Sweep 20 Hz to 20 kHz, record it back, and see where your sound falls away
Processed in your browser. Nothing is uploaded. Updated on 24 Oct 2026.Sweeping from 20 Hz to 20 kHz and recording the result shows where a speaker or a room falls away. A dip at the bottom is usually the box; a spike in the middle is usually the room. Runs through your own microphone, and no audio is uploaded.
Press Allow the microphone and continue so the sweep can be heard coming back.
Pick a Sweep length: 4, 8, 15 or 30 seconds; longer sweeps give a smoother curve.
Set the volume to a normal listening level, keep the room quiet and press Run the sweep.
Read the curve from 20 Hz to 20 kHz and the Within 3 dB and Within 10 dB ranges shown under it.
This tool plays a slow logarithmic sweep from 20 Hz to 20 kHz through your speakers, records it back through your microphone, and draws what survived the journey as a curve in third-octave bands, relative to 1 kHz.
The curve is the speaker, the microphone and the room multiplied together, and nothing on this page can separate them. A dip at 200 Hz might be the speaker, or a standing wave between two walls, or the microphone. Telling them apart needs a calibrated measurement microphone and either an anechoic chamber or gated measurement, none of which a laptop has. Anyone who shows you a browser-made curve and calls it a speaker specification is overselling it.
What the measurement is genuinely good for is comparison. Run it twice in the same place with two different speakers and the difference between the curves is real, because the microphone and the room were the same both times. It is also good at finding the point where a device stops producing bass at all, at showing a broad treble roll-off, and at catching a rattle.
A driver a centimetre across cannot move enough air to make low frequencies, so a typical laptop is already falling away by 200 Hz and has essentially nothing below 150 Hz. The bass you hear in music on those speakers is being implied by its harmonics rather than played. Phones are usually worse. A steep fall at the top of the curve, above 8 kHz, is often the microphone rather than the speaker, because many laptop microphones are filtered there by their driver.
Frequency runs left to right on a logarithmic axis, so each octave takes the same width, which is how hearing works. The curve is normalised so 1 kHz sits at 0 dB. The tool reports the points where the response last stands within 3 dB and within 10 dB of the middle of the curve, and the slope above 5 kHz in decibels per octave.
Both, plus the room, multiplied together. A browser has no way to separate them, because that needs a calibrated measurement microphone and either an anechoic chamber or gated measurement. Treat the curve as a comparison between two setups in the same place, never as a specification for a single component.
Small drivers cannot move enough air to produce low frequencies, so a typical laptop speaker is already falling away by 200 Hz and has almost nothing left below 150. Phones are usually worse. The bass you hear in music on those speakers comes from harmonics higher up rather than from the fundamental.
Room reflections cause it. A reflection arriving a few milliseconds after the direct sound cancels some frequencies and reinforces others, producing peaks and dips that move when you move the laptop. Smoothing reduces the comb; moving half a metre and measuring again shows how much of the shape belongs to the room.
Not by chart, because the sound never reaches the microphone. Use the sweep by ear instead: note where the bass turns into a faint rumble and where the treble vanishes, then compare those two points between one pair and another. Judging headphones that way is more useful than most published specifications.
A logarithmic sweep spends the same amount of time in every octave, so 20 to 40 Hz takes as long as 10 to 20 kHz. Hearing and audio measurement both work that way, which means every part of the range gets an equal number of readings rather than the treble getting almost all of them.
No. Levels are read frame by frame while the sweep plays and only the resulting curve is kept, which is a few dozen numbers. No audio buffer is stored, no file is written, and nothing is uploaded. Stopping the sweep releases the microphone immediately.
Still stuck? Ask about Frequency Response Sweep and anybody who knows can answer. No account needed.
Same job, other tools: Sweep Tone Generator – Sweep smoothly between two frequencies over a time you choose · Frequency Generator – Generate a steady test frequency for speakers, headphones or instruments · High Frequency Hearing Check – Screen the top tones, 4 kHz and 8 kHz, ear by ear, then test higher with a tone generator · Play Audio Backwards – Hear a recording backwards, then forwards, for the reverse singing challenge · Increase Audio Volume – Raise a quiet recording as far as it will go before anything starts to distort · Speed Up Audio – Make audio faster or slower without the chipmunk voice.
In the collection Audio Test Tools.