AI Image Generation Research Protocols — Bulk Image Generator
Transparent study plans for prompt behavior, visual variation, failure modes, four-image choice, and diminishing returns.
8 articles
- Short Prompt or Detailed Prompt? A Test for Finding the Useful Middle
A transparent test of prompt length, instruction density, adherence, diversity, and the point where added detail stops helping.
- 12 Visual Hooks for Static Ads to Test on the Same Offer
Compare twelve visual hook mechanisms on one controlled offer and judge clarity, distinctiveness, truth, and testability before beauty.
- FLUX Schnell Prompt Benchmark: What Changes Across Four-Image Batches?
A reproducible benchmark plan for adherence, diversity, composition, typography, counting, and decision value across four-image batches.
- Prompt Specificity vs. Creative Diversity: A Controlled Image Test
Measure how increasing prompt specificity affects instruction adherence, creative diversity, usefulness, and contradiction across matched batches.
- One Image or Four? Measuring the Value of Choice in AI Generation
Compare one- and four-output workflows using time to acceptable choice, confidence, duplicate rate, generation cost, and reviewer workload.
- A Taxonomy of Visual Variation: 24 Ways an Image Direction Can Change
A coded taxonomy of subject, scene, composition, palette, light, material, era, medium, emotional tone, and other variation axes.
- The AI Image Failure Atlas: 50 Failure Modes and What to Try Next
A transparent atlas of prompt omission, anatomy, counting, spatial, typography, product, cultural, and production failures with repair paths.
- How Many Creative Variations Do You Actually Need? A Diminishing-Returns Experiment
Measure unique-direction yield, duplicate concepts, review fatigue, decision confidence, and time across sequential creative batches.