Extract Audio Channels

Take a recording apart into one mono file per channel — or keep only the side you need.

Runs on your deviceAudioWorkingNo account needed

What this tool does

Extract Audio Channels decodes your file into raw samples and writes each channel out on its own. A stereo file becomes two mono files, named after the original with -left and -right on the end. A six-channel file becomes six, labelled left, right, centre, lfe, surround-left and surround-right in the order the channels appear, and anything past the sixth is named channel-7 upwards. A channel is copied out sample for sample rather than recalculated, but the file it lands in is still written to the output settings you chose — WAV at 16-bit until you raise the depth, so a 24-bit or float source is requantised on the way out.

Four choices sit behind the Extract setting. Every channel as its own file is the default and the stereo split most people arrive for. Left only and Right only write a single file and ignore the rest, which is the short route when a recorder put one microphone on each side. One mono mix of all channels does the opposite: it averages the channels together, so material identical on both sides keeps its level while anything out of phase between them cancels instead of adding. That is arithmetic on the samples, not a repair of the recording.

What it will not do is worth saying before you start. It separates channels, not sounds: there is no vocal removal here, no instrument isolation, and splitting a finished stereo mix hands you two files that each still contain the whole song. A multi-channel file can be taken apart in full, or have its first or its second channel pulled out one run at a time, but there is no picker for a single named channel, so pulling out the centre of a 5.1 file means writing all six and deleting five. Ask a mono file for every channel, or for its right channel, and it is refused with an error saying there is only one; ask the same file for Left only or for the mono mix and it hands back a copy of the channel it already had under a new name.

The work happens in this browser tab and your audio is never uploaded. WAV files are read by a RIFF reader built into the page, which handles uncompressed PCM and float WAVs at 8, 16, 24 and 32 bits and passes compressed ones on to the browser rather than decoding them itself; everything else goes straight to your browser's own audio decoder, so MP3, M4A and FLAC usually work, OGG and Opus vary between browsers, AIFF opens only in Safari and WMA opens nowhere. Output is WAV or MP3 and nothing else, since no browser exposes an AAC, FLAC or OGG encoder to a page — and asking for MP3 fetches the encoder from this site the first time you use it, so that one path wants a connection even though the audio itself never crosses it. The whole file also sits in memory as 32-bit floats while it is split — more than a gigabyte for an hour of 44.1 kHz stereo, and three times that across six channels.

How to use it

  1. Add the audio file. MP3, WAV, M4A, AAC, FLAC, OGG and Opus are accepted, as far as your browser's decoder reaches. Queue several and each is split in turn, with every output collected into one list.
  2. Choose what to extract. Every channel as its own file, Left only, Right only, or one mono mix of all channels. Left and Right read the first and the second channel of the file, so on a surround recording each setting hands back one member of the front pair.
  3. Pick the output format. WAV is the default and writes 16, 24 or 32-bit float, staying at 16-bit until you change it. MP3 encodes each channel with the encoder loaded in the page at the bitrate you choose, resampling to the nearest rate MP3 allows if the source sits between them, and Match the input where possible keeps an MP3 as MP3 while writing everything else as WAV.
  4. Download the results. Each channel arrives as its own file, named with the channel on the end. Two or more give a Download all as ZIP button, and the summary states how many channels went in and how many files came out.

Processing & privacy

Runs on your device. Your file is processed directly in this browser.

Common use cases

An interview recorded on two microphones

Field recorders routinely put each lapel mic on its own side of the stereo pair. Splitting the file gives one track per speaker, so you can gate, level or cut them separately.

Feeding a tool that only accepts mono

Speech recognition and some analysis software refuse anything with two channels. The mono mix folds the recording into one track rather than leaving a second channel to be ignored.

Finding out which side of a recording failed

A dead microphone or a faulty cable hides inside a combined waveform. Two separate files make it obvious in seconds which channel carries the fault.

Taking apart a multi-channel bounce

A four- or six-channel file that carries separate sources rather than a surround mix comes back as individual tracks, each named for its position in the file.

Questions

Is my audio uploaded?
No. Decoding, splitting and encoding all happen inside this browser tab, so the file is never transmitted. The only thing that touches the network is the MP3 encoder, which this site sends to the page the first time you ask for MP3 output; a WAV split needs nothing beyond the page already in front of you.
Can it remove vocals or isolate an instrument?
No. It copies whole channels out of the file. Vocal removal works by cancelling one channel against the other, and instrument separation needs a model of the sound; neither happens here, and a stereo split leaves both files carrying the full mix.
Why was my mono file refused?
Because there is nothing to take apart. A single-channel recording has no left and right, so asking it for every channel, or for the right channel, is rejected with an error saying it has only one. Left only and the mono mix are not blocked in the same way — both simply write out the single channel it already has under a new name.
Which formats can it write?
WAV or MP3 only, because browsers ship no AAC, FLAC or OGG encoder. WAV is the default and adds no lossy re-encoding, though its 16-bit default depth does requantise a 24-bit or float source unless you raise it. MP3 re-encodes every channel, which is a lossy pass on top of whatever the source has already been through, so keep it for files you are sending on rather than editing further.
Why are the split channels at a different sample rate to the source?
A WAV goes through this page's own reader and keeps the rate it had. Other formats go through the browser's audio decoder, which returns samples at the rate its audio engine runs at — commonly 48 kHz — so a 44.1 kHz MP3 can come back as 48 kHz files. Choosing MP3 output can shift it once more, since the encoder accepts only a fixed list of rates and moves anything else to the nearest. The result panel reports the rate the file was decoded at, which is the rate a WAV output is written at.
What are the channels named?
By position: left, right, centre, lfe, surround-left, surround-right, then channel-7 upward. The names come from the order the channels appear in, not from anything read out of the file, so a recording that stores its channels unconventionally is still labelled by position.
Do titles and cover art survive?
No. The output holds audio and nothing else, so ID3 tags and artwork are dropped and a player falls back to the filename. Audio Metadata will show you what the original carries if you want to put it back.
Can I split several files at once?
Yes. Each file is decoded and split in turn with the same setting, and the outputs gather into one list with a ZIP button. A long queue takes time rather than running in parallel, since every file is decoded in full before it is written out.

Related tools