Frosted glass, with the legibility measured
The risk of this effect is text on a surface nobody designed. This measures the composited panel and says whether the text survives.
Frosted glass, with the legibility measured
The risk of this effect is that text ends up on a surface nobody designed. This computes the contrast against the composited panel and says whether the text survives.
What the contrast figure means here
The panel's effective colour is the tint composited over the backdrop at the tint's opacity. The ratio is measured against that, not against the tint alone and not against the backdrop alone.
It is a single-point measurement. A real frosted panel sits over whatever scrolls under it, so the honest question is the worst case across everything that can appear there — try the darkest and the lightest thing on your page, not the average.
backdrop-filter is unsupported in some browsers and disabled by some privacy settings. Where it does not apply, the panel is a flat tint at the same opacity, which is exactly the surface this measures.
The effect hides its own failure
Blur makes anything look soft and expensive, including text that is a long way below any threshold. That is why so much glassmorphism fails an audit despite looking considered.
Raising the tint opacity is usually the fix, and it costs less of the effect than people expect — the blur is doing most of the visual work.
Check it against the worst thing behind it
A panel over a hero image passes over the dark corner and fails over the bright sky. Design for the worst case or constrain what can be behind it.