- The tone is meant to be almost inaudible, and it will still move the cone further than anything in your music does. A woofer can bottom out – a hard mechanical knock – without your ears warning you first.
- The tone level starts at zero. Leave it there until the music is playing at your normal listening level.
- Bring it up one step at a time and watch the cone. Stop at the first knock, scrape or hard chuffing from a port.
- A ported box is the risky one. Below its tuning frequency the port stops loading the driver and excursion climbs fast, which is exactly why the test works – and exactly why it bites.
Mixer
Playing:
<waiting audio context>
Bias tone: off
Sample rate: <waiting audio context>
Output level clipping
Digital level leaving the browser, not sound pressure. If clipping lights up, the two sources together have run out of headroom and the distortion you hear is ours, not your speaker's – pull a fader down until it goes out.
Spectrum
The plot starts at 5 Hz here, so the bias tone is visible even when you cannot hear it.
How to run it
- Play music alone, at your normal listening level. Listen to it properly first. It should be clean. If it is already rough, stop here – the test has answered itself and the system is at its limit before the tone has even started.
- Bring the bias tone up slowly until you can see the woofer moving. You should hear little or nothing from the tone itself.
- Listen to the music again. New fuzz, roughness or a wobble in sustained notes is intermodulation: one cone is doing two jobs, and the large slow movement is smearing the small fast one.
- Press the space bar to switch the tone on and off. That comparison, back to back, is the whole test – the memory of a sound fades in seconds.
- Keep an eye on the clipping indicator. If it lights, lower a fader. Digital clipping sounds exactly like a speaker in trouble, and it is the one way this test can lie to you.
More
What this is
Music and a very low tone at the same time, each with its own fader, so you can hear whether large cone movement makes your speakers distort. The tone is low enough to be nearly inaudible and low enough to demand a lot of excursion, which is a combination almost nobody has to hand at home. Everything is generated in your browser; nothing is recorded and nothing is uploaded.
What you are listening for
Intermodulation. A driver only reproduces one signal at a time in the sense that matters here: while the cone is swinging several millimetres for the bass tone, it is also trying to make the midrange, and it is doing that from a moving starting point. The result is frequency and amplitude modulation of everything else – heard as roughness, grain, or a slight warble on sustained notes. Suspension and motor non‑linearity add their own harmonics on top.
Choosing the tone frequency
Lower is not automatically better; the useful range is below the tuning frequency of your box. A ported speaker is usually tuned somewhere between 25 and 45 Hz, a subwoofer lower. Above tuning, the port does the work and the cone barely moves. Below it, the port unloads and the driver takes everything, which is why 10 or 12.5 Hz produces so much movement for so little sound. Sealed boxes have no such corner: excursion still rises as the frequency drops, but the air spring keeps it in check, so a sealed system is the more forgiving one to test.
What this test cannot tell you
It says most about speakers where one cone covers both jobs – full range drivers, small two ways, desktop monitors. On a three way with a separate subwoofer, the driver making the excursion is not the one making the midrange, so the modulation you are listening for largely does not happen and a clean result means less than it appears to. It also cannot separate the driver from the amplifier: a small amplifier asked for a lot of current at 10 Hz may be the thing that gives up first.
A steady tone is the most revealing programme material
Music hides modulation, because there is always something else happening. Pick a sine at 500 Hz or 1 kHz as the main source instead and the effect is unmistakable: a pure tone that starts to warble is doing so because the cone carrying it is moving. Use music to judge whether the effect matters, and a sine to find out whether it is there at all.
Why the tone is so much hotter than the music
At the same fader position the tone sits far above the programme material in dBFS, and that is deliberate. At 10 Hz digital level buys you cone movement, not loudness: your ears barely register the tone however high it goes, while the driver is working harder than it ever does with music. So do not try to match the two by ear or by number – set the music where you normally listen, then bring the tone up until the cone is visibly moving.
The tone stops at −6 dBFS rather than full scale, which leaves room for the programme underneath it. The limit that matters is not the fader position but the clipping indicator.
The master level is the panic control
The two faders here set the balance between the two sources. Audio source is the panel docked at the bottom of the screen – the same one that holds the generators, the sample loops and your own file – and its fader and mute here are the same switch as the play button down there. The level in that panel sits above both faders, so it stays within reach however far down the page you have scrolled. If anything starts to sound wrong, that is the one to pull.