Microphone Test: Level, Device and Playback

What This Microphone Test Checks

Three things go wrong before a call, and they look identical from the outside: the microphone is muted, the wrong device is selected, or the right device is selected and producing almost nothing. A meeting app will show you the same silent participant in all three cases. This page separates them — the bar moves or it does not, the device name is right or it is not, and the recording sounds like you or it does not.

Everything happens in your browser through the standard audio APIs. Nothing is uploaded, nothing is stored, and the recording exists only until you close the tab. That is worth stating plainly, because granting microphone access to a web page is a real decision and you deserve to know what happens next.

People arrive here having searched a dozen ways: a mic test, a mic test online, a microphone test online, a mic check before a meeting, a mic test tool for a new headset, an online mic test tool before an interview, or simply a tester mic page. Some want to check microphone hardware after a driver update; others need to check microphone online or check mic online because the meeting starts in four minutes. It is one page and one button, and a test of microphone takes about ten seconds.

Test Your Microphone in Three Steps

1. Press Start test and allow the browser prompt. The prompt appears because you pressed the button — this page never asks for the microphone on load.
2. Speak normally and watch the level. A conversational voice should push the bar into the middle of its range and leave it there. Silence means no signal; a bar pinned to the right means clipping.
3. Record and listen. This is the part the level meter cannot do for you.

That is the whole procedure whether you came to test my mic after plugging in a new one, to test my microphone before recording something, or to test your mic on a borrowed machine. The advice people give each other — test your microphone before the call, check your mic when nobody can hear you, check your microphone settings when they can hear the room instead of you — all comes down to these three steps.

Reading the Level

The bar is the average level of the last few dozen milliseconds, and the number beside it is that same average in decibels relative to full scale: 0 dB is the loudest signal the input can carry, and everything quieter is negative. The colors are fixed to the scale, so a normal voice stays in the green and only a loud one reaches amber and red.

Two markers sit on top of the bar. The gray one follows the peaks — the single loudest moments, which run several decibels above the average — holding each high point briefly before it drifts back. The red one marks the highest peak since you started the test or switched devices. It never moves back, so a pop that clipped while you were looking away is still there when you look.

• No movement at all — look at the number beside the bar. At −∞ nothing is arriving: the microphone is muted, disconnected, or the browser selected a device that is not there, so check the device list above. A figure such as −75 dB means the microphone works and the room is simply quiet, since the browser's noise suppression pushes background hiss below the bottom of the bar. Speak, and it will move.
• Barely moving, around −40 dB — the signal exists but is too quiet. Raise the input gain in your operating system's sound settings, or move closer.
• Sitting near the top, touching 0 dB — the signal is clipping. Peaks are being cut flat, which is what makes a voice sound crunchy and distorted. Lower the gain.
• Peaks between −18 and −6 dB when you speak — this is where you want to be.

Microphone Playback Test

A microphone playback test answers the question the meter cannot: does this actually sound acceptable? Record ten seconds at your normal speaking distance and listen back. A mic playback test catches the faults that never show up as a level problem — a fan under the desk, a keyboard too close to the microphone, a room with too much echo, or the hollow, distant quality of a laptop microphone picking you up from a meter away.

Doing a mic test and playback in one sitting is also the fastest way to compare two devices: record on one, record on the other, and the difference is obvious in a way that no meter reading conveys. The recording is deliberately not downloadable — this is a diagnostic, not a recording studio, and files that vanish when you close the tab are one less thing to worry about.

Headsets and Headphone Microphones

A headset mic test has one extra trap. Plugging in a headset often adds two devices — the headset microphone and the built-in one — and the browser may pick either. If the level looks wrong after connecting a headset, the device list above is the first place to look.

A headphone mic test for earbuds with an inline microphone has the same issue, plus a physical one: the microphone module sits on the cable, near your collarbone, and rustles against clothing. To test headphones and microphone together, record on this page and listen back through the headphones themselves — that is the combination you will actually be using, and problems in either half show up immediately.

If the Browser Refuses

A blocked microphone is remembered per site, and once blocked, the prompt does not come back. The permission has to be reset in the browser's own site settings — in Chrome and Edge through the icon at the left of the address bar, in Firefox through the permissions panel in the same place, and in Safari through Settings for This Website. Full instructions appear on this page if the request is refused, because being told "access denied" and nothing else is the most common reason people give up on a test like this.

Two other requirements are worth knowing. The page must be served over HTTPS — browsers refuse microphone access on plain HTTP entirely. And on shared machines the permission is stored per browser profile, so allowing it in one profile does nothing for another.

Questions

• Is my voice recorded or uploaded?
  The audio stays in the tab. The level meter reads it directly from the audio stream, and the optional recording is held in memory and attached to the player on this page. Nothing is sent to a server, and closing the tab discards it.
• Why does the bar move when nobody is speaking?
  Room noise. A fan, an air conditioner, or traffic outside will register at around −50 dB, which is normal. If the resting level is much higher than that, your input gain is probably too high and the noise will be audible to everyone on the call.
• The device list shows names like "Default". Why?
  Browsers hide device names until you grant permission once. After you allow access the real names appear; if they still do not, the operating system is reporting the device without a label, which some USB interfaces do.
• Does this test my speakers too?
  Only indirectly — you need working output to hear the playback. This page is about the input side; a speaker check is a different test and would only muddle this one.
• The recording sounds processed. Why?
  Browsers apply echo cancellation, noise suppression and automatic gain by default, and the device panel above shows which are active. They usually help on a call and can make a recording sound artificially flat.