Controller Latency Test

This controller latency test measures sample age: how old the newest gamepad reading is when the page draws it. It also notes the age at the moment a button goes down. That is the browser’s share of input lag. It is not your monitor, and it is not the full delay inside a game.

This page focuses on sample age and input timing. Every other check is still here.

Inputs stay in this browser. Nothing is uploaded.

Press any button on the controller. Browsers hide gamepads until you press a button on this tab.

Browser-observed rate—
Sample age (ms)—
Left stick0.00
Right stick0.00
Offset—

This is how often the browser hands us a new sample. It is not the controller’s USB poll rate. How old the newest sample is when the page draws it. This is not display lag and not in-game lag.

Buttons

Face

YTriangleXB30.00
BCircleAB10.00
ACrossBB00.00
XSquareYB20.00

D-pad

Up0.00
Right0.00
Down0.00
Left0.00

Shoulders and triggers

LBL1LB40.00
RBR1RB50.00
LTL2ZLB60.00
RTR2ZRB70.00

Center and sticks clicks

ViewCreate−B80.00
MenuOptions+B90.00
XboxPSHomeB160.00
LSL3LSB100.00
RSR3RSB110.00
ShareTouchpadCaptureB170.00

Sticks

Left stick0.00, 0.00 · 0.00
Right stick0.00, 0.00 · 0.00

Up on the plot is up on the stick. The numbers keep the browser’s sign: positive Y is usually down.

Vibration

Motion

Gyro, accelerometer, touch position, and battery are not part of the standard Gamepad reading. Chrome and Edge on desktop can request a DualSense or DualShock 4 over WebHID.

Full checkup

  1. Buttons
  2. Sticks
  3. Triggers
  4. Rumble
  5. Review

A short pass over buttons, both sticks, both triggers, and vibration.

Raw data
Raw values appear here after a controller connects.

How to measure input lag in the browser

  1. Wake the pad. Press a button so the browser starts delivering samples.
  2. Watch sample age. The chip is the median age of fresh timestamps. Lower is a fresher report.
  3. Tap a face button. The press reading is the sample age at the frame we first saw it down.
  4. Change the link. Repeat on USB and on Bluetooth before you blame the pad.

Gamepad timestamps and `performance.now()` share a time origin in browsers that implement the spec cleanly. Subtracting them gives the age of the sample at the moment we read it. If a browser’s timestamp is not comparable, the chip says so instead of showing a nonsense millisecond value. The polling rate test is the companion measurement.

Sample age vs end-to-end lag

End-to-end lag is everything from your finger to the photons: the pad, the radio or the cable, the OS, the game, the GPU, and the display. Sample age is one slice, the slice this website can see. A 4 ms wired age and a 20 ms Bluetooth age is a meaningful difference in that slice. A 2 ms difference is noise. A 60 Hz television can add a frame on its own, which this page will never show.

How to read your result

  • Single-digit milliseconds on USB in Chromium is a healthy browser path.
  • A jump when you switch to Bluetooth is the radio and the stack, which is what you wanted to see.
  • A number that changes wildly while the pad sits still means the timestamp source is noisy. Compare presses instead of one frame.
  • “Timestamp cannot be compared” means this browser is the wrong tool for a millisecond claim. Buttons and drift still work.

Ways to lower controller latency

Prefer USB when the milliseconds matter. Close remappers you do not need. Keep the test tab focused. In a game, use a wired pad, a high-refresh display, and the game’s low-latency mode if it has one. Those last two do not move the number on this page, and they do move what you feel. The practical list is How do you reduce controller input lag?.

Frequently asked questions

Is this the same as input lag?

It is the browser-visible part of input lag, also called sample age and press-to-frame timing here. Full input lag includes the game and the display.

What is a good sample age?

On desktop Chromium over USB, a small single-digit figure that stays stable is a good browser path. Do not chase 0. Bluetooth is often higher and can still be fine for most games.

Why does the number differ from a YouTube latency video?

Those tests usually photograph a screen and include the display. This page stops at the browser. Both can be correct and still disagree.