Extract audio from video

How to Extract the audio

1. 1

   Add your video

   MP4, MOV, MKV, AVI, WebM and eight more container types, up to 300 MB each.

2. 2

   Choose a format

   MP3 plays everywhere, AAC is smaller at the same quality, WAV is uncompressed for editing.

3. 3

   Click Extract and download

   If the track is already in the format you asked for, it is copied out untouched.

Often there is nothing to re-encode

A video file does not contain sound woven into the picture. It contains a separate audio track sitting alongside it, already encoded. When the format you ask for matches what is already there, that track is copied out byte for byte instead of being decoded and encoded again. This is not a small optimisation. Re-encoding costs a second generation of loss and minutes of processing; copying costs neither and finishes in seconds. It happens more often than you would expect: an AVI from the 2000s almost always carries MP3 audio, so asking for MP3 copies it straight out. An MP4 or MOV carries AAC, so asking for AAC does the same. The page tells you when this happened, because it changes what the result is worth.

Choosing between MP3, AAC and WAV

MP3 is the answer when the file is going somewhere unpredictable: a car stereo, an old player, a piece of software you do not control. AAC is technically better at the same bitrate and is what most videos already contain, so choosing it is frequently both the higher-quality and the faster option. WAV is uncompressed and is what you want if the audio is going into an editor: it decodes once, and everything you do to it afterwards starts from unquantised samples rather than from something already compressed twice.

Videos with more than one audio track

Films and recordings often carry several: an original language, a dub, a commentary, sometimes a descriptive audio track. The track number selects which one comes out, counting from one in the order they appear in the container. The details of the track that was used, its language tag, its title where the container carries one, its codec and channel layout, are reported with the result, so a wrong guess is obvious immediately rather than after listening to ten minutes of the wrong language.

When there is no audio at all

Screen recordings made with the microphone off, exports from editors where the audio was muted, and footage from cameras without a microphone all produce videos with no audio track whatsoever. There is nothing to extract, and the page says so plainly rather than handing back an empty file, and links to the tool that adds a soundtrack, which is usually what was wanted.

Frequently asked questions