This project explores using generative tools to make a coherent set of quiet abstract images. I treated it as a way to learn how constraints shape taste and consistency.
The Question
I wanted to see if I could make images I’d want to live with - simple, restrained, and suited to a real space rather than a screen. The space I live in is minimal and calm. I used it as a reference point for the prompts and constraints.
The Approach
The goal was to see whether a small set of constraints could produce a consistent, gentle series.
I focused on:
- Defining style through colour, composition, and medium
- Translating material qualities such as texture and wear into AI-generated images
- Testing whether repetition could produce a coherent body of work
Process
I began by curating and refining a set of abstract references, removing anything decorative or trend-driven. The remaining works shared muted palettes, horizontal field-based compositions, visible texture, and a quiet emotional register.
I then decomposed the style into three elements:
- Colour: Muted blues, greys, off-whites, and warm neutrals with low contrast
- Composition: Minimalist abstract fields with strong horizontal divisions and no narrative elements
- Medium: Painterly surfaces with visible texture, abrasion, and signs of time
Using this framework, I developed a single prompt and generated multiple variations in Midjourney, making small language and parameter adjustments to test consistency. Material and surface descriptors proved more effective than stylistic references.
The Outcome
The images formed a coherent set rather than individual "hero" pieces. That felt more appropriate for the kind of work I wanted to make.
Reflection
For me, this project was about choosing constraints, iterating patiently, and learning what felt right.
Questions of authorship still matter, but here the work lives in the framing, the curation, and the decision to keep things restrained.
Constraint isn’t limitation; it’s a way to stay focused.