Catt

Record and transcribe on iPhone and Android. Same account.

Get

Remove Background Noise from Audio and Video, Free

Drop a recording, pick how much noise to take out, download clean speech

Drop a file here or click to browse

MP3, WAV, M4A, MP4, MOV, and more · up to 100 MB per file, no sign-up

Rated 4.8 out of 5 from 1,646 customer ratings

To remove background noise, drop an audio or video file here, choose Strong, Medium or Light noise reduction and press Clean Audio. A speech enhancement model keeps the voice and takes out hiss, hum, fans and street noise, then evens out the volume. Download MP3, WAV, M4A, FLAC or the video with clean sound. Free, no sign-up, files up to 30 minutes.

Last updated · By the ConvertAudioToText team

What the noise remover does

Keeps the voice, removes the rest

A speech enhancement model separates the voice from steady and changing noise: hiss, electrical hum, fans, air conditioning, traffic and room echo tails.

You choose how much

Strong removes the most noise. Medium and Light leave some room sound in, which can sound more natural on music or a busy scene. Even out volume brings quiet and loud speakers to a steady podcast level.

Video keeps its picture

Upload an MP4, MOV, MKV or WebM and get the same video back with the cleaned sound track. The picture is copied as it is, so quality and size stay the same.

How it works

1

Drop your file

MP3, WAV, M4A, FLAC, OGG and other audio, or a video. Up to 100 MB and 30 minutes.

2

Pick the strength

Strong, Medium or Light, volume leveling on or off, and the format you want back.

3

Download clean audio

Cleaning usually takes less than a quarter of the recording length. Listen, then download.

How well it works, measured

What we tested

We took clean read speech in English and Spanish, mixed in five kinds of noise at levels from as loud as the voice to a tenth of its power, and ran each mix through the tool on the Strong setting. Then we scored how clearly the words come through against the clean original, on the standard intelligibility scale from 0 to 1.

NoiseMixesClarity beforeAfter cleaning
Hiss (white noise)40.810.89
Fan and air (pink noise)40.770.87
Electrical hum20.960.97
Rumble (brown noise)21.000.99
Other people talking40.770.77

Higher is clearer. Hiss and fan noise gain the most. Hum and rumble leave the words clear already, so the tool takes the noise out without changing how clear the voice is. Background chatter does not improve: the model keeps part of it as speech, so for a recording in a busy room use Light, listen, and keep the original if it sounds better. Very quiet speech is lifted before cleaning so it is not mistaken for noise.

Frequently asked questions

Yes. It is free with no sign-up and no watermark. You can clean 10 files a day, each up to 100 MB and 30 minutes long.

Steady noise such as hiss, hum, fans and air conditioning goes almost completely. Changing noise such as traffic, keyboard clicks and wind is reduced strongly. Other people talking in the background is the hardest case: the model treats it partly as speech, so use Light or Medium there and listen before you publish.

It does not generate new speech, but on very noisy recordings a quiet word under loud noise can come out softer or sound slightly different. If you need a transcript, transcribe the original file: our transcription reads the original audio, not the cleaned copy.

Yes. Upload the video and keep the output on Same as input. You get the same MP4, MOV, MKV or WebM back with the clean sound track. Choose MP3, WAV, M4A or FLAC instead if you only want the audio.

Your file and the cleaned copy are processed on our own servers and deleted automatically within two days. They are never used to train a model.

Need to convert video files?

Try ConvertIntoMP4. 268 formats supported, free to use.

Ready to transcribe?

Start transcribing for free. No credit card required.

Free for files up to 30 minutes · No credit card required