Microphone Test

Check that your microphone works and is set to a sensible level before a call, a class or a recording. The test records a few seconds, plays them back to you and measures the peak level, the average level, the background noise and any clipping, then gives a plain verdict. The sound stays in your browser and is not uploaded.

How to use this tool

1. Press Calculate and allow microphone access when the browser asks.
2. Talk normally at your usual distance for the length of the recording; the live meter shows the level.
3. Read the verdict and the numbers, and play the recording back.
4. To test a different microphone, change the input device in the browser's site settings (or your system sound settings) and run it again.

Formula

Level (dBFS) = 20 × log10(amplitude), where 1.0 (0 dBFS) is the loudest a digital signal can be

Peak is the single loudest sample; RMS is the root-mean-square average over the whole recording. The noise floor is the RMS of the quietest 10% of 100 ms slices, and signal-to-noise compares the loudest 10% with it. Samples at or above 0.999 of full scale count as clipped. When the quietest moments are within 10 dB of the loudest and above −40 dBFS, the verdict is "Noisy": a fan, hum or a noisy input is nearly as loud as the voice. Voice processing is off by default so you see the microphone itself. Switching to another tab stops the test and the microphone.

Example

A recording whose loudest moment reaches half of full scale peaks at 20 × log10(0.5) = −6 dBFS. If the quiet gaps sit at −62 dBFS and the speech at −20 dBFS RMS, the signal-to-noise ratio is about 42 dB: a clean, well-set microphone.