DeviceBenchCups and boom in one run
Headset test that checks both cups and the boom
A 1.5-second pink-noise burst goes to one channel at a time and you answer where it arrived — left ear, right ear, both, or nothing — which is the smallest question that separates a reversed pair from a silent cup from a headset that has stopped carrying stereo at all. Then the microphone opens and the same two bursts run again, because that is the moment a wireless link gives up its music profile, and a run that only ever plays sound with the input closed can never see it happen. A web page has no ears, so those four answers have to be yours; the boom half is measured for you, down to the rate its track is running at.
- 100% free
- No signup
- 6 stations
- ±1.0 hard pan
- ~90 seconds
0 of 4 cup answers in.
Put the cups on the right ears first
The side markings sit where you cannot see them while wearing it — moulded into the inside of the band, or printed on the inner face of a cup. Turn the thing round in your hands and find them before the first burst plays.
Six stations, about 90 seconds. The first two play sound and nothing else. Nothing asks for the microphone until station 3, and you are the one who presses that button.
- Left burst, mic closedWaiting
- Right burst, mic closedWaiting
- Left burst, mic openWaiting
- Right burst, mic openWaiting
- Boom microphoneNot opened
- ChannelsBoth cup stations not answered yet.
- With the microphone openThe second pass has not been answered yet.
- Boom microphoneThe input has not been opened yet.
- Where sound can goThe output list fills in after station 3, because browsers withhold device names until an audio permission has been granted once.
Two of those rows are your ears rather than a measurement, because nothing in a browser can listen back to what it just played. Nor is a page told which Bluetooth codec a link settled on, whether a cable carries a mute switch, or how much charge is left in the cups. The one thing it can read is the input track, and 16,000 Hz or below there is the fingerprint of a headset that has left its music profile behind.
How to test a headset
Sides first, then the same sides again with something listening — the second pass is where a wireless headset gives itself away.
Check which cup is on which ear before anything plays
L and R are printed inside the headband or on the inner face of each cup, often in gray on black where nobody looks. This matters more here than on a speaker check because a headset can be worn backwards, and a headset worn backwards passes every test that only asks whether you heard something. Settle it now and the four answers below mean what they say.
Answer where the burst arrived, not whether it arrived
Each station sends 1.5 seconds of pink noise to one channel and gives you four buttons: left ear, right ear, both, or nothing. That is the whole measurement. Yes-or-no cannot tell a correctly wired pair from a reversed one, and neither of them from a headset that has collapsed to mono — the four-way answer separates all three in a single press. If a burst is too faint to place, use Louder rather than guessing.
Open the microphone, then do the same two bursts again
Station 3 is the one that matters. Opening an input renegotiates a wireless connection, so the two replays afterwards are asking a question nothing else on this site asks: does your headset still put left in your left ear once something is listening? Finish by reading a sentence into the boom, and the run tells you whether anything reached the track at all.
Technical specifications
| Stations in one run | Six — left cup and right cup with nothing listening, the microphone opened, the same two cups again, then a spoken check |
|---|---|
| Burst | 1.5 seconds of pink noise panned to ±1.0, so the opposite channel is silent rather than merely quieter, with a 20 ms fade at each edge |
| Why pink and not a tone | Pink noise carries equal energy per octave, so it is audible on any driver; a single sine can land in a dip in a small cup and read as a dead side |
| Answers per station | Four — left, right, both, nothing — which is the smallest set that separates a reversed pair, a silent cup and a lost stereo image |
| Profile check | The cup answers are taken twice, before and against a live input; an answer that moves from one side to both is the link renegotiating rather than a fault |
| Input read back | What rate the live track is running at and how many channels it carries — 44,100 or 48,000 Hz on a wired or USB boom, 16,000 Hz on wideband speech, 8,000 Hz on the older narrowband profile |
| Silent-input threshold | A peak below -45 dBFS after a spoken sentence is reported as nothing arriving; a boom sitting 2 cm from a mouth normally peaks between -12 and -6 dBFS |
| Price | Free, no signup, and about 90 seconds start to finish |
Frequently asked questions
Both bursts came out of both ears. What does that mean?
Nothing in the chain is carrying two channels any more, and there are three usual reasons. A single four-contact plug pushed into a three-contact headphone socket loses the microphone ring and often sums the sides; a cheap Y-splitter can do the same; and a wireless headset that already had the microphone open is running a mono voice link. Try the run again through a different socket, and if the answer changes the socket was the fault rather than the cups.
The left burst came out of the right cup.
The pair is reversed, and the headband is the first thing to check rather than the cable. After that it is the cable: extension leads with a three-contact plug at one end and a four-contact at the other reverse channels on some layouts, and a few gaming headsets ship a detachable cable that only works one way round. If the sides are reversed with the original cable and worn correctly, the wiring inside the cup has been mixed up at the factory, which is a returns matter and not something a setting can fix.
