The Sandwich Method: Subject + Environment + Light
AI Creative Lead
Sintetografía Avanzada Instructor.
When working with Generative AI like Midjourney or DALL-E 3, coherence is key. A prompt isn't just a wish list; it is a set of instructions. To get consistent results, we use Layering.
1. The Subject Layer (Who?)
This is your anchor. Be specific. "A woman" is too vague. "A 90-year-old grandmother wearing a spacesuit" gives the AI a concrete visual to latch onto.
2. The Environment Layer (Where?)
Without an environment, AI defaults to a generic void or a simple gradient. Force the context. "In a bustling cyberpunk market" or "In a serene zen garden". This sets the color palette of the background.
❌ Bad Practice
Cool robot city night lights.
✔️ Good Practice
A rusty robot sitting on a bench, in a neon-lit futuristic city at night, rain falling.
3. The Lighting/Style Layer (How?)
This creates the mood. Is it "cinematic lighting"? "Soft volumetric fog"? "Hard flash photography"? This layer transforms a simple image into a piece of art.