How to test your microphone before recording
A moving meter is only the first check. Record a short sample, replay the saved sound, then repeat the test in your recording or calling app.
Make a useful 30-second microphone test
Connect the microphone or headset, check its mute switch and put it where you will use it. Wear headphones for playback. Leave your computer cooling as normal: the test should represent the recording setup, not a temporarily silent machine.
- Open Mic Test, select Start mic test and allow microphone access when asked.
- Choose the intended microphone from the list after permission is granted. Do not assume the default is your USB mic.
- Speak normally and check that the live input level responds. Select Record sample.
- Leave a few seconds of room sound, say a normal sentence, a quiet sentence and the loudest line you expect. Stop recording before 30 seconds.
- Play the saved sample. Download it if you want to compare takes, or delete it and record again.
Mic Test keeps its meter, recording and playback on your device. Each recording is up to 30 seconds; the live test stops after two minutes. The download format depends on your browser: WebM, M4A or OGG. These are short checks, so use a recording app for the full session.
Confirm that the selected microphone is the one you hear
A device name alone is not proof. Speak close to the chosen mic, then move away while keeping your speaking effort similar. Watch whether the level and recorded sound change as expected. Do not strike the capsule. If unplugging an external mic causes an app to switch inputs, choose it again after reconnecting and make another saved test.
With a USB microphone, check the hardware gain or mute control as well as the computer’s input setting. With an audio interface, verify the correct input and channel in its control panel and your recording app. A headset may expose more than one audio route; compare the saved file rather than judging by a device label.
On a Mac, the sound input settings show the selected device and available input controls. In Chrome, the microphone permission settings let you review allowed sites and microphone selection. Device and menu names vary, so use the controls your system actually shows.
Read the level meter without chasing a magic number
The meter shows digital input level in dBFS, with zero as the digital ceiling. It is not a calibrated room sound-pressure reading. The Mic Test display follows a short average of the incoming waveform, so it is not a true-peak meter and a modest average does not rule out clipped peaks. Listen to the loudest line and check the saved file.
In a recorder with a peak meter, a practical starting point for ordinary spoken takes is to leave the loudest expected peaks roughly 6–12 dB below full scale. This is working headroom, not a delivery loudness target or a threshold to chase on the browser’s average meter. Raise or lower capture gain while testing your actual loud line. If the recording already crackles, lowering playback volume cannot restore the missing shape.
Use the LUFS Meter on a downloaded file when you need file measurements. Keep that separate from the live input check: the browser and recording app may have different gain or processing.
Listen for the fault before choosing the fix
Use these two real voices to practise the comparison. Each pair is the same narration with added hiss, followed by actual noise cleanup at a matched speech-band level. The examples are not microphone hardware measurements. Notice both the background in the pause and the quieter details during the sentence.
| What your saved sample reveals | Change one thing next |
|---|---|
| Steady hiss or a tonal buzz | Check the input chain; distinguish hiss from electrical hum. |
| Air blasts on “p” or “b” | Try a pop filter or a slightly off-axis position, then repeat. |
| A distant voice with a room tail | Move closer or change the room position, keeping adequate gain headroom. |
| A doubled voice only while monitoring | Keep one monitoring route, then replay the saved sample. |
| Crackle on the loud line | Lower capture gain and record again before adding any cleanup. |
Follow the microphone hiss guide for a steady “shhh,” the hum guide for a buzz, or echo vs reverb for repeated or trailing words. Comparing these symptoms is more useful than applying every effect to every recording.
If there is no sound, isolate the failed step
- No device listed: connect and unmute the microphone, then confirm the operating system sees it.
- Permission denied: review the site’s microphone permission and the operating system’s permission for the browser. Restart the test after changing it.
- Device listed but meter stays still: select the intended input, check gain and mute, and close another app’s microphone session if it is holding the device.
- Meter moves but playback is silent: check headphones, output device and playback volume. Download the sample and try another player.
Apple’s microphone access instructions explain the system permission layer on macOS. Chrome’s microphone help covers the browser layer. Both can matter: granting one does not necessarily grant the other.
Repeat the check where the session will happen
Open your recorder, video app or call and select the same intended microphone. Record the same short script there. Automatic gain, suppression and echo cancellation can change the result between apps. Save the good settings and a reference sample, then recheck after switching headsets, reconnecting an interface or moving rooms.
For a long interview, agree on a backup recording and check free storage and power before starting. A successful browser test confirms that this browser captured a sample; the final saved test in your session app confirms the path you are about to rely on.
Recording credits
Garth Comira and Heather Barnett, LibriSpeech, © 2014 Vassil Panayotov, CC BY 4.0. Excerpts from The Ashiel Mystery and Sense and Sensibility, with added hiss and actual Loudness.ai cleanup, aligned and speech-band level-matched. These pairs also appear in the hiss guide.