Skip to content
epub2audiobook Convert

M4B or MP3: which audiobook format to pick

Both hold the same audio. The difference is what the file remembers, and that turns out to matter more than anything about sound quality.

What M4B does that MP3 cannot

An M4B is one file with a table of chapters inside it, plus a flag saying “this is an audiobook, not music”. Those two things buy you:

  • Chapter navigation. Skip buttons move a chapter at a time instead of fifteen seconds, which is what you want when you lose the thread and need to go back.
  • Position memory. Players treat audiobooks differently from songs. Close the app, come back a week later, and it starts where you stopped.
  • A tidy library. One entry, one cover, one progress bar.

A folder of MP3s can carry a cover and track numbers, but nothing in the format says the files belong together as a book, so each player invents its own approach and many of them forget.

When MP3 is still the right answer

Compatibility, mostly.

  • Car stereos reading from a USB stick. Most will refuse M4B outright.
  • Cheap or old players, including gym and running players.
  • Media servers you have already set up around per-file libraries.
  • Anywhere you want to edit or re-order the audio afterwards.

If your player takes M4B, take the M4B. MP3 is the fallback, not the upgrade.

What about M4A, AAC and MP4?

Mostly the same thing wearing different names. The audio inside an M4B is AAC, in an MP4 container. Rename an M4A to M4B and most players will start treating it as a book — the extension is doing a lot of the work.

There is one thing the extension does not do, which is set the media type flag inside the file. Apple Books reads that flag to decide whether a file belongs under Audiobooks or in your music library. We set it, which is why a file from here lands in the right place.

Bitrate

Speech does not need much. 64 kbps mono is the sensible target and what we produce: about 29 MB an hour, so a ten-hour book is near 290 MB.

Going higher does not make a synthetic narrator sound better, because the limit is the model rather than the encoding. Going much lower starts to hurt sibilants — the “s” sounds go glassy — which is tiring over a long book.

Two conventions, one file

There is an awkward historical detail worth knowing about if you ever build these yourself: there are two competing ways to store chapter marks in an MP4, and players disagree about which to read. Apple Books and CarPlay read one. A lot of Android and desktop software reads the other.

Writing only one gets you a file that works perfectly on your machine and shows no chapters on somebody else’s. We write both.

← All guides