What the Hz estimate means
Hz means cycles per second. Here, the estimate describes browser animation callbacks per second, which may differ from the display's selected or advertised refresh rate.
The test collects consecutive animation callback intervals, discards a half-second warm-up, then uses the median interval to estimate cadence: 1000 divided by median milliseconds. The raw estimate is shown without snapping it to 60, 90 or 120 Hz. Enough consistent samples improve repeatability; they do not turn a browser estimate into a panel specification.
The spread between the 10th and 90th percentile intervals shows how much the typical timings differ. The mean and worst interval retain slower callbacks that a median alone could hide. Runs with fewer than 120 intervals or a wide spread are labeled inconclusive or variable.
A 120 Hz display can produce a 60 Hz browser result
- The operating system may use a lower mode or power-saving policy.
- Adaptive displays may change cadence according to content; this low-workload page may not trigger the fastest mode.
- The browser, compositor or window placement can constrain callback scheduling independently of the advertised panel rate.
- Timer precision and other activity can shift the measured intervals.
Check the selected refresh mode in system display settings, move the window entirely onto one monitor and repeat. Do not interpret a nonstandard estimate as a newly discovered hardware mode. For animation under an actual canvas workload, compare the FPS test.
If only pointer movement or clicks seem wrong, use the mouse button, scroll and movement test to inspect input events separately from animation timing.
What invalidates a comparison
Switching tabs or losing window focus cancels the sample. Changing display mode, monitor, browser or power state between runs changes the conditions. This tool cannot measure panel response time, certify variable-refresh range or identify skipped physical scanouts. A mismatch is a reason to investigate settings, not proof of damage.
Only browser timing intervals are measured; no camera or microphone access is needed. Saved results contain timing statistics and your expectation check.