How do I record and hear myself in this microphone test?

Start, grant access and select the intended input. Press Record 5 s, say Testing, one, two, three, and wait for recording to finish. Press Play in the audio player. The recording is local and may use a compressed codec, so playback is a practical sound check rather than a bit-for-bit copy of raw microphone samples. The page does not play your live microphone through the speakers.

The level moves but I cannot hear the recording. What should I check?

Check that a sample was produced, press Play, and inspect your system output selection and volume. The selected microphone is an input; the player uses your current default output, which can be different. Headphones can help you distinguish playback from room sounds. This workflow does not select output devices inside another app or test individual speaker channels.

Can I test headphone, headset or earbud sound with this sample?

Choose the connected headset microphone in Input, record a normal phrase and play the sample through your selected system output. That provides an audible input-to-recording-to-output check. It does not calibrate headphones, diagnose a cable, compare a named product with review samples, or measure a Bluetooth profile. Some devices expose a different input when their microphone is enabled.

What mic test phrases or script should I use?

Try Testing, one, two, three, followed by Please check the sound of this microphone. Please pass the paper provides plosive consonants; Six soft sounds provides sibilants. Keep your distance and speaking volume consistent when comparing inputs or echo settings. These are short original sample phrases, not standardized lab or intelligibility tests.

Is this a live loopback or microphone latency test?

It is a buffered recording and playback workflow. There is no live speaker monitoring, hardware loopback or calibrated round-trip/input latency result. The capture timer requests about five seconds, and its elapsed window is shown separately from input settings. The waveform bars represent sample slices rather than pitch or frequency.

What happens when I Stop, cancel or replace a recording?

Stop or Cancel closes the current microphone and abandons an unfinished recording. A completed sample remains available until Discard, another recording or leaving the page. Canceled, stale and failed recorder callbacks cannot create a new sample. If a native recorder fails, an actionable error appears and you can try recording again while the meter remains available.

Where is my recording stored, and can I download it?

The sample stays in browser memory and is not uploaded by this tool. This workflow has no recording download or file conversion. Starting another recording, Discard or leaving the page releases it. Only an explicitly changed echo preference is remembered locally; device identifiers and microphone permission are not stored.

Will this confirm my mic works in Discord, Teams, Zoom or Google Meet?

It confirms only what this browser captured from the selected local input and what its player can decode. Use each app's own device selection, permission and test call to check that environment. Stop this page's stream if another app cannot open the microphone. A local sample does not establish network, remote-call or hardware health.

Microphone Playback Test

Record your microphone and hear the sample back in your browser. This workflow opens with record-and-listen instructions and example phrases; nothing is captured until you press Start and grant permission, and recording begins only when you press Record 5 s. Select your intended microphone, speak for about five seconds and press Play in the audio player. The player uses your current system output, which may differ from your selected microphone. Use normal voice and distance to compare samples. Audio stays local in memory, with no live microphone-to-speaker routing or media upload. Discard clears the completed sample; another recording replaces it. Stop cancels an unfinished recording and releases the microphone while retaining a completed sample to hear.

Record, then listen

Start and select the microphone you want to check.
Press Record 5 s, say a short phrase, and wait for the sample.
Press Play in the audio player. It uses your system's current output; check output volume and routing if you hear nothing.

Try: "Testing, one, two, three. Please check the sound of this microphone." Use your normal speaking voice and distance. These are sample words, not a standardized quality test.

Test your microphone

Your browser asks for permission first. Nothing is recorded until you press Record, and nothing audio is uploaded by this tool.

Audio stays localRecordings stay in memoryThe stream closes when you stopNo account, nothing installed

Input

Sample rate: waiting

Channels: waiting

Echo cancel requested

The microphone is closed. Nothing is being listened to until you press Start.

Playbackyour current recorded sample, held until replaced or discarded
This browser cannot record a supported audio format. The meter and the levels can still be tested.
Levelloudest reading since the test began
no readings yet: press Start
Devicethe input the browser opened
no input open yet
Signal observationwhat the measured levels show
press Start and speak to find out

Permissions and processing. Allow microphone access for this site and your browser in system privacy settings. Choose the intended input; a headset, USB microphone or webcam mic can have a different label from the system default. Noise suppression, automatic gain and echo cancellation are requested off by default. Echo settings can be remembered from a previous visit. The browser or hardware may still process audio; the Echo row separates the requested switch from what the track reports. These settings cannot certify another application's processing.

What the numbers mean

RMS and peak describe digital samples in dBFS. The held peak covers this capture run. The low-level estimate is the mean of the quietest tenth of the readings collected over up to the last 20 seconds, not a measured acoustic noise floor or guaranteed signal-to-noise ratio. Frames are sampled roughly every 33 ms while the page runs. Values below -100 dBFS, including digital zero, display at the floor. A steady level can come from speech, a tone, noise or a stuck input; the page cannot identify which. Listen to playback for a practical check.

Peak reading	Digital observation	Next check
-1 dBFS and above	Near or above digital full scale.	Listen for distortion and adjust input gain if needed.
-12 dBFS to -6 dBFS	An example of headroom below full scale, not a quality score.	Judge clarity and background sounds from playback.
Below -30 dBFS	A lower sampled peak; audibility depends on playback gain and other sounds.	Speak normally, check your distance and selected input.
At or below -60 dBFS	A very low digital reading, not proof no samples are arriving.	Check mute, gain, permissions and the chosen device.

Microphone sample phrases

Say "Testing, one, two, three" to check ordinary speech. Add "Please pass the paper" to listen for plosive consonants, and "Six soft sounds" to hear sibilants. Use the same voice and distance for comparisons. These original example sentences are not a standardized intelligibility, frequency-response or manufacturer test.

The sound envelope shows time-domain sample slices, not a spectrum or pitch reading. There is no calibrated latency, hardware loopback, speaker-channel, live speaker-monitoring or acoustic loudness test here. The recording player uses your current system output. For simultaneous camera preview and microphone levels, open Camera and microphone.

Accuracy. Levels describe browser-supplied digital samples in dBFS, not calibrated room loudness or hardware health. Recorded audio stays local in memory; recording codecs may compress it. The stream closes on Stop or leaving the page. No audio is uploaded by this tool. Input sample rate/channels are reported settings, and analysis rate comes from the audio engine. They do not establish the encoding parameters of a compressed recording. The capture timer requests about five seconds; scheduling can change its elapsed window. Playback lets you inspect what this browser captured, not certify headset health, call quality or the inputs selected inside another app.

Common questions

How do I record and hear myself in this microphone test?: Start, grant access and select the intended input. Press Record 5 s, say Testing, one, two, three, and wait for recording to finish. Press Play in the audio player. The recording is local and may use a compressed codec, so playback is a practical sound check rather than a bit-for-bit copy of raw microphone samples. The page does not play your live microphone through the speakers.
The level moves but I cannot hear the recording. What should I check?: Check that a sample was produced, press Play, and inspect your system output selection and volume. The selected microphone is an input; the player uses your current default output, which can be different. Headphones can help you distinguish playback from room sounds. This workflow does not select output devices inside another app or test individual speaker channels.
Can I test headphone, headset or earbud sound with this sample?: Choose the connected headset microphone in Input, record a normal phrase and play the sample through your selected system output. That provides an audible input-to-recording-to-output check. It does not calibrate headphones, diagnose a cable, compare a named product with review samples, or measure a Bluetooth profile. Some devices expose a different input when their microphone is enabled.
What mic test phrases or script should I use?: Try Testing, one, two, three, followed by Please check the sound of this microphone. Please pass the paper provides plosive consonants; Six soft sounds provides sibilants. Keep your distance and speaking volume consistent when comparing inputs or echo settings. These are short original sample phrases, not standardized lab or intelligibility tests.
Is this a live loopback or microphone latency test?: It is a buffered recording and playback workflow. There is no live speaker monitoring, hardware loopback or calibrated round-trip/input latency result. The capture timer requests about five seconds, and its elapsed window is shown separately from input settings. The waveform bars represent sample slices rather than pitch or frequency.
What happens when I Stop, cancel or replace a recording?: Stop or Cancel closes the current microphone and abandons an unfinished recording. A completed sample remains available until Discard, another recording or leaving the page. Canceled, stale and failed recorder callbacks cannot create a new sample. If a native recorder fails, an actionable error appears and you can try recording again while the meter remains available.
Where is my recording stored, and can I download it?: The sample stays in browser memory and is not uploaded by this tool. This workflow has no recording download or file conversion. Starting another recording, Discard or leaving the page releases it. Only an explicitly changed echo preference is remembered locally; device identifiers and microphone permission are not stored.
Will this confirm my mic works in Discord, Teams, Zoom or Google Meet?: It confirms only what this browser captured from the selected local input and what its player can decode. Use each app's own device selection, permission and test call to check that environment. Stop this page's stream if another app cannot open the microphone. A local sample does not establish network, remote-call or hardware health.

Levels describe browser-supplied digital samples in dBFS, not calibrated room loudness or hardware health. Recorded audio stays local in memory; recording codecs may compress it. The stream closes on Stop or leaving the page. No audio is uploaded by this tool.