Check a stereo pair before a call, a mix, or a hardware return. Everything here is generated at the output, so it tests the path from this tab to the drivers: routing, channel order, cabling, and what the speakers can actually reproduce.
Channel identification
The channel test drives one side at a time, and the panel on screen names the side being driven. Sound and screen should agree.
When they disagree, the cause is almost never the speakers. A reversed pair usually comes from the cables at the back of a desktop amplifier, a headphone adapter wired for a different standard, or a surround configuration in the operating system that has mapped the outputs to the wrong pair.
The third step of the cycle drives both. On a correctly wired pair that images as a single source between the speakers, not as two separate sounds.
The sweep
The sweep is a tone climbing from 20 Hz to 20 kHz over the duration you set, repeating until you stop it.
Three things are worth listening for. Where it becomes audible at the bottom tells you what the cabinet can do: a laptop speaker often produces nothing usable below 150 Hz, and a small monitor below 50 Hz. Where it disappears at the top is mostly about your ears, since hearing above 15 kHz declines with age. Anything between those two points that buzzes, rattles, or suddenly jumps in level is a mechanical problem, either in the driver or in whatever the speaker is sitting on.
Slow the sweep down when you are chasing a rattle. A 60 second pass gives you time to move your head and find where it is coming from.
Steady tone
A held tone is the tool for a resonance you have already found. Move it slowly around the frequency and the problem will usually peak within a few hertz.
The waveform selector changes the harmonic content. A sine is one frequency and nothing else. A square wave adds odd harmonics, which makes it far brighter and much better at revealing a buzz, and far more unpleasant at the same meter reading.
Polarity
Polarity is about which way the cone moves first. Two speakers wired the same way push together; one reversed pushes while the other pulls.
Correlated noise makes that audible. In phase, identical noise in both channels images as a solid block between the speakers. With one channel inverted, the centre cancels and what is left spreads outward and sounds hollow.
The test is a comparison, not a reading. Switch between the two settings while listening. The one that sounds centred is how your pair is wired, so if the centred one is the inverted setting, a speaker cable is reversed.
Level, and why a tone is not music
The level control is in dBFS, the same scale a meter uses, where 0 is full scale.
Music at -18 dBFS and a sine at -18 dBFS are not the same amount of energy. Music has peaks and gaps; a continuous tone has neither, so it delivers far more power to the driver and to your ears. A tweeter fed a sustained high frequency at a level that feels fine for music is the most common way small speakers are damaged.
Start at the default and move up only as far as you need.
What a browser test cannot do
This checks the output path. It does not measure the room, the speaker’s frequency response, or how any of it sounds at the listening position, all of which need a microphone and a calibrated reference.
It also cannot tell you what an operating system does after the browser hands over the audio. Spatial audio, loudness normalization, and channel upmixing all sit downstream and can change the result, which is worth turning off before you trust a channel test.