Remove Background Noise from Audio — Free, In Your Browser
Take the hiss, mains hum, fan, and air-conditioning noise out of a recording and keep the voice. The audio is decoded and re-encoded on your own device with FFmpeg's FFT denoiser, so nothing is uploaded and there is no account, no watermark, and no length limit beyond your device.
Noise reduction of this kind works by learning the constant background hum of a recording and subtracting it from every moment of the audio. That is why it is very good at steady noise — a laptop fan, an air conditioner, tape hiss, the 50 Hz or 60 Hz buzz that mains electricity leaves in cheap microphones — and far less able to help with one-off sounds such as a door slamming, a dog barking, or two people talking over each other. Pushing the strength higher than the noise really needs is what makes a voice sound thin and underwater, so start at Balanced and only go up if the hum is still there.
Common ways to use it
- A voice memo recorded in a noisy room — remove the fan or the traffic behind the words.
- A meeting or lecture recording — cut the air-conditioning rumble before sharing or transcribing it.
- A podcast or voiceover take — clean the noise floor before editing.
- An old tape or vinyl transfer — reduce the constant hiss underneath the music.
It pairs with the rest of the audio tools: record straight in the browser, trim the ends, turn the speech into text, or pull the audio out of a video first.
How to use Remove Background Noise
- Choose an audio file — MP3, WAV, M4A, AAC, OGG, Opus, WebM, or FLAC.
- Pick a strength: Light, Balanced, or Strong.
- Tick "Even out the volume" if the voice drifts quieter and louder.
- Click "Remove Background Noise", listen to the result, then download it.
Features
- Built for steady noise — Fan, air conditioning, mains hum, tape hiss, and room rumble.
- Three measured strengths — On a noisy test recording Light cut the steady noise by about 9 dB, Balanced by 23 dB, and Strong by 40 dB.
- Compare before you save — Play the original and the cleaned version side by side.
- Optional volume levelling — Evens out a voice that moves closer to and further from the microphone.
- Common formats in, MP3 or WAV out — No conversion step first.
- 100% browser-based — The recording is never uploaded, which matters for interviews and meetings.
Frequently Asked Questions
What kind of noise can actually be removed?
Constant background noise: a fan, air conditioning, a computer, traffic rumble, tape hiss, and the mains hum that sits at 50 Hz in Europe or 60 Hz in North America. The tool learns that steady noise and subtracts it. Sudden sounds — a door, a cough, a bark, someone else talking — are not steady, so they stay.
Will it make the voice sound robotic?
It can, if you use more reduction than the recording needs. That thin, underwater sound is the usual sign of over-processing. In our own tests Balanced left the speech level untouched, while Strong also took about 3 dB off very quiet speech — so start at Balanced, listen, and only move to Strong when audible hum remains. You can clean again with different settings as often as you like.
Is my recording uploaded anywhere?
No. FFmpeg runs as WebAssembly inside your browser tab, so the file is processed on your own device and nothing is sent to a server. Interviews, medical notes, and internal meetings stay private.
Why does the first run take a moment?
The audio engine is downloaded once on first use and then cached by your browser, so the next file starts immediately. Processing speed after that depends on your device and the length of the recording.
Which formats can I use?
MP3, WAV, M4A, AAC, OGG, Opus, WebM, and FLAC go in. WAV files come back as WAV; everything else comes back as MP3, which keeps the file small and plays everywhere.
Can it remove echo or reverb?
No. Echo is the room reflecting your own voice back, not a separate noise, so subtracting a noise profile does not remove it. Recording closer to the microphone, or in a room with soft furnishings, is the only reliable fix.