Refresh rate test that shows the spread behind the number

The operating system is reporting what the panel is capable of and this page is timing what it is actually doing, and those disagree more often than anyone expects. The usual causes, in the order worth checking: the rate was never raised from the default in the display settings, the cable or port cannot carry that rate at your resolution — HDMI 1.4 tops out around 144 Hz at 1080p and 30 Hz at 4K — or a second monitor with a lower rate is dragging the whole desktop down on some driver versions. Change one thing, press reset here, and you will know within ten seconds whether it took.

Almost certainly not — that is what a variable-refresh display looks like from inside a browser. G-Sync, FreeSync and the VESA Adaptive-Sync standard underneath them let the monitor change its own timing to match whatever is being drawn, so the interval between frames is supposed to move. The steadiness figure is the giveaway: a fixed panel sits above 95%, an adaptive one running an unsteady workload sits well below it, and the distribution spreads across several columns instead of stacking into one.

Because a short sample is dominated by whatever else happened during it. One garbage collection, one compositor hiccup or one frame lost to a background tab adds a single enormous interval, and over 60 samples that is enough to pull a mean off by several Hz — which is exactly why the estimate uses a median and trims anything past three times it before the second pass. Six hundred intervals make the middle value stable enough that the last digit stops flickering, and the trade is that the number reacts slowly when the conditions change under it.

The tab was in the background while it counted. Browsers cut a hidden tab's animation callbacks down to roughly one a second, and some stop them entirely, so a page that carried on measuring would report a 1 Hz display with complete confidence. This one watches for that, refuses to publish the number and says why — bring the window back to the front and the count restarts from zero.

It can, and the honest way to put it is that the measurement is of the schedule your browser was given rather than of the cable. Some browsers cap animation callbacks below the panel rate to save power on battery, some throttle when the window is partly covered, and a few historically capped everything at 60 regardless of hardware. If two browsers on the same machine disagree, run both with the window fully visible and on mains power, and treat the higher of the two as the panel's real behavior.

Yes, and the number will legitimately move while you watch. A 120 Hz phone drops to 60, 24 or even 10 Hz when the system decides the content on screen is static, and raises it again when you scroll — that is a battery feature, not a fault, and it is why a handset that advertises 120 Hz can read 60 here while sitting still. Keep a finger moving on the glass during the count if you want to see the top of its range.

No: this is how often the display can accept a new picture, and frame rate is how often something manages to produce one. They coincide when a machine has work to spare, which is why an idle browser tab usually reports the panel rate exactly, and they separate the moment something struggles — the display keeps refreshing at its own rate and simply shows the previous picture again. That is why this page deliberately draws nothing at all while it counts.