Sound moved to my laptop speakers the moment I opened the microphone.
The system changed output device underneath you, which a headset invites in a way no other peripheral does. A USB headset registers as an output and an input in one go, and Windows lists a Bluetooth headset twice — once as Headphones for music and once as Headset for calls — then switches between the two entries as applications open and close the microphone. Set the output explicitly in the sound panel rather than trusting the default, and pin the meeting app to the same entry.
My headset has mute and volume on the cable. Does the page see them?
No, and that is exactly why an inline mute is worth checking first. The switch cuts the signal on its way to the sound card, so the browser opens a completely healthy track and simply receives silence down it — indistinguishable, from a page's point of view, from a room where nobody is speaking. The same is true of the inline volume wheel, which is why a cup can be at zero while the operating system swears it is at eighty.
The second stereo pass gave a different answer from the first. Is the headset broken?
No — that is the fault this run was built to catch, and it is normal behavior for a wireless headset rather than a defect. A music link runs one way and can afford stereo; the moment anything opens the microphone the connection has to carry both directions, so it renegotiates to a hands-free profile that is mono and low-bandwidth for as long as the input stays open. Close the microphone and the cups return to stereo within a second or two. The only real fix is to stop taking input from the headset: use a separate microphone and leave the cups doing output only.
One plug, two sockets on my laptop. Which one?
The combined one, marked with a headset icon rather than a headphone icon. A headset plug carries four contacts — left, right, ground and microphone — and a laptop with separate headphone and microphone sockets expects three-contact plugs in each, so a headset in the headphone socket gives you sound with no microphone at all. The adapter you want is a Y-splitter that goes four-contact female to two three-contact males; the ones sold for phones do the opposite and will not help.
Do I have to allow the microphone to test the earcups?
Not for the first two stations — they play sound and ask you a question, and playing sound needs no permission from anyone. The run only asks for the microphone at station 3, from a button you press, and the reason it asks at all is to repeat the cup stations while an input is held open. If you only came to check that both sides work, stop after station 2 and you will have your answer.
About the faults that only a combined device has
A headset is two speakers and a microphone sharing one cable, one connector and — when it is wireless — one radio link, and nearly everything that goes wrong with one is a consequence of that sharing rather than of a part failing. The connector is the first example. A headset plug carries four contacts where a headphone plug carries three, and the extra ring is the microphone; push it into a socket that expects three and the microphone contact lands on ground, which is why a laptop with separate sockets gives you perfect sound and no input from the same plug. Worse, the two competing pinouts that existed before the phone industry settled on one still turn up in adapters, so a splitter bought for the wrong standard can hand you a working microphone and a dead left channel at the same time.
The wireless case is stranger, and it is the reason this page asks the same question twice. Bluetooth audio has two entirely different modes: a one-way, higher-bandwidth music mode that can afford stereo, and a two-way hands-free mode built for telephony that is mono and low-bandwidth in both directions. There is no way to have both at once, because the radio budget will not stretch to it, so the instant any application holds the microphone the link drops to the second mode — and a burst that was clearly in your left ear a minute ago now arrives in both, dull and narrow. Nothing is broken, no setting will avoid it, and the only real escape is to stop using the headset as an input and take the microphone from somewhere else. The one visible fingerprint from a web page is the input sample rate, which is 16,000 Hz on the wideband version of that mode and 8,000 Hz on the older one, against 44,100 or 48,000 on any wired or USB device. That is the same collapse people meet five minutes into a meeting, and the video call test checks it in the company of the camera and the line.
What this run deliberately does not do is settle the questions that belong to a single device. If both cups are fine here and you want to know how the drivers behave — channel balance, a rattle when a low note pushes them, where the bass gives out — the headphone test is built for that, and the sound test answers the more basic question of whether the browser is sending audio to this device at all. The input troubleshooting ladder — a blocked permission, an operating system switch above the browser, a level that never leaves the bottom of the meter — lives on the mic test, which also measures a noise floor properly. And if the headset came with a camera in the same call, the webcam test and the webcam mirror cover that side.
What leaves this page while the boom is open
Every number on this page is worked out by JavaScript running in the tab you are reading it in. Nothing you type, paste or open is uploaded, logged or kept, which is also why the tools carry on working after you disconnect from the network.
The run keeps four button presses and a single number — the loudest level your voice reached — and neither of those is audio. Nothing is recorded here at all: the boom check reads a buffer, takes a peak from it and lets it go, and there is deliberately no playback on this page, because a clip played through open cups while the microphone is live is how you get a howl rather than a diagnosis.