Test your microphone for speech capture

This tool uses the same browser capture function as PewTranslate’s personal translator. It requests echo cancellation and noise suppression, then saves the audio as a 16 kHz, mono, 16-bit WAV file. Your browser decides whether to apply these settings; after you start the test, open “Microphone and browser settings” to see what it reports for this test.

This is a speech-capture format, not a high-fidelity recording or a microphone-quality grade. Listen for understandable words, unwanted noise and gaps. A moving meter shows only that sound is arriving. Recordings stop at 15 seconds.

The microphone turns off before playback. Audio stays in this tab unless you choose Download WAV. Delete recording removes the clip from this page; it cannot remove a file already downloaded. No account, transcription service or translation server is needed for this test.

A microphone test can reveal missing or unclear sound before speech-to-text, but it cannot tell you whether a recognizer will transcribe the words correctly. Capture, recognition and translation are separate checks.

1. Capture: record a short sentence, pause briefly before stopping, then listen. Check that the intended speaker is audible and words are not cut off. Audio is saved in tenth-of-a-second pieces, so the last fraction of a second before you press Stop may be missing.
2. Recognition: when your chosen speech-to-text service is available, test the same words there and compare the transcript with what was said. This page does not produce a transcript or test service availability.
3. Translation: ask someone who understands both languages to compare the meaning, especially names, numbers and negations. Clear audio does not guarantee a correct translation.

This test stops at the capture step. In PewTranslate’s personal translator, while a translation is read aloud, the microphone sends silence unless the “Earbuds connected” switch is turned on. This test does not repeat that back-and-forth.

The separate live translation service was unavailable when this page was reviewed on November 7, 2026. This local microphone test works without it. For background on that account-based workflow, read the Spanish-to-English sermon translator guide.

1. Begin with someone speaking clearly near the microphone. Record one sentence, leave a short pause, then stop and play it back.
2. If you plan to listen to a speaker across a room, repeat from your intended position during a rehearsal or sound check. A close-up test does not establish that distant speech will be clear.
3. Change one thing at a time: move the device, reduce nearby noise, or choose the intended input in your browser. Record again and listen. A short test does not establish reliability throughout a long service.

If the input is silent, check the physical mute switch and microphone permission. If the browser cannot start capture, try opening the link in your main browser instead of an in-app browser. The on-screen error gives the next step; the test never treats an unsupported browser as a successful recording.

Another website or session can select a different input or apply different processing. The browser’s input sample rate can also differ from this tool’s 16,000 Hz output. A successful clip here is evidence about this test, not a guarantee for another application.

The MDN microphone permission guide explains how websites ask for microphone access. The browser’s getSettings values describe the current track; an omitted property means it was not reported, rather than proving that the feature is off.