Mic Quality Test

This mic quality test records you reading two short lines and scores the sound out of 100. Instead of a single "your mic works" message, it measures seven things that listeners hear: how far your voice stands above the room, how noisy the room is, how close your peaks come to clipping, whether anything clipped, low rumble, mains hum and how much treble survives. Each one gets points, a bar and, when it is weak, a specific fix. It works with USB, headset, laptop, webcam and phone mics, and with XLR mics through an audio interface.

How to run the mic quality test

1. Pick your mic and choose Raw mic to rate the hardware, or Call processing to rate what calls apps would send.
2. Press Start quality test and stay silent for three seconds while the tester measures the room.
3. Read the two lines aloud at your normal distance and volume. The test stops on its own when it has enough speech and you pause.
4. Read the score, the breakdown and the advice under the chart.
5. Play the recording back or save it as a WAV file. Change one thing, such as distance or gain, and run again. Your runs stay listed side by side until you close the page.

The lines are Harvard sentences, phonetically balanced sentences published by the IEEE in 1969 for speech quality testing. The same pair stays on screen until you press Other lines, so you can read identical lines on every run and compare the scores fairly.

How the score is built

90 or more is excellent, 78 to 89 good, 60 to 77 fair, 40 to 59 poor and anything lower bad. High frequencies, clipping and the gap between your voice and the room carry the most points, because they are what listeners notice first. The chart under the breakdown shows your voice in green against the room in grey, with a yellow line where your sound stops.

Your voice is not far enough above the noise

This costs the most points in ordinary rooms. Each halving of the distance between your mouth and the mic adds about 6 dB of voice without adding noise, so moving from arm's length to a hand width adds about 12 dB, usually the biggest single gain in the score. Then work on the room: switch off fans, keep the mic away from the computer, close the window, and put soft things such as curtains or a sofa near you. A dynamic mic or a headset mic hears much less of the room than a sensitive condenser on the desk.

Peaks clip or the recording is too quiet

Clipping means the loudest parts of your words were cut flat. It sounds harsh and cannot be repaired, so lower the gain until your peaks land between -12 and -3 dB. Switch off Microphone Boost on Windows and turn down the gain knob on a USB mic or audio interface. If your peaks stay under -24 dB, the opposite applies: raise the gain, because any boost added later raises the noise with it.

Your recording sounds muffled

The high frequency score falls when the sound stops early:

• Near 3.5 to 4 kHz: phone line quality, from an 8 kHz connection such as old Bluetooth hands-free mode.
• Near 7 to 8 kHz: a 16 kHz connection, typical of Bluetooth headsets in call mode. Use the computer's mic or a wired or USB headset instead.
• Between 8 and 12 kHz: often a built-in mic with heavy processing, a thick foam cover, or speaking to the side of the mic.

Check the format your system uses as well. On Windows, open the Sound control panel, then Recording, then your mic, Properties, Advanced, and pick 48000 Hz under Default Format. On a Mac, open Audio MIDI Setup and set the input format there.

Low rumble or mains hum

Rumble under 80 Hz comes from desk bumps, typing through the mic stand, air conditioning and breath pops on words with P and B. Put the mic on an arm or in a shock mount, turn on its low cut switch if it has one, and speak slightly across the mic instead of straight into it. To see how loud each key of your keyboard is through the mic, try the typing sound test.

Hum is a steady tone at the mains frequency, 50 Hz or 60 Hz depending on your country, often with its multiples. It usually comes from a ground loop between a computer, a charger and audio gear. Unplug the laptop charger for a moment to see if it goes, plug everything into one power strip, and keep mic cables away from power bricks. The mic test also listens for hum live, which helps while you move cables around.

Raw mic or call processing

Raw mic turns off the browser's echo cancellation, noise suppression and automatic gain, so the score rates the mic, the room and your technique. Call processing turns them on, as Zoom, Teams and Discord do. Noise suppression lowers the noise floor and often raises the score, but it can also thin out your voice and cut some treble. Run both to see what processing costs your sound. The discord mic test shows how Discord's voice gate would treat the same mic.

Comparing two mics fairly

Use the same lines, the same distance and the same room, and set the gain so the peaks of both mics land at a similar level. Run each mic twice, because a single run can be thrown off by a car passing outside. For a headset, the headset mic test also checks the earcups and how much of their sound leaks into the mic.

What the score cannot tell you

The score rates one recording: your mic plus your room, your distance, your gain and your browser. It is not a lab measurement. A browser cannot separate the mic's own hiss from the room's noise, measure its pickup pattern, or measure its frequency response, which needs a reference mic and a quiet test room. Singing and instruments use a wider range than speech, so a mic that scores well here can still fall short for music. To see the full spectrum of any sound, use the spectrum analyzer, and to estimate how loud your room is, the decibel meter.

Mic quality test questions