Free tool

AI audio enhancer

This audio enhancer makes a voice recording clearer in four steps: an AI speech model removes the background noise, a gentle EQ evens out the tone, a de-esser softens harsh S sounds, and the loudness is set to −16 LUFS, the level Apple recommends for podcasts. It all runs in your browser.

  • Free
  • Loudness to −16 LUFS
  • Nothing is uploaded

Choose your voice recording

MP3, M4A, WAV, FLAC, OGG, AAC, or the sound of an MP4, MOV or WebM video, up to 1 GB and 2 hours. Everything happens in your browser; nothing is uploaded.

MP3 encoding by LAME, open-source software.

Enhancing a recording so you can transcribe it? NoteOwl transcribes recordings and meetings on your Mac with each speaker labelled, and keeps them searchable as notes.

How it works

How do you enhance audio online?

Add a voice recording and the enhancer runs at once, with every step on. Press Listen and switch between Original and Enhanced to compare. Turn off any step you’d rather skip, press the button to make it again, and save the enhanced file as MP3, WAV or M4A.

  1. 1

    Add the recording

    A voice memo, an interview, a lecture, a podcast take or a video’s sound: MP3, M4A, WAV, FLAC, OGG, MP4, MOV or WebM.

  2. 2

    Compare

    Fifteen seconds from the playhead, switching between the original and the enhanced sound while it plays.

  3. 3

    Adjust and download

    Untick the EQ, the de-esser or the loudness step, or choose how much noise to remove. Then save the file.

What does the audio enhancer do to a recording?

Four things, in this order: it removes background noise, applies a gentle EQ for speech, softens harsh S sounds, and sets the loudness. Under the result, the page shows what changed: the background level, the loudness before and after, and how long the de-esser worked.

  • Noise removal

    AI noise reduction takes out steady noise such as fans, hum, air conditioning and traffic. In our tests it was about 30 to 40 dB quieter in the pauses.

  • Speech EQ

    Rumble under 80 Hz goes (9 dB less at 50 Hz), the boxy sound around 250 Hz is 2 dB lower, and 3 kHz, where consonants are, is 2 dB higher.

  • De-esser

    While an S or SH carries more energy above 4.5 kHz than the voice below it, that top band is turned down: by 2.7 dB on average in our tests, and never more than 6 dB.

  • Loudness

    The whole recording is measured the way broadcasters do (ITU-R BS.1770) and set to −16 LUFS. A gentle compressor evens out the loudest syllables, and a limiter keeps the peaks under −1 dB.

What loudness should a voice recording have?

Around −16 LUFS for speech played online: Apple recommends that podcasts sit within a decibel of −16 LUFS, with true peaks no higher than −1 dB. A quieter recording makes listeners turn it right up, and they still strain to follow it.

In our tests, a quiet recording with a desk fan behind the voice measured −43.4 LUFS, and an interview over street traffic −26.8 LUFS. Enhanced, both came out within half a decibel of −16 LUFS on ffmpeg’s loudness meter, with true peaks under −1 dB.

What can’t an audio enhancer fix?

Sound damaged when it was recorded: a voice distorted by an input set too loud stays distorted. It is also made for speech: instruments are treated as noise, and in our tests music came down by 12 to 22 dB, so it isn’t a tool for songs.

How is this different from removing background noise only?

The background noise remover takes the noise out and leaves the level and tone as they were. The enhancer does that too, then evens the tone, softens S sounds and sets the loudness, which suits a podcast, a voice-over, or a recording you’ll send to someone.

How do you turn an enhanced recording into notes?

Bring it into NoteOwl on your Mac. NoteOwl transcribes recordings with each speaker labelled, writes a summary of what was said, and lets you search everything you’ve recorded later. All of it runs on the Mac, with nothing sent anywhere.

FAQ

Questions about enhancing audio

Is the audio enhancer free, and is anything uploaded?

It’s free, with no sign-up and no watermark. Nothing is uploaded: the recording is enhanced inside your browser, and the only download is the noise remover, a 197 KB file from this site, fetched once when you add a file.

Will it make my recording louder?

If it’s quieter than −16 LUFS, yes: it is raised by up to 40 dB, with the peaks kept under −1 dB. A recording louder than −16 LUFS is turned down to it. Untick the loudness step to keep the level as it was.

Can I turn steps off?

Yes. The EQ, the de-esser and the loudness step each have a tick box, and noise removal has three strengths, from all of it to 10 dB down. Press the button again to make the file with your settings.

Does the enhancer work on music?

Not well. It is made for voices and treats instruments as noise, so part of the music is taken out. A voice over background music comes through with the music lower.

Which files can it enhance?

MP3, M4A, WAV, FLAC, OGG and AAC files, and the sound of MP4, MOV and WebM videos, up to 1 GB and 2 hours long. The result is saved as MP3, WAV or, in browsers that can write AAC, M4A.

How long does enhancing take?

The enhancer reads the recording twice: once to measure its loudness, once to make the file. On a Mac with Apple silicon, a minute of mono audio took 5 to 6 seconds in Chrome and 6 to 8 in Safari’s engine; a stereo recording takes about twice as long.

Clear recordings in, searchable notes out, on your Mac

Free to download. Runs entirely on your Mac.

macOS 26 or later · Apple Silicon (M1 or newer) · 16 GB memory minimum, 24 GB recommended

About NoteOwl

NoteOwl is private AI meeting notes and memory for the Mac. It captures meetings without a bot joining the call, labels every speaker, writes summaries and action items, and builds a memory of every conversation that you can chat with — fully offline, because the speech recognition and the AI run on the Mac itself. Dictation into any app is free; Pro is a one-time purchase.

NoteOwl pricing: what’s free and what Pro adds

Last updated