DeviceBenchFree online bass and subwoofer test
Bass test that falls until your speakers give up
A sine tone drops from 200 Hz to 10 Hz while you mark two moments: the frequency where the note stops being audible, and the frequency where it stops being a note and becomes a rattle. The second mark is the more useful one, because a speaker that buzzes is being driven past the travel its suspension allows and pays for every minute of it. The browser hears none of this — the marks are yours, and the optional microphone reading carries its own roll-off — so on a laptop, which produces essentially nothing below about 150 Hz, the whole run is a demonstration of that single fact.
- 100% free
- No signup
- 200 Hz to 10 Hz
- Sine only
- Optional mic reading
Turn the volume down before you press play
A sustained sine below 40 Hz is the single most effective way to destroy a small speaker. Nothing in music holds one note at full level for thirty seconds, so a driver never has to survive it; the voice coil heats up and the cone runs out of travel at the same time. Start quiet, raise the level only while the tone is clean, and stop the moment it turns into a rattle rather than turning it up to hear whether it clears.
Ready at 200 Hz. The sweep falls from here to 10 Hz.
Optional: measure what comes back
With the microphone open, the page reads the level at the exact pitch the sweep is on and keeps the loudest reading in each of the 10 bands below. What you get is your speakers and your microphone multiplied together: it can prove a speaker produces nothing at 40 Hz, and it cannot prove one is flat. Wearing headphones? Leave this off — the microphone will only hear the room.
On a phone this page is a demonstration rather than a test: a handset speaker is a driver a few millimeters across with no enclosure behind it, and it produces essentially nothing under 300 Hz — the whole sweep happens below the point where it starts working. Plug the phone into something, or run this on the machine the speakers are actually attached to.
How to test bass response without wrecking the speaker
Three moves, and the first one is the reason this page carries a warning at all.
Turn the level down before anything starts
Nothing in music holds one low note at full level for half a minute, so no driver has ever had to survive it, and this sweep asks exactly that of yours. Start with the level slider near the bottom and raise it only while the tone is clean. The instinct to turn it up when the note gets faint is precisely the instinct that destroys a small speaker: the voice coil heats and the cone runs out of travel at the same moment.
Mark the two moments as they arrive
Press the first button the instant the note stops being audible, and the second the instant it stops being a note. Those are different events and they rarely happen at the same frequency — plenty of speakers keep making noise well past the point where the noise has anything to do with the pitch being sent. Both marks record the frequency the sweep was on at that millisecond, and either can be replaced by pressing again.
Hold on anything that sounded wrong
Press Hold and the sweep parks on the frequency it had reached and keeps sounding it as a steady tone, which is the only way to work out whether a buzz is coming from the cone, the port, the grille or the desk it is standing on. Press a finger on each suspect while the tone runs; the one that changes the sound is the one at fault. Carry on down picks the sweep up from that pitch with the remaining time.
Technical specifications
| What a browser can observe | Nothing about the sound in the room — the two marks come from you, and the optional bars come from a microphone with a response of its own |
|---|---|
| Sweep | 200 Hz down to 10 Hz, 4.32 octaves, exponential so each octave takes the same time |
| Sweep length | 15, 30 or 60 seconds, which is 3.5, 6.9 or 13.9 seconds per octave |
| Waveform | Sine only — a harmonic-bearing shape would be reproduced through its harmonics and prove nothing |
| Level | Adjustable from -34 dBFS, capped at -8 dBFS, and changeable while the sweep runs |
| Hold | Parks the sweep at the pitch it had reached as a steady tone, then resumes with the time that was left |
| Microphone reading | Narrowband, from the FFT bin nearest the tone at 2.9 Hz per bin, in 10 bands from 200 Hz to 20 Hz |
| Microphone processing | Echo cancellation, noise suppression and auto gain all requested off, because each one would erase the effect being measured |
Frequently asked questions
I can feel something but I cannot hear a pitch. Does that count?
That is a different sense answering, and it deserves its own mark. Below roughly 20 Hz the ear stops resolving a pitch and the sensation moves into your chest, the desk and the chair — real, physical, and produced by a speaker that is genuinely working. It is also how a subwoofer demonstration convinces people: what they remember as hearing 15 Hz was pressure they felt. Mark the point where the pitch went, not the point where the sensation did.
My speaker buzzes at 60 Hz but is clean again at 40 Hz. How can it get better going down?
Because what buzzed was almost certainly not the speaker. Resonance is frequency-specific: a loose grille, a port tube, a panel on the cabinet, a coin on the desk or the desk itself all have one pitch at which they rattle and are silent either side of it. A driver failing from excursion behaves the opposite way and gets steadily worse as the frequency falls. Use Hold at the buzzing frequency and press on things one at a time — whatever silences it is the source.
My laptop produces nothing below 150 Hz. Is it faulty?
No, it is arithmetic. Sound pressure at low frequency depends on how much air a driver moves, which is its area multiplied by how far it travels, and it has to move four times as much for every octave down at the same loudness. A laptop speaker is a driver a couple of centimeters across with two or three millimeters of housing behind it, so there is nothing to move and nowhere to move it. Every laptop maker fakes the rest with harmonics, which your ear reassembles into a bass note that was never there.
The tone gets quiet near the bottom. Should I turn it up so I can hear it?
No — that is the mistake this page exists to warn about. Your ear is far less sensitive at 30 Hz than at 300, so getting equal loudness down there needs vastly more power at exactly the frequency where the driver has the least headroom left. Turning it up until you hear something is how a speaker is destroyed in one sitting, and the result would be worthless anyway: you would have found the level at which your speaker distorts, not the frequency at which it stops.
My subwoofer is advertised down to 20 Hz but this says it stops at 35. Who is lying?
Both numbers can be true, because the specification does not say at what level. A serious measurement quotes the frequency at which output has fallen 3 dB, an optimistic one quotes the 10 dB point, and a marketing figure often quotes the lowest frequency at which the driver makes any sound at all. Room position moves it as much again — a subwoofer in a corner gains several decibels at the bottom that the same box loses in the middle of the floor.
Why is the sweep a sine? A square wave would be easier to hear.
It would be easier to hear and it would be worthless. A square wave at 30 Hz also contains 90, 150 and 210 Hz, and those your speaker can reproduce perfectly — so you would hear a clear tone, conclude the speaker reaches 30 Hz, and have heard nothing of the kind. A sine contains one frequency and nothing else, which makes it the only waveform that can answer the question this page is asking.
With the microphone on, the last two bands always read as nothing.
Those bands are usually your microphone rather than your speakers. Almost every laptop and headset microphone has a deliberate high-pass filter around 80 to 100 Hz, put there to remove handling noise, air conditioning and desk thumps, and a MEMS capsule the size of a grain of rice has very little response left underneath it. The bar chart is your speakers and your microphone multiplied together: it can prove a speaker produces nothing at 40 Hz, and it cannot prove one is flat.
About roll-off, excursion, and the difference between hearing bass and feeling it
No speaker has a frequency below which it produces silence. What it has is a roll-off: a slope where output falls away octave by octave, gently in a sealed box and steeply in a ported one below the port's tuning. That is why a specification is meaningless without the level attached to it, and why the same cabinet is honestly described as reaching 45 Hz or 30 Hz depending on whether the writer used the 3 dB point or the 10 dB point. The physical limit underneath all of it is displacement. Radiated pressure depends on the volume of air a cone moves, and holding loudness constant means moving four times as much air for every octave you descend — so a driver has to travel four times as far at 40 Hz as at 80 Hz to sound equally loud, and at some point the suspension simply runs out. That is the moment your ear hears as a buzz, and it arrives before the driver goes silent, which is why this page asks for two marks instead of one.
The second thing worth separating is hearing from feeling. Sensitivity collapses at the bottom of the range: a 30 Hz tone needs far more sound pressure than a 1 kHz tone to seem equally loud, and by 20 Hz the pitch sensation has largely handed over to your chest, the desk and the floor. Manufacturers exploit the gap with psychoacoustic bass, which synthesizes the harmonics of a note the speaker cannot produce and lets your auditory system reconstruct a fundamental that is not in the air at all — a real effect, and the reason a phone can seem to have bass. It is also why the tone here is a bare sine and why the tone generator is the page to use if you want a shape with harmonics in it. At the opposite end of the same argument, the hearing range test walks up to 20 kHz to find where your ears stop rather than where your speakers do.
One routing caveat decides whether this page is testing the thing you think it is. A browser sends stereo to whatever output your system has selected, so a subwoofer fed from a receiver's own bass management is being tested through the crossover rather than directly, and the dedicated low-frequency channel in a 5.1 layout is not addressed at all — that one belongs to the surround sound test. Monitor speakers deserve their own warning: two watts behind a plastic bezel means most of what sounds like distortion at 80 Hz is the chassis resonating rather than the cone misbehaving, and pressing a hand on the frame while the tone holds will usually settle the question. If the monitor is new enough that you are still deciding whether to keep it, the black screen and white screen fields find the panel faults in the same sitting, and the speaker test confirms both channels are alive before you conclude one of them is weak at the bottom.
What the microphone does and does not do here
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 optional microphone is read as numbers and never as audio: each frame takes the level at one frequency out of the spectrum analyzer, keeps the loudest value per band, and discards the samples. No recorder is created, nothing is written to a buffer that could be replayed, and pressing Turn the microphone off ends the track, which is the recording indicator in your browser tab going out.