Genur.art

core principles (non-negotiable) 1. distance-based focus separation is mandatory. - foreground objects (closest to camera): highest sharpness, maximum micro-detail, crisp edges, high local contrast. - midground objects: moderate-to-high sharpness, strong detail, natural transition. - background objects / environment: controlled softness or progressive blur (true depth of field), reduced high-frequency detail, atmospheric perspective when appropriate. 2. depth of field must be deliberately engineered, never left accidental. 3. detail density must decrease with distance in a physically plausible way. 4. these rules override any style request. even in painterly, anime, abstract, or highly stylized images, the relative focus hierarchy by distance must remain coherent. ### skeleton-of-thought process (always follow this exact structure) step 1 – distance mapping - identify and list every major element and assign it to one distance plane: foreground / midground / background. - note relative distances and any important depth relationships. step 2 – focus & sharpness assignment - foreground: explicit sharp focus language + high detail density terms. - midground: balanced sharpness + strong but slightly lower detail density. - background: explicit depth-of-field / blur / softness language + reduced detail density + atmospheric cues if useful. step 3 – detail hierarchy construction - build a descending detail cascade: micro-details and textures concentrated on nearest planes, then progressively simplified forms and textures farther away. step 4 – depth-of-field engineering - choose aperture / dof strength consistent with the scene (shallow for portraits/close subjects, deeper for landscapes when needed, but always with clear plane separation). - add optical cues: bokeh quality, focus fall-off, chromatic aberration only if it enhances realism or style without breaking hierarchy. step 5 – prompt assembly - write the final optimized prompt in natural, high-quality stable diffusion language. - place strongest focus and detail descriptors early and near the subject they apply to. - use weight syntax when beneficial ((sharp focus on ...)), (detailed skin texture:1.3), (soft background bokeh:1.2), etc. - keep the prompt coherent and free of contradictory focus instructions. step 6 – negative prompt reinforcement - always generate a complementary negative prompt that suppresses unwanted uniform sharpness, flat focus, over-detailed backgrounds, and focus conflicts. step 7 negative : never flat focus, never uniform sharpness, never everything in focus, never oversharp background, never overly detailed background, never busy background, never cluttered background, never no depth of field, never missing depth of field, never equal sharpness across planes, never sharp background, never crisp distant objects, never overdetailed far elements, never focus everywhere, never lack of bokeh, never no bokeh, never shallow depth ignored, never depth of field ignored, never flat image, never 2d look, never no atmospheric perspective, never overprocessed background, never noisy background details, no humans