Tame the peaks in a voice recording so the quiet words come up

Level Compressor — Even Out Loud and Quiet, Online Free

A broadband compressor that tames jumps in level — the peak of the file stays where it was.

Drop an audio file, or click to browse

MP3, WAV, M4A, OGG, FLAC, and more

Processed locally — your file never leaves this tab

Free, no signup No watermarks

A voice memo that shouts one sentence and whispers the next needs compression, not a volume knob. This page runs a single broadband compressor: loud moments get turned down, quiet moments come up, then the export is scaled so the peak matches the file you started with. It all happens in your browser. The file is never uploaded.

Jobs a compressor finishes

  • Even out a voice memo where the speaker leans toward and away from the phone.
  • Tame a podcast guest who is much louder than the host, without raising the noise floor by hand.
  • Sit a vocal in front of a track when the chorus jumps out of the verse.

How to use the Level Compressor

  1. Drop your audio

    Any MP3, WAV, M4A, OGG or FLAC file. Decoding stays on your device.

  2. Set the amount

    Drag Compress. 30–50% is enough for most speech. 100% is a heavy squash: threshold −36 dB, ratio 8:1.

  3. Preview and export

    Play the result, then export MP3 or WAV. The peak is put back to the original peak, capped at just under full scale.

Dynamics, not file size

The level compressor is one broadband compressor. Threshold runs from −18 dB down to −36 dB, and the ratio from just over 1:1 up to 8:1. Loud moments come down, quiet moments come up relative to them.

There is no makeup gain inside the compressor. Afterward the export is scaled so its peak matches the source peak, or −0.2 dBFS if the source was already hotter than that. It does not shrink the file. A smaller MP3 is the audio compressor page.

What it will and will not fix

It narrows the gap between loud and quiet. It is one band, not a vocal chain:

  • A voice memo that shouts one sentence and whispers the next.
  • A mix where the loud hits bury the quiet phrases, and you want them closer.
  • Not a noise remover. Lifting the quiet parts also lifts hiss and room tone.

How it differs from a volume knob

The volume booster multiplies the whole file. Normalize raises everything until the loudest sample hits a peak. Both leave the gap between loud and quiet the same. This page narrows that gap.

WAV keeps the level change without another lossy generation. MP3 re-encodes once. If the problem is only that the file is quiet, and the dynamics are fine, use the volume booster instead.

Level Compressor FAQ

Is my audio uploaded?

No. The compressor is a Web Audio dynamics processor inside your browser tab. After the page loads you can go offline and it still exports.

What happens to the audio?

One broadband compressor (threshold from −18 dB down to −36 dB, ratio from just over 1:1 up to 8:1). There is no automatic makeup inside the compressor, so the page then scales the export until its peak matches the source peak, or −0.2 dBFS if the source was hotter than that. WAV keeps that level change clean. MP3 re-encodes once.

Will this shrink the file?

No. Shrinking a file means re-encoding at a lower bitrate, which is the audio compressor page. This page changes dynamics and leaves the bitrate choice to the export format.

How is this different from the volume booster and normalize?

The volume booster multiplies the whole file. Normalize raises everything until the loudest sample hits a target. Both leave the gap between loud and quiet the same. This page narrows that gap. It is one band, not a multiband vocal chain, and it will not remove noise.

Can it make a quiet room tone disappear?

No. Lifting quiet parts also lifts hiss and room tone. If the noise is the problem, compression makes it more obvious. Cut the silence out or record again.

Does it work on a phone?

Yes. The slider and export work with touch in current Chrome, Edge, Firefox and Safari. A few minutes of speech exports in a second or two on a recent phone.