Extract the audio from a video

The audio inside an MP4 is almost always AAC, and an M4A is an MPEG-4 container holding exactly that. So choosing M4A copies the audio across without a decoder ever seeing it: no quality is lost, and it finishes in about a second whatever the length of the video.

MP3 and WAV cannot hold AAC, so those two decode the sound and encode it again. MP3 costs a generation of quality; WAV costs none but produces a file roughly ten times the size, because it stores the samples raw. Pick MP3 for a phone, WAV for editing, M4A when you simply want the sound as it was.

The picture is discarded, which is usually the entire point — a one-hour lecture drops from a gigabyte to about thirty megabytes with nothing lost that you were going to listen to.

Chapter marks, subtitle tracks and any second audio language do not come across. If a film has an English and a German track, this takes the first one. That is a limitation worth knowing before you use it on something dubbed.

Quinvert also has pages for specific pairs — MP4 to MP3, MOV to WAV and so on — and they do the same work. This page exists because the question arrives differently: somebody with a video does not always know or care what container it is in, and they want the sound out of it.

So this one takes any of the five common video containers and lets you choose the destination afterwards, where a pair page decides both for you. Neither is better; they answer different sentences.

The audio track is already a finished, compressed stream inside the video, and what comes out is that same stream in its own container. So the size follows from the bitrate the camera or the encoder chose: roughly 128 kbit/s is about one megabyte a minute, and 256 kbit/s is about two.

A one-hour lecture at a typical 128 kbit/s therefore lands near sixty megabytes, down from a video that may have been two gigabytes. That ratio is why extracting is worth doing at all for anything you intend to listen to on a phone rather than watch.

A film can carry several audio streams — a second language, a director's commentary, a descriptive track — and a screen recording can carry the microphone and the system sound as separate streams rather than mixed together. In all of those, “the audio” is not a single thing.

The first stream is taken, which is the main programme audio in nearly every file that has more than one. If what you got is a commentary or a silent microphone channel, the video has an unusual layout and needs a tool that lets you pick the stream by index — that is a real limitation of this page and worth knowing before you conclude the file is broken.

Technically this works on any video whose file you hold, including one downloaded from a streaming site. Whether you may keep the result is a separate question and it is not a technical one.

Your own recordings, lectures you were given, material licensed for reuse and anything in the public domain are yours to do this with. Commercial music and film are not, and no property of this tool changes that. Saying so is more useful than a disclaimer nobody reads, because the distinction is about the video and never about the software.