Audio Transcriptor
Start

Make an audio file smaller

Recordings of people talking are almost always ten times bigger than they need to be. Drop yours and see what it weighs at the right bitrate.

  1. 1File
  2. 2Settings
  3. 3Processing
  4. 4Download
  • MP3
  • WAV
  • M4A
  • AAC
  • OGG
  • FLAC
  • OPUS

What this tool does

A voice recorder app, a WAV export or a phone memo can easily produce 60 MB for twenty minutes of someone talking. That size isn't quality, it's a default: those files are usually recorded in stereo, at a bitrate meant for music, capturing a range of frequencies a human voice never reaches. Bring both numbers down to what speech actually needs and the same recording sits at three or four megabytes with nothing audibly different about it.

The two dials that matter are here. The bitrate is how much detail each second gets to keep. Mono is the bigger one and the one nobody thinks of: an interview recorded in stereo is the same voice written twice, so folding it into one channel halves the file before the bitrate does anything at all.

What we handle that others don't

  • Mono in one click, which is the single biggest saving on any recording of people talking and the one every other compressor buries.
  • The before and after side by side, so you can try 96, look at the number, and try 64 if you feel like it.
  • It takes uncompressed WAV, which is where the real savings are: a WAV usually drops by 95%.
  • The file is never uploaded, so a two-hour meeting recording doesn't have to leave your laptop to be made smaller.

How it works

  1. 1

    Drop the audio

    MP3, WAV, M4A, FLAC, OGG — anything, including the giant WAV your recorder made.

  2. 2

    Choose how small

    96 kbps for speech, 128 if there's music in there. Mono halves it again.

  3. 3

    Download it

    You'll see the before and after in megabytes, and how much you saved.

In detail

What bitrate is, and why speech needs so little of it

Bitrate is how much data gets stored for each second of sound. A CD spends 1,411 kbps because it has to carry all of the music, top to bottom, from 20 Hz to 20,000. A human voice doesn't come close to either end: it lives between roughly 80 and 3,500 Hz, which is exactly why a telephone is perfectly intelligible on a fraction of that data.

So a recording of people talking at 96 kbps sounds the same as at 320 to anyone listening, and weighs a third as much. Below that it starts to show: at 48 kbps the voice turns metallic, though it stays understandable. With music underneath, the floor rises, and there it's worth staying at 128 or above.

The mistake of recording an interview in stereo

Almost every recorder ships set to stereo, because they're built with music in mind. In an interview, a meeting or a lecture, that means storing exactly the same sound twice: both channels carry the same thing, or near enough, and the file is twice the size for nothing.

Folding it to mono is by far the biggest saving available on a spoken recording, and it's the one no other compressor puts in front of you. It's a single switch, it works on anything already recorded, and it changes nothing you can hear — there was nothing different between the two channels to lose.

Questions

And if what you have is the video?

Upload it and you'll get the transcript with timestamps, the speakers separated by name and a summary with the key points. You can start without an account.

Transcribe my recordingSee prices

Related tools