What the readability test does
The tester renders your thumbnail at 100%, 75%, 50%, 25%, 15% and 10% of YouTube's recommended 1280-pixel width — from 1280 px down to 128 px. Each tile is drawn at its true size so you see exactly what a viewer sees.
For every size it measures detail retention: how much of the original edge contrast survives. When text-like regions are detected, retention is measured there specifically, because small text is usually the first thing to disappear.
How it works
- The image is scaled down to each test width with high-quality browser resampling.
- The small version is scaled back up to the original size (up to 1280 px) so it can be compared pixel-for-pixel with the original.
- A Sobel filter measures edge strength in both. Retention is the fraction of original edge strength still present, capped to ignore sharpening artefacts.
- You mark each size as readable or not. Your judgement is shown separately from the measurement.
Example
A thumbnail with a bold two-word headline and a smaller subtitle might keep 80% of text-like detail at 50%, 55% at 25% and 30% at 15%. The headline is probably still readable at 15%; the subtitle probably isn't. Mark what you can actually read, then try a version with the subtitle removed or enlarged and compare. The Thumbnail Analyzer shows the contrast values behind the result.
Limitations
- Detail retention is a proxy. It doesn't read text and can't know which words matter.
- Displays with high pixel density show small thumbnails with more physical pixels than this test assumes.
- Thresholds (70% and 40%) are guides for where to look, not pass/fail rules.
- Nothing here predicts click-through rate or views.