Initializing... drag & drop files here
Supports: MP3
This walk-through is for anyone with a long spoken-word MP3 — an audiobook, a lecture series, a recorded course — that their player keeps treating as a music track and restarting from zero. .m4b is the MPEG-4 Audiobook extension, and audiobook players use it as a signal to file the recording as a book and remember where you stopped. The guide covers the upload-to-download flow, which of the bitrate controls on this page gives you a result you can predict, and the one thing a single MP3 can never give you no matter what settings you pick.
.m4b — the converter does not stitch several MP3s into one book..m4b. Files upload over an encrypted connection, are processed on our servers, and are deleted automatically after a few hours — no sign-up, no watermark.The output codec here is always AAC, encoded with the Fraunhofer FDK encoder — there is no codec choice to make. What matters is which rate-control mode you use, because they do not all behave the same way:
| Variable Bitrate band | Encoder quality level | What it suits |
|---|---|---|
| 20k-32k | 1 (lowest) | Thin but intelligible speech; only for very long books |
| 32k-40k | 2 | A sensible floor for mono narration |
| 48k-56k | 3 | Comfortable spoken word, the usual pick |
| 64k-72k | 4 | Books with score, effects, or dramatised casts |
| 96k-112k | 5 (highest) | More than any narration needs |
The ranges in those labels are inherited from the encoder's documentation, which describes them as per-channel targets — but that is documentation rather than behaviour. Measured through this pipeline the delivered rate barely moves with channel count, and the top band comes out above its printed range. Treat the five bands as level 1 through level 5, higher is better, and expect the material to move the result as well: a quiet single-voice reading encodes well below its band, a busy dramatisation near the top of it. When you need a rate you can plan around, use Constant Bitrate.
| Property | MP3 source | M4B output |
|---|---|---|
| Codec | MP3 (MPEG-1/2 Audio Layer III) | AAC, written by the FDK encoder |
| Container | MPEG audio stream | MPEG-4 (the ipod flavour, same family as M4A) |
| Filed as an audiobook by players | No — usually treated as a music track | Yes — the .m4b extension is the signal |
| Playback position remembered | Rarely | Yes, in players that honour the convention |
| Chapter markers | None in a plain MP3 | None either — see below |
| Title / artist / album / genre tags | Present as ID3 | Carried across into MP4 metadata atoms |
| Efficiency at the same bitrate | Baseline | Better — AAC holds more detail per kilobyte |
| Plays on basically anything | Yes | No — iOS and macOS natively, elsewhere via VLC or an audiobook app |
.m4b. VLC and dedicated audiobook apps handle it; those apps are also the ones that implement the resume behaviour you converted for.If the MP3 came from a DRM-protected store download, or the file is truncated or corrupt, the conversion will not complete — that is a property of the source, not a setting. If your book is already spread across dozens of MP3 files and you want one navigable volume with chapter names, this converter is the wrong tool: it produces one M4B per upload with no chapter authoring. And if the real goal is simply a smaller file rather than audiobook behaviour, staying on MP3 and re-encoding at a lower bitrate with compress MP3 avoids a codec change entirely.
That is the whole point of the format convention, and it works in players that implement it — audiobook apps, and Apple's ecosystem, treat a .m4b as a book rather than a song. What the extension cannot do is force a player that ignores the convention to behave. If your app still restarts from zero after the conversion, the app is the limiting factor, not the file.
No. Chapters are metadata that has to be authored, and a single MP3 has none to carry over. The output is one continuous AAC track in an MPEG-4 container with your title, artist, album and genre tags preserved. If you need named chapters or want to merge a multi-file book, use a dedicated audiobook-building application for that step.
Variable Bitrate, whose five bands map onto the encoder's own quality levels, or Constant Bitrate when you want an exact figure to plan a storage budget around. The Quality Preset dropdown displays "Highest" when the page loads, but nothing is pre-selected there — leave it untouched and the server falls back to its own default rather than the value you can see. The presets are also the least predictable control here, so reach for them only when you genuinely do not mind what rate comes out, and check the finished .m4b in your player's file-info panel either way.
Set Audio Channel to Mono and pick the 32k-40k or 48k-56k Variable Bitrate band. Narration is a single voice in a small frequency range, so it compresses far better than music — a mono book at that setting stays clear while keeping a twelve-hour recording to a sensible size. Reserve the higher bands for books with score, sound design, or multiple dramatised voices.
No. Both formats are lossy, so this is a transcode: the AAC encoder can only work from what the MP3 already kept. AAC is more efficient than MP3, which means a well-chosen bitrate loses very little — but the ceiling is your source. The reason to convert is the audiobook behaviour and the smaller file at equal quality, not fidelity.
Not internally. The file is an MPEG-4 audio container — the same family as M4A — written through FFmpeg's ipod muxer, and it identifies itself with an M4A major brand. What makes it an audiobook is the .m4b filename, which is the convention players key off. Renaming an M4B to .m4a produces a file that plays identically but gets filed as music.
Yes, with M4B to MP3. Be aware that this is a third lossy generation — MP3, then AAC, then MP3 again — so keep your original MP3 rather than round-tripping if you might need it. Converting back also drops the audiobook filing behaviour, since that came from the extension.
Files travel over an encrypted connection, are processed on our servers, and are deleted automatically after a few hours. There is no account requirement, no watermark, and nothing is shared or made public. A multi-hour MP3 is a large upload, so the wait is mostly transfer time rather than encoding time.