How to reduce audio file size

Audio2 min read

Four levers in order of effect, and the one that most people forget — which halves a spoken-word recording on its own.

In short

Trim what you do not need, downmix speech to mono, pick a bitrate that suits the material, and choose a lossy format. For a podcast, mono at 96 kbps is a fraction of the size with no audible cost.

Audio is small next to video, which is why people ignore it until an hour-long recording turns out to be 600MB.

In order of effect

  1. 01

    Trim the dead air

    Linear and free: remove a fifth of the length and remove a fifth of the file. The silence before someone starts talking and after they finish is pure weight. Use Trim Audio.

  2. 02

    Downmix speech to mono

    The forgotten one. A single voice has no meaningful stereo content, and many devices play one channel anyway. This halves the file before you change anything else.

  3. 03

    Pick a bitrate that matches the material

    96 kbps is generous for speech; music generally wants 192 kbps or more. Using a music bitrate for a voice note doubles the size for nothing.

  4. 04

    Use a lossy format

    If it is still a WAV, that is where the size is. MP3 or AAC is roughly a tenth.

Removing silence in the middle

For interviews and recorded meetings, the gaps between speakers add up. Shortening them tightens the recording and takes real size off — though do not remove them entirely, because speech with no pauses is exhausting to listen to.

Mistakes worth avoiding

  • Lowering the bitrate of an already-compressed file when you could have trimmed instead.
  • Downmixing music to mono. There is real stereo information in a mix; this is a speech technique.
  • Re-encoding from an MP3 when the original WAV is still available.