Convert MKV to M4A

People who convert an MKV for its audio are frequently not after the main mix. They want the director’s commentary, the audio description track that narrates what is on screen, the original language under a dub, or the interpreter channel from a recorded conference. Matroska is where all of that lives, because it is the container that stops nobody from adding a fourth soundtrack.

Every other audio conversion here answers that request by taking the first track and throwing the others away without saying so. That is the right behaviour for a format with room for one, and it is exactly wrong for this reader — which is why this page exists as something other than a second version of the MP3 page.

An M4A is not a codec, it is a container: AAC audio inside the same MPEG-4 box that an MP4 uses, with the extension changed so that software knows to expect sound and no picture. That box was designed for films, so it counts audio tracks in the millions rather than stopping at one. Ogg and Opus are built the same way and share that headroom, which is why all three sit on the "keeps everything" side of this page.

MP3, raw AAC, WAV and FLAC are all built the other way. Each of them is a single stream of audio with a header on the front, and there is nowhere in the specification to put a second one. It is not a limitation of this converter; it is what those formats are. When the requirement is "keep them all", the container has to come from the video world, and M4A is the smallest, most widely supported thing in that world that is still an audio file.

Every audio track the browser can decode is written into the M4A, in the order the MKV declared them. Nothing is reordered and nothing is promoted — including Matroska’s own "default track" flag, which is not consulted. A rip with an English mix, a commentary and a Spanish dub produces a three-track M4A in that same sequence.

A player with an audio menu is what turns that into something usable. VLC lists them under Audio › Audio Track on every platform it runs on, and mpv and Infuse do the same. Apple’s Music app is the awkward one, because its model is one stream per file and it will generally hand you the first — so the practical arrangement is an M4A as the file you keep and a capable player as the thing you open it with.

Even when the MKV already holds AAC, the audio is decoded back to samples and compressed again. There is no packet-copy path in this converter, so AAC to AAC costs a second generation of lossy compression rather than nothing. It is audible on music at the small-file band and effectively inaudible on speech at the balanced one.

For a commentary or a description track — speech, usually mono, recorded in a booth — the loss is not worth thinking about. For a concert mix it is, and the high band is the lever. If nothing at all should change, a desktop remuxer will lift the same AAC stream out of the Matroska container and into an MPEG-4 one in seconds without decoding a frame, and that is the correct tool for an archive copy.

A track is dropped before conversion if the browser cannot decode it, and the list of codecs that fall into that hole is short and predictable: AC-3, E-AC-3, DTS and Dolby TrueHD. Those are the soundtracks discs carry, and browsers do not ship decoders for them — DTS and TrueHD are not even recognised as codecs by the reader.

This lands hardest on precisely the files this page is for. A commentary track on a Blu-ray rip is very often AC-3 while the main feature is DTS-HD, so a file can come back with one track where you expected three, or refuse entirely because none of them could be read. If that happens, converting the audio to AAC once with a desktop tool and returning here is the shortest way through.

The band applies to all of them, so it has to suit the most demanding one. Small, balanced and high are offered, defaulting to balanced, and the band is resolved per codec rather than being a bitrate you have to guess at.

Balanced is right for a file that is mostly speech tracks. Go to high when one of the tracks is a music mix, because the setting is not per track and a band chosen for a commentary will be applied to the score as well. The size cost is real but modest against the video file the audio came out of — an hour of AAC at the high band is still a small fraction of a gigabyte.

Title, artist, album, date, description and comment are read out of the Matroska tags and written into the M4A, which is what the Music app, Podcasts and a file inspector read to show something other than a filename. Cover art carried in the source comes across the same way.

Chapter marks do not. Matroska chapters are one of the format’s selling points and this conversion does not read them, so an audiobook or a lecture series that was navigable inside the MKV becomes one long track in the M4A. Where that matters — and for a five-hour recording it matters a great deal — the chapters have to be rebuilt afterwards in a tool that writes them, or the file has to be split.

Inside the browser tab, on your own processor, with the decoders the machine already has. Nothing carrying the file is sent to a server, and the network tab will show you that in ten seconds if you want to check rather than take it on trust.

The free tier accepts an MKV up to 100 MB. That is generous for a recorded talk and restrictive for a film, which is the awkward shape of this pair: the files with the most interesting alternate tracks are usually the largest ones. For those, extracting the tracks with MKVToolNix on the desktop costs nothing in quality and no time at all, and this page is the browser-sized version of the same job.