▶ Try the live tool · critique.lbishop.app
A self-initiated exercise in structured visual critique: an eight-dimension rubric (1-5 scale, convergent vs divergent taste) applied to six practice images drawn in code. The samples are not AI-generated and are not an evaluation of any vendor's output.
A working version of the rubric you can use on any image. Score it, pin notes to it, and check your eye against real measurements.
- Score against an 8-dimension rubric, 1 to 5, with the anchor text on hover. A convergent dimension scored 1 caps the verdict at Revise, live.
- Tag every note C or D: convergent craft that reviewers should agree on, or divergent taste that depends on direction.
- Drop numbered pins on the image, with a note and a tag each.
- See what the image does with detail stripped away: grayscale, squint at three blur levels, a 64 px thumbnail, edges, and a focus map.
- Measure, don't guess: luminance histogram, dominant palette with a palette-control score, a two-point contrast probe, and a focal-peak count. Plain image processing, no ML models.
- Compare with my critiques: seven constructed practice samples, each with my written critique to reveal after you score.
- Export as Markdown, JSON or a PNG one-pager. Sessions save in your browser only.
- Private by construction: your image is decoded and measured in the page, never uploaded, and a Content-Security-Policy blocks network requests.
A self-initiated exercise. One rater's rubric, version 1. It does not evaluate any vendor's output and does not replace human judgment.
© 2026 lbishopdev. All rights reserved. View only; see LICENSE.