at the start |
| Keys look flat | Missing material description | ivory-textured white keys and glossy black keys |
| Figures are wrong scale | AI isn't sure how small "tiny" is | figures approximately 1cm tall relative to the piano keys |
| Colors too washed out | No light direction specified | soft warm studio lighting from the upper left |
Each fix is appended to the end of the original prompt—no rewriting needed.
Not sure about the results? Try both the base and fixed versions in nanobanana pro to compare the difference yourself.
FAQ
What's the difference between tilt-shift and regular blur?
Regular blur (Gaussian blur) softens the entire image evenly. Tilt-shift blur is directional—a horizontal band in the middle stays sharp while the top and bottom gradually soften. This specific blur distribution is what tricks your brain into seeing a model. Use tilt-shift instead of blurred to trigger this specific pattern.
What scenes work best for tilt-shift miniature besides pianos?
Any object with repeating structures and flat surfaces: bookshelves (figures walking on book spines), keyboards (figures camping between keycaps), sushi platters (figures sunbathing on fish slices). The key is choosing surfaces with enough room for figures to "live on."
Do figure actions affect the overall result?
Significantly. Labor scenes—cleaning, repairing, carrying—succeed most often because these actions involve clear tools and poses. Social scenes—dancing, chatting—are harder because the AI must handle complex character interactions. Start with simple actions.
How can I use these images in real projects?
Tilt-shift miniature style works well for: music festival posters, instrument brand social media content, creative studio portfolio covers, and any visual that needs a "double-take" effect. After export, add text and branding in Photoshop or Canva.
Can I shift the focus zone away from center?
Yes. Add with the sharp focus zone on the far right keys to move the sharp area. But note: the most natural tilt-shift focus position is center to lower-third of the frame. Extreme shifts may look unnatural.