What M4B Has That MP3 Doesn't
M4B is the audiobook file format. Technically it's an MPEG-4 audio file with a specific extension, but the format adds three things that matter for audiobooks:
- Chapter markers with timestamps and titles
- Resume-on-reopen behavior (the player remembers where you stopped)
- Embedded cover art that displays in the player
MP3 can have ID3 chapter frames, but most players ignore them. M4B's chapter system is universally supported by every audiobook app from Audible to Apple Books to Bookplayer.
If you're producing an audiobook for self-publishing or ACX (Audible's submission platform), M4B is the deliverable format. This post covers the encoding pipeline, chapter authoring, and ACX submission requirements.
For the underlying audio production workflow, see Logic Pro Bounce Settings or Audacity Export Settings.
ACX Audio Specs
ACX (Audiobook Creation Exchange) is the dominant self-publishing audiobook platform. Their specs are the de facto industry standard:
| Spec | Value |
|---|---|
| Format | MP3 (CBR) for upload, ACX converts to AAC for delivery |
| Bit rate | 192 kbps CBR |
| Sample rate | 44.1 kHz |
| Channels | Mono |
| Bit depth | 16-bit |
| Peak level | -3 dB or below |
| RMS level | -23 dB to -18 dB |
| Noise floor | -60 dB or lower |
| Track structure | One MP3 per chapter (not one for the whole book) |
| Track length | Each MP3 between 0:30 and 120:00 |
| Lead-in silence | 0.5 to 1.0 second |
| Lead-out silence | 1.0 to 5.0 seconds |
| Cover art | Square JPG/PNG, 2400x2400 minimum |
| ID3 tags | Title, artist, album, track number, year |
Even though ACX wants MP3 upload (not M4B), the chapter-per-file structure mirrors M4B's chapter markers. ACX combines them into the M4B-style audiobook on their backend.
For non-ACX distribution (Findaway Voices, AuthorsRepublic, direct sales), M4B with embedded chapters is the correct deliverable.
Building an M4B File
The pipeline from individual chapter MP3s or WAVs to a single M4B file:
Step 1: Render or assemble each chapter as a separate audio file
In your DAW, bounce each chapter as a separate WAV or MP3. Name them by chapter number for sortability:
01_introduction.wav
02_chapter_one.wav
03_chapter_two.wav
...
Step 2: Concatenate to a single audio file
Use FFmpeg to combine while tracking chapter boundaries:
# Create a file list
ls -1v *.wav > files.txt
sed -i "s|^|file '|;s|$|'|" files.txt
# Concatenate
ffmpeg -f concat -safe 0 -i files.txt -c copy combined.wav
Step 3: Generate a chapters file
Calculate cumulative time at each chapter boundary:
import subprocess
import os
chapters = []
current_time_ms = 0
for i, filename in enumerate(sorted(os.listdir("."))):
if filename.endswith(".wav") and not filename == "combined.wav":
# Get duration
result = subprocess.run([
"ffprobe", "-i", filename, "-show_entries", "format=duration",
"-v", "quiet", "-of", "csv=p=0"
], capture_output=True, text=True)
duration_ms = int(float(result.stdout.strip()) * 1000)
title = filename.replace(".wav", "").replace("_", " ").title()
chapters.append({
"start_ms": current_time_ms,
"end_ms": current_time_ms + duration_ms,
"title": title
})
current_time_ms += duration_ms
# Write chapters in FFMETADATA format
with open("chapters.txt", "w") as f:
f.write(";FFMETADATA1\n")
for ch in chapters:
f.write(f"\n[CHAPTER]\nTIMEBASE=1/1000\n")
f.write(f"START={ch['start_ms']}\n")
f.write(f"END={ch['end_ms']}\n")
f.write(f"title={ch['title']