Input Lag Test

The input lag test measures how much time passes from a key press, mouse click or screen tap to the moment the browser starts handling that event and draws the next frame. You get the result in milliseconds, separately for keyboard, mouse, touch and pen.

How to run the test

1. Click the test box. From then on it is active and accepts key presses.
2. Press any keys, click the box or tap it with a finger. Every press is one sample; holding a key down does not count more than once.
3. Collect at least 20 samples. The verdict appears from the fifth sample, but a median of twenty is much more stable.
4. The "Copy result" button puts a summary with the median, 95th percentile, minimum, maximum and your browser name on the clipboard, which is handy for comparing two computers or settings.

At the start the tool also measures your screen refresh rate from the intervals between frames. That value is needed to judge the result: on a 60 Hz monitor a frame lasts about 16.7 ms, on a 144 Hz one about 6.9 ms.

What exactly the tool measures

Every browser event carries a timestamp that the system assigns when it hands the input over to the browser. The test compares that timestamp with two moments:

• input delay: the time from the event timestamp to the start of its handling by the page. On a light page this is usually a fraction of a millisecond; more means the browser's main thread was busy,
• time to next frame: the time from the event timestamp to the moment the browser prepares the frame with the changed box (measured with requestAnimationFrame),
• estimated time to display: the time to the next frame plus one frame, because the finished picture reaches the screen on the following refresh.

The tool shows the median rather than the average, because single hiccups (for example when the system updates something in the background) would inflate the average. The worst cases are described by the 95th percentile: 95 out of 100 samples were faster.

What the test does not cover

The browser cannot see what happens before the timestamp is assigned or after the picture is sent to the monitor. Outside the measurement are:

• USB polling: the keyboard or mouse reports to the computer at a fixed interval,
• matrix scanning and switch debouncing inside the keyboard,
• image processing in the monitor or TV and pixel response time.

Full "finger to pixel" latency can only be measured with a high-speed camera or a light sensor on the screen. This test shows the part that depends on the browser and the system, which is exactly the part you can most easily improve with settings.

Frame time at common refresh rates

The time to the next frame depends mostly on where in the refresh cycle you pressed the key. That is why the test rates the result in frames rather than in fixed milliseconds.

A result above 2.5 frames is marked as high latency. If, in addition, the handling of the event starts more than 8 ms after its timestamp, the tool tells you that the page's main thread was busy.

USB polling: the part the browser cannot see

Mice and keyboards send data at a fixed rate. The higher the rate, the shorter a press waits to be sent. The average wait is half of the polling period.

How to reduce input lag

• Close heavy tabs (video, online editors, browser games) and turn off unneeded extensions, because they compete for the same CPU.
• Turn on hardware acceleration in your browser settings.
• On a laptop, plug in the charger and choose a performance power mode. Power saving lowers CPU clocks and can cap the frame rate.
• If your monitor supports a higher refresh rate, enable it in the system settings, which shortens every frame.
• Repeat the test in another browser. Differences between browsers on the same computer can be larger than between mice.

Human reaction time is a different measurement; check it with the reaction time test. You can verify individual keys with the keyboard test and frame smoothness with the screen FPS test.