Use this brief when: Choose gouache for opaque brush color, paper cut for layered shapes, or screen print for distinct flat ink layers.
Acceptance check: Look for the specified medium in edges and surfaces, and check that the florist, village or flower stall remains legible without copied marks or text.
If it fails: If the medium appears only as a filter, describe the physical layers, brush blocks or ink registration more concretely.
Diagnose a medium mismatch before rewriting the brief
The three prompts use different physical cues. Judge the requested medium in the image surface rather than trusting a style label:
| Prompt subject |
Visible medium cue |
If the output misses it |
| Florist in gouache |
Opaque blocks and brush texture around the flowers and storefront |
Replace a broad 'painterly' adjective with opaque pigment and visible brush edges. |
| Dusk mountain village in paper cut |
Separate tree, house and hill layers with readable cut edges |
Specify the foreground-to-background layer order; remove blur language. |
| Flower stall in screen print |
Three flat ink colors with slight registration at edges |
Remove blended gradients and name the three ink layers. |
These are visual tests for future results, not observations of images generated by the copied briefs. A reference to a medium does not grant permission to copy a particular artist's composition, signature or poster.
Hold one flower stall fixed across three media
To compare a medium rather than three different scenes, keep the same small flower stall, two bouquets, blank sign and front three-quarter crop in every written variant. In gouache, describe opaque coral petals and visible brush edges; in paper cut, describe separate bouquet, counter and wall layers with narrow shadow gaps; in screen print, use navy outlines, coral flowers and a cream background as three flat ink passes. The subject and object positions stay fixed while only the surface-making method changes.
If the paper-cut version invents a second stall or the screen-print version moves the sign, reject that comparison before deciding which medium looks stronger. This controlled brief set is an editorial exercise, not a three-image test result. It gives a future reviewer one cause to inspect when the appearance changes.