Sound test for speakers and headphones

Headphone test: what to listen for, in order

Almost every headphone test online is the same page: two buttons marked left and right, a stock beep behind each, nothing else. If both sides make a noise you are told everything is fine — even when the drivers are wired out of phase, and even when the right earcup is rolling off everything below 200 Hz because a solder joint cracked.

Work through the panels above in order and each rules out a different fault. Start with left, right and both: the card on screen lights up for the channel the browser is feeding, so you are comparing what you hear against what the page says it is sending, not against your memory of which button you pressed. Then the stereo comparison, which must sound obviously different in its two halves. Then polarity. Then the sweep, where a failing driver gives itself away.

What this sound test checks, and what each result means

The value of a sound test is not the noise. It is knowing what a particular noise failing to arrive tells you about the hardware.

• Left / right / both confirms routing and that both drivers are alive. A dead channel here is a cable, a connector, or a balance slider left at one extreme — in that order of likelihood.
• Stereo versus mono confirms that two independent channels are actually reaching you. It catches a mono downmix, a bad adapter and a headset stuck in call mode.
• Polarity confirms that both drivers push outward at the same moment. A speaker wired backwards passes every other test on this page.
• The sweep shows the frequency range that reaches your ears, and shakes loose any resonance in the cabinet or the desk.
• The tone generator holds one frequency still so you can locate a buzz you heard go past during the sweep.

Left right speaker test: a swap is nearly always a cable

When a left right speaker test comes back reversed — the page says left, the sound arrives on the right — the speaker is almost never the problem. Drivers do not swap themselves. Work backwards from your ears: a 3.5 mm adapter with the ring and sleeve reversed, speaker wire on the wrong terminals behind the desk, an interface where outputs 1 and 2 were patched in the order they reached you rather than the order they are labelled, or headphones simply put on backwards. That last one sounds facetious and is the most common cause of all.

A channel imbalance — both sides working, one noticeably quieter — has a different shortlist: the operating system balance control, then per-application volume, then the connector. A 3.5 mm plug pushed in only part of the way attenuates one channel and adds a faint buzz, and the last two millimetres fix it. The driver itself is the last suspect, not the first.

An audio test that asks for no permissions at all

Most online audio test pages fetch an MP3 or a WAV from a server, so they cannot tell you anything the encoding did not preserve: a 128 kbps MP3 has already discarded most of what is above 16 kHz, which means a sweep delivered as an MP3 is testing the file rather than your speakers. Everything here is generated by an oscillator inside your browser — exact, instant, and with no permission prompt of any kind. On a managed laptop where microphone access is blocked by policy, a mic-based test page cannot run at all and this one still can. For the input side you need the microphone test, which does ask for the microphone because it has no other way to hear you.

Stereo test: are two channels really reaching you?

The stereo test alternates two notes placed hard left and hard right against the same two notes summed to the centre. The difference should be unmistakable: in the wide half the sound occupies space either side of you, in the centre half it shrinks to a point. If both halves sound the same, you are hearing one channel duplicated. Bluetooth is the usual cause: the A2DP profile carries stereo, the HFP and HSP call profiles carry a single narrowband channel, and the system switches automatically when an application claims the microphone. Conference software often holds the mic open after the meeting ends, which is why music sounds thin for the rest of the afternoon until you quit it.

What a failed polarity check means physically

A driver moves outward when the signal goes positive and inward when it goes negative. If one speaker’s wires sit on the wrong terminals, it moves inward at the exact moment its partner moves outward. At high frequencies the wavelengths are short and the two arrive too far out of step to interact much. At low frequencies, where the wavelength is metres long, the pressure one speaker builds is removed by the other and the bass largely cancels.

That is what the inverted button demonstrates: the tone loses body and becomes hard to place. If your wiring is correct the two buttons sound clearly different. If they sound the same, one channel is already inverted somewhere in your chain and this test is cancelling that fault back out — check the red and black terminals at both ends of each cable. The effect is much weaker on headphones, where the two pressure fields never meet in the air: the centre image falls apart, but the bass does not vanish.

Bass test: what your speakers physically cannot reproduce

Run the bass sweep, 20 to 200 Hz, and expect to be underwhelmed on a laptop. A laptop driver is a couple of centimetres across in a sealed cavity and produces essentially nothing below about 200 Hz. Small Bluetooth speakers reach perhaps 100 Hz, bookshelf monitors 50 or 60 Hz, and 30 Hz at a useful level needs a subwoofer — moving that much air takes displacement a small driver cannot deliver. A silent bottom end is normal, not a fault.

What you are listening for instead is a frequency where something buzzes, rattles or distorts: a loose grille, a resonating desk panel, a cabinet joint. When you hear it, note where the readout was, stop the sweep and dial that frequency into the tone generator to hold it steady while you press on surfaces to find the source.

Using the sweep as a rough hearing-range check

The full sweep doubles as a crude upper-limit hearing check: watch the readout and note the number at the moment you stop hearing anything. Loss at the top of the range is a normal part of ageing and the figures are steadier than people expect — teenagers commonly hear to around 17–20 kHz, people in their twenties to roughly 16–17 kHz, thirties 15–16 kHz, forties 14–15 kHz, falling by about a kilohertz per decade after that.

Treat that as interesting rather than diagnostic. Your result is capped by whichever part of the chain gives out first, and on cheap earbuds that is the earbuds. The hearing that matters for conversation sits between roughly 250 Hz and 4 kHz, which this page cannot assess at all.

What this sound test deliberately does not do

It tests your output path and nothing else. It cannot hear you, so it says nothing about your microphone, your gain staging or your room — that is the microphone test, and if you need to cut a section out of something already recorded, the audio trimmer does that in the browser too.

It also cannot measure loudness. Nothing calibrates the level slider against the sound pressure at your ears: that depends on your system volume, amplifier, driver sensitivity and seating distance. A web page can only estimate decibels through a microphone: the decibel meter does that, labels its figure as an estimate, and gets closer only once you calibrate it against a real meter. Nor is this a frequency response measurement — hearing the sweep at 8 kHz tells you 8 kHz arrives, not that it arrives at the same level as 1 kHz. This is a pass-or-fail check you make with your own ears: the right tool for “is my headset broken before this call starts?” and the wrong one for “how flat is this monitor?”

Where VoiceSnap Pro fits

This page exists because we build dictation software, and dictation lives or dies on audio hardware nobody checked. VoiceSnap Pro is a voice-to-text app for macOS and Windows: hold one keyboard shortcut, speak, and clean punctuated text lands in whatever field your cursor is already in — an email, a Slack message, a code editor. It adds punctuation and paragraph breaks as it transcribes, strips filler words, and saves every dictation to a searchable notes library. It is a one-time $39 purchase, not a subscription.

It has not shipped yet, so there is nothing to download today. Join the waitlist for a single email on release day. In the meantime you can see what dictated text reads like with the browser-based voice typing tool, or compare how fast you speak against how fast you type with the voice typing test.