Mic test

The test listens to the microphone you choose and measures the sound level as you speak or make a sound. A live level shows whether sound is reaching the page. The peak shows the loudest level measured during the test.

You can switch between raw input and call processing. Raw input asks the device to turn off echo cancellation, noise suppression and automatic gain where the browser allows it. Call processing uses those features, which can change the sound level and background noise.

The page can record up to 10 seconds, then play the recording from memory. It discards the recording when you leave the page. You can use playback to check whether the sound is present and whether it sounds as expected to you.

The live level and peak use dBFS, a scale where zero is the maximum digital level. The peak is the highest level seen so far. A clipping count increases when samples reach or exceed −0.1 dBFS, which means the input is near the top of that scale and may sound distorted.

The noise floor is measured from 3 seconds of silence that the page asks you to provide. It estimates the sound level while you are not speaking. Background sound or room noise can affect it, so compare it with the live level while speaking rather than treating it as a pass or fail by itself.

The verdict can say working, too quiet, clipping, silent, not found, blocked or busy. Silent means the page is receiving little or no sound, often because the microphone is muted by hardware or the system. The device details can show sample rate, channels and whether processing is active; these details describe the input, not a quality score.

If the verdict is too quiet, clipping or silent, check that the selected device is the microphone you intend to use and speak toward it during the test. A quiet result can also come from a muted input or a low system input level. Clipping can happen when the sound reaching the microphone is too loud.

If the test says not found, there may be no available microphone, or a device may have been unplugged while the test was running. If it says blocked, access was denied. Busy means another use of the device may be interfering with this test.

The page links to a microphone fix guide for next steps. If the microphone disappears, select it again after reconnecting it. A Bluetooth headset can switch to a lower-quality call mode during use, and an in-app browser can behave differently, so try the test again in a regular browser if the result remains unclear.

The device picker lets you select an available microphone. The list appears after you allow access. If you connect or disconnect a microphone during the test, the available devices can change, so choose the current device again before continuing.

Try the raw input and call processing settings separately if you want to see how processing affects the readings. The raw setting asks the browser to disable echo cancellation, noise suppression and automatic gain, but the device or browser may not honour that request. Check the processing details shown by the page to see what settings it reports.

A useful check combines the readings with playback. Speak during the live test, note whether the level moves and whether clipping is counted, then record a short sample and play it back. If the readings and playback do not match what you expect, use the linked fix guide to check the microphone and system settings.