Refresh Rate Estimate
Estimate the browser’s visible rendering cadence from requestAnimationFrame timestamps. A one-second warmup is followed by six seconds of measured intervals. The headline value is 1,000 divided by the median interval in milliseconds, while additional statistics show variation and long gaps.
Key features
- Explicit start, one-second warmup and six-second measurement
- Median-based frequency estimate alongside p5, p95 and p99 intervals
- Mean interval, interval standard deviation, long-gap count and sample count
- Stops when the page is hidden, the window loses focus or its size changes
- Automatic cleanup on cancel, completion and a 15-second watchdog timeout
How to use
- Keep the browser window on the display you want to observe and close intensive background tasks if practical.
- Press Start measurement and keep this tab visible and focused for about seven seconds.
- Read the estimated cadence together with interval variation, sample count and long gaps.
- Repeat after changing browser or display settings; a cancelled run must be started again.
Use cases
- Compare browser rendering cadence before and after a display-setting change
- Observe frame interval stability under different system workloads
- Check whether a browser is rendering at a lower cadence than expected
- Compare repeated measurements on different displays without treating them as hardware certification
Frequently asked questions
Is the reported Hz my monitor’s physical refresh rate?
Not necessarily. requestAnimationFrame reports browser rendering opportunities. Power saving, browser limits, variable refresh rate, system load, multiple displays and background policies can change its cadence. Check operating-system display settings for the configured hardware mode.
How is the main estimate calculated?
The tool sorts positive timestamp differences and calculates their median. Estimated Hz equals 1,000 divided by that median in milliseconds. It does not round the result to a preset mode such as 60, 120 or 144 Hz.
What do percentiles and jitter mean?
The p5, p95 and p99 values are interpolated percentiles of the measured intervals in milliseconds. Jitter here is the population standard deviation of those intervals; it is not a universal hardware jitter metric.
What counts as a long gap?
An interval greater than 1.5 times the median counts as a long gap. These intervals stay in the mean, percentiles and standard deviation. The count is a diagnostic hint, not a confirmed dropped-frame or defective-screen count.
Why did the measurement stop without showing a result?
Hiding the page, moving focus away or resizing the window invalidates the run. A run also stops if callbacks do not complete within 15 seconds or if sample limits are reached. Return to the visible tab and start a new measurement.
Can two runs give different results?
Yes. Browser timing precision, GPU and CPU load, power mode and variable refresh behavior affect results. Compare several stable runs and read the interval variation rather than relying on a single headline number.
Privacy
Frame timestamps and aggregate statistics are processed in this page’s memory. This tool does not transmit the measurements, read screen contents or store results in browser storage. Leaving the page clears the current session.
Comments & questions