How to record your voice
- Choose the microphone and the mode: Voice for talking, Music or singing for anything with held notes.
- Press Record and allow the microphone. Speak at your normal distance and keep the meter in the green; red means the sound is too loud and will distort.
- Press Pause (or the space bar) to stop for a moment, and Stop and save when you are done. Listen on the page, then press Download.
What the browser's voice processing does to your recording
When a web page asks for the microphone, Chrome, Edge and other browsers switch on three filters unless the page turns them off: noise suppression, echo cancellation and automatic gain control. They were made for video calls, and most online recorders keep them on without saying so. We measured what they do.
The test played a 36-second recording into Chrome 155 as its microphone: a fan-like background hiss, 15 seconds of speech, an eight-second sung note with vibrato and four plucked guitar notes. It was recorded once with the browser's default settings and once with all three filters off, and both recordings were lined up against the original:
| Part of the recording | Default processing (Voice mode) | No processing (Music mode) |
|---|---|---|
| Background hiss between sentences | 12 dB quieter | Unchanged (0.1 dB) |
| Speech | 7.9 dB louder | Unchanged |
| Held sung note, first second | 2.9 dB louder | Unchanged |
| Held sung note, eighth second | 5.4 dB quieter | Unchanged |
| Ring of guitar notes as they fade | Up to 9 dB quieter | Unchanged |
For speech the processing helps: the hiss between sentences drops away and a quiet voice is lifted. For music it works against you. The note started louder and sank by about 8 dB over eight seconds, because noise suppression learns that a steady sound is "background" and turns it down; the natural fade of the guitar strings was cut short in the same way. Noise suppression on its own, without the level control, lowered the held note by 6.7 dB and the guitar's fading ring by 14 to 25 dB.
That is why this recorder has two modes. Voice keeps the browser's processing on; Music or singing turns all three filters off, and in our test that recording matched the input within 0.1 dB. If the microphone you chose has its own noise cancelling, built into a headset or set in Windows or macOS sound settings, it works before the browser and stays on in both modes.
Getting a clean recording
- Distance: a hand's width to a forearm's length from the microphone. Closer sounds boomy and pops on "p" and "b"; much further picks up the room.
- Level: in Music mode nothing evens out your volume, so do a short test and keep the peaks around −12 to −6 dB on the meter. If the page says the recording clipped, move back or lower the input level in your computer's sound settings.
- The room: soft furnishings absorb echo. A closet full of clothes is a classic home vocal booth.
- Headphones: if you play a backing track while you sing, use headphones, or the microphone records the track as well.
After recording
The MP3 is 96 kbps for mono speech and 128 to 192 kbps for music; the WAV is 16-bit at the microphone's own sample rate. To cut silence from the start or end, use Trim Audio; to cut a long recording into pieces, use Split Audio; and if it came out too quiet, use Increase Audio Volume.
The recording is made and saved in your browser. Nothing is uploaded, and the microphone is released the moment you press Stop.