Audio Reaction Test

Tap the pad to start, then click or tap it or press any key the moment you hear the beep. Your average of five tries, in milliseconds.

More options

The standard run. A 1,000 Hz beep lasting 150 ms plays in both ears after a random wait of 2 to 5 seconds, and the pad looks exactly the same when it sounds. The result is the average of five tries in milliseconds.

Turn your device down before the first beep, then raise it until the beep is clear and comfortable. Nothing sounds until you press a button. Headphones on a wire answer sooner than wireless ones.

–This run

–Your best

–Last run

5 signals. After a random wait of 2 to 5 seconds the page plays a 1,000 Hz beep of 150 ms in both ears, and the pad does not change when it sounds. The result is the average of your times in milliseconds. A press before the signal is a false start and the try is repeated. Your sound output adds a delay that the page cannot measure and never subtracts.

No runs yet. Your last ten will show up here.

Saved in this browser

Next step Now react to a light: the lamp test

All 24 rounds and notes

Change the sound 6

  • Beep, five tries The standard run. A 1,000 Hz beep lasting 150 ms plays in both ears after a random wait of 2 to 5 seconds, and the pad looks exactly the same when it sounds. The result is the average of five tries in milliseconds.
  • Click, one millisecond A single click lasting 1 ms, the length of one cycle of the standard beep. There is no pitch to recognise, only an edge. A click this short is easy to lose in a noisy room, so take this round on headphones.
  • High tone, 4,000 Hz Two octaves above the standard beep and inside the 2 to 5 kHz band where human hearing is most sensitive, so it sounds louder at the same setting. Research on reaction time finds that stronger signals tend to draw quicker presses. See whether yours do.
  • Low tone, 250 Hz The same five tries with a beep two octaves lower. One cycle of a 250 Hz tone lasts 4 ms, against 1 ms for the standard beep. It keeps its own best, so you can see whether pitch moves your time on your own equipment.
  • Noise burst 100 ms of white noise, which has equal power at every frequency, so catching it does not depend on hearing one particular pitch. Try it if a pure tone is hard for you to pick out of the room around you.
  • Quiet beep The standard beep at a quarter of the amplitude, about 12 dB lower. Weaker signals tend to give longer reaction times, and this round shows you the size of that effect on your own ears. Leave the volume where the sound check put it.
Show all 24 rounds and notes (18 more)

Ears and eyes 4

  • Eyes closed, five tries A run you can take without looking. Press the space bar to start. A soft tick marks the start of each wait, a low double buzz marks a false start, and a rising two-note chime ends the run. Open your eyes to read the result.
  • Left ear only Every beep plays in the left channel, which on headphones means the left ear alone. If you hear it on the right, the headphones are on backwards or the channels are swapped somewhere along the way. Keeps its own best.
  • Random ear Ten tries, each beep in the left or the right channel at random, so you cannot lean on one side. The result lists the left average and the right average separately as well as the overall figure.
  • Right ear only Every beep plays in the right channel. Run it straight after Left ear only and set the two averages side by side. A small gap between them is ordinary run-to-run spread and says nothing about your hearing.

Listening tasks 3

  • Beep over background noise A bed of quiet white noise plays through every wait and the standard beep arrives on top of it. The signal is the same but stands out less, and research on reaction time finds that less salient signals draw slower presses.
  • High beep go, low beep hold Ten signals in random order: seven beeps at 1,000 Hz and three decoys an octave lower at 500 Hz. Press on the high beep and keep still on the low one. The result is your average on the high beeps plus a count of presses on decoys. Deciding takes longer than simply reacting, so expect a slower average.
  • Tick, then beep one second later A tick plays and the beep follows exactly 1,000 ms later, every time. With a fixed gap you can time the press and stop reacting, so averages usually fall. A press inside 100 ms of the beep is flagged as anticipated and the try is taken again.

Set up and compare outputs 4

  • Output A: ten beeps Ten tries scored by the median, which one slow try cannot drag. Take it on your first output, for example wired headphones. The page keeps this result and sets it beside Output B.
  • Output B: ten beeps The same ten tries on a second output, for example a wireless headset or the monitor's speakers. When both are done the page shows B minus A in milliseconds. A gap of a few milliseconds is ordinary run-to-run spread. A gap of tens of milliseconds is mostly the equipment.
  • Sound check, three beeps Three tries and nothing is saved. Start with the volume slider low and bring it to a level where the beep is clear but comfortable, then leave it alone. Every other round assumes the level you set here.
  • Twenty beeps, first five against last five Twenty tries in one run, the longest spell of listening on the page. The result prints the average of tries 1 to 5 beside the average of tries 16 to 20. If the last five are slower, check first that the level you set still sounds clear: quieter signals tend to draw slower presses.

Between the page and your ear 7

  • The browser's audio buffer Browsers process sound in blocks of 128 samples, which is 2.7 ms at 48,000 samples a second, before handing it to the operating system. The page prints the base latency the browser reports for that hand-over.
  • Loudspeakers across the room Sound travels at about 343 metres a second in air at 20 °C, which is 2.9 ms for every metre. Speakers 3 metres away add about 8.7 ms to every try, before the speakers' own electronics are counted.
  • Reported output delay Some browsers estimate the time between handing sound to the operating system and the output device playing it. The page prints that estimate under your result, or says it was not reported. It is never subtracted, because it is an estimate that varies with platform and hardware.
  • Sound blocked until your first press Browsers hold back sound from a page until the visitor has clicked, tapped or pressed a key on it. That is why the first beep can only follow your first press, and why the page tells you if the browser still has sound suspended.
  • Television or soundbar Many televisions, soundbars and AV receivers have an audio delay setting that holds sound back to stay in step with a picture the set has delayed. Whatever it adds goes on every try here. Set it to zero for this test if the menu allows.
  • Wired headphones The shortest path this page can use. The sound goes down a cable and the drivers sit against your ears, so the air adds next to nothing. Make wired headphones your Output A and judge every other output against them.
  • Wireless headphones Over a wireless link the sound is encoded, sent by radio and buffered in the headset before it plays, and every step takes time. How much depends on the headset and the device, so no single figure is honest. Measure your own pair with Output A and Output B.

How this list was compiled: Rounds and notes written for this page. Every round changes one thing about the standard beep run (the pitch, length, loudness or ear of the signal, what you must listen for, or the output you listen on), so two results can be compared fairly. The arithmetic was computed on 2026-09-19: one cycle of a 1,000 Hz tone lasts 1 ms and one cycle of a 250 Hz tone 4 ms; 250 Hz and 4,000 Hz are each two octaves from 1,000 Hz and 500 Hz is one octave below it; a quarter of the amplitude is 20 x log10(0.25), about 12 dB lower; sound at 343 m/s covers one metre in 2.9 ms and three metres in 8.7 ms; a block of 128 samples at 48,000 samples a second lasts 2.7 ms. Facts checked on 2026-09-19. Signal conduction times (auditory 8 to 10 ms, visual 20 to 40 ms), the direction of the sound and light difference, the effect of signal intensity and salience, warning signals and fixed foreperiods, Donders' ordering of simple, recognition and choice reactions, and the 100 to 200 ms floor for a valid response: https://en.wikipedia.org/wiki/Mental_chronometry. The 0.1 second false-start rule set by World Athletics: https://en.wikipedia.org/wiki/False_start. Loudspeakers behind each lane and about 3 ms per metre for the report of a pistol: https://en.wikipedia.org/wiki/Starting_pistol. Speed of sound in air at 20 degrees C, about 343 m/s: https://en.wikipedia.org/wiki/Speed_of_sound. Hearing at its keenest from 2 to 5 kHz: https://en.wikipedia.org/wiki/Equal-loudness_contour. White noise has equal intensity at different frequencies: https://en.wikipedia.org/wiki/White_noise. A2DP audio over Bluetooth is encoded with a codec (SBC is the mandatory one): https://en.wikipedia.org/wiki/List_of_Bluetooth_profiles. No Bluetooth delay figure is quoted because no consistent sourced figure was found. Audio delay settings on televisions, soundbars and AV receivers: https://en.wikipedia.org/wiki/Audio-to-video_synchronization. What the browser's outputLatency and baseLatency values mean: https://developer.mozilla.org/en-US/docs/Web/API/AudioContext/outputLatency and https://developer.mozilla.org/en-US/docs/Web/API/AudioContext/baseLatency. Browser versions that report outputLatency (Chrome 102, Firefox 70, Safari 18.4): https://github.com/mdn/browser-compat-data/blob/main/api/AudioContext.json. Audio processed in blocks of 128 frames: https://developer.mozilla.org/en-US/docs/Web/API/AudioWorkletProcessor/process. Sound held back until the visitor interacts with the page: https://developer.mozilla.org/en-US/docs/Web/Media/Guides/Autoplay. Licence: own work

How it works

This audio reaction test plays a short beep after a random wait of 2 to 5 seconds. The pad looks the same when it plays, so this is an auditory reaction time test: only your ears can help. Tap the pad to start, then click or tap it or press any key at the beep. Five tries give an average in milliseconds, with your browser's reported sound delay underneath.

Read the full guide

How the sound reaction test runs

Turn your device down, press Play the beep, then raise Volume to a comfortable level. Pick Sound check, three beeps in the Round menu and tap the pad. Press at each 1,000 Hz beep, and the next follows by itself. After a false start, tap again. Then play Beep, five tries, Noise burst or Random ear.

What the sound path adds

The clock starts when the page schedules the beep, not when you hear it. In between sit the browser's audio buffer, the computer's mixer, the output device and the air, about 2.9 ms per metre from a loudspeaker. A wireless headset encodes and buffers the sound, so it adds more than a cable. No page can measure that path. Output A: ten beeps and Output B: ten beeps show the gap between two outputs.

  • Set the level once, then leave it. Louder signals tend to draw quicker presses, so a volume change spoils comparisons.
  • Sprinters react to sound too. At major meets the start signal plays from a loudspeaker behind each lane.
  • Use wired headphones if you can, and keep one output, browser and seat between runs.

Questions and answers

Is reaction to sound faster than reaction to light?

In laboratories, usually. Early researchers found an auditory signal reaches the brain's processing centres in 8 to 10 ms, against 20 to 40 ms for a visual one. A web page cannot time the audio hardware the beep passes through, so this sound reaction time test never sets its result beside one from the lamps.

Where does the sound delay in this audio reaction time test come from?

From the browser. Web Audio reports an output latency: the browser's estimate of the time between handing sound to the operating system and the device playing it. Chrome has it from version 102, Firefox from 70 and Safari from 18.4. It varies with platform and hardware, so it is shown and never subtracted.

Can a hearing reaction time test check my hearing?

No. It times a press after a sound you can already hear, at a level you chose, and it diagnoses nothing. A slow run can come from a low level, a loose earbud or a Bluetooth headset's delay. If you are worried about your hearing, see a doctor.

More tests

Show 6 more

Browse Reflex Tests