Microphone Test

What is this tool?

The Microphone Test captures your microphone through navigator.mediaDevices.getUserMedia() and shows what it is receiving: a live level meter in dBFS, a waveform, a frequency spectrum, and a peak indicator you can reset.

It also records locally using MediaRecorder so you can hear yourself back. Playing back a recording is the only way to judge the things a meter cannot show — muffling, room echo, a fan, a keyboard, or a sibilance problem.

Nothing is uploaded. The engine contains no fetch call, no XMLHttpRequest and no WebSocket; the recording exists as a blob in the page and disappears when you close the tab or press DELETE. The page's own FAQ answers “Does the microphone test upload my audio?” and “Where is my microphone recording stored?” on the same terms.

How to use it

1. Press START MICROPHONE TEST and allow accessThe browser will ask for microphone permission — this is a browser prompt, not a page element, and it must be allowed for the tool to work at all. If you previously blocked it, the prompt will not reappear; clear the site permission in the address-bar icon first.
2. Pick the right input deviceIf more than one microphone is available, select the one you actually want to test. A laptop with a headset plugged in typically exposes both, and the built-in one is often the default.
3. Speak normally and watch the levelAim for peaks in the region of −18 to −6 dBFS. Consistently near 0 dBFS means the input gain is too high; barely moving off the floor means it is too low. Use RESET PEAK between attempts so the peak reading reflects the current try.
4. Run the silence testSTART SILENCE TEST measures your noise floor with nobody speaking. This is the number that explains a hissy or noisy recording, and it is measured rather than guessed — stay quiet for its full duration.
5. Record, then play it backSTART RECORDING captures locally; PAUSE and RESUME work mid-take; playback lets you hear the result and DELETE removes it. Listen for problems the meter cannot show: room echo, background hum, plosives, a keyboard.
6. Use the guided test and the checklistSTART QUICK TEST walks through a seven-step sequence with PREVIOUS and NEXT; a fifteen-step full test goes deeper. The checklist covers microphone detected, permission granted, input level responds, voice detected, recording works, playback works, no clipping, background noise acceptable, input not intermittent, channel configuration correct and microphone sounds clear.

What the results mean

The dBFS level reading

dBFS is decibels relative to full scale, so it is always negative and 0 dBFS is the digital ceiling. It describes the signal arriving in the browser, not loudness in the room — there is no calibrated reference and therefore no SPL figure, which the page states in its own “Can this test measure microphone SPL?” FAQ.

RMS versus Peak

RMS follows the sustained average of your voice; Peak catches the single loudest instant. Peak is what tells you whether a plosive or a sudden laugh is going to clip even when the average looks safe.

Clipping

The engine flags samples at or above 0.99 of full scale and reports a clipping rate per second of test time. Anything above zero means the waveform is being cut flat, which sounds like crackle or distortion and cannot be repaired afterwards — lower the input gain and try again.

Noise floor

The silence test reports the level present when you are not speaking. A low floor is the goal; a high one is what you hear as hiss behind your voice on calls. Fans, air conditioning, a nearby laptop and a high input gain all raise it.

Voice activity and intermittent input

The tool detects whether speech is present, which is how it separates “the microphone is silent” from “the microphone is working and you are not talking”. If the detection blinks in and out while you speak steadily, that is the intermittent-input case on the checklist and usually points at a cable, connector or driver.

What it can detect

• A microphone that is not detected at all, or whose permission was never granted
• Input level that is too hot (clipping) or too quiet to be usable
• Digital clipping at 0 dBFS, reported as a rate rather than a single yes/no
• A noisy environment or a hissy input, measured as a noise floor
• Intermittent capture — audio that cuts in and out while you speak steadily
• Recording or playback failing on your browser/device combination

Limits worth knowing

A browser can only report what the platform gives it. These are the honest boundaries of this page, so a result is never read as more than it is.

• Nothing you say leaves your device. The engine has no fetch, no XMLHttpRequest and no WebSocket; recordings are held in memory by the page and are gone when you delete them or close the tab. No account, no upload, no server-side processing.
• It cannot measure microphone sensitivity in mV/Pa, frequency response, or SPL. Those need a calibrated reference source and a measurement rig; the browser only receives the digitised signal your system chose to hand it. The page says the same under “Browser Microphone Test Limitations” and in its sensitivity, frequency-response and SPL FAQs.
• What you see has already passed through your operating system's processing. Echo cancellation, noise suppression and automatic gain control are usually on by default and will change the level, the noise floor and the spectrum. Turning them off is a system setting, not something this page can do for you.
• There is no way to tell a hardware fault from a driver or OS problem from inside a browser tab. A microphone that reads silent here may be muted at the OS level, routed to a different device, or physically switched off.
• The spectrum is an FFT of the incoming signal, useful for seeing where energy sits. It is not a calibrated frequency-response plot of your microphone.

Tips

• Speak from about a hand's width away and slightly off-axis. That single change fixes more plosive and boominess complaints than any setting in this page.
• Test in the room you actually work in, with the same fan, air conditioning and window state, so the noise floor you measure is the one you live with.
• If your system has echo cancellation, noise suppression or auto gain control, note whether they are on. A recording that sounds thin or gated is often the processing, not the microphone.

Troubleshooting

Related documentation

• AudioMic Latency TestObserved acoustic loopback latency over ten rounds, with jitter and P95.
• AudioHeadphone TestEars, balance, separation, bass and treble checks built for headphones and earbuds.
• AudioSpeaker TestLeft, right, stereo, bass, treble and sweep tones for laptop or desktop speakers.
• AudioAudio Frequency TestA 20 Hz to 20 kHz tone generator with sweeps, steppers and A/B comparison.
• AudioStereo TestChannel isolation, balance, separation and phase checks for any stereo setup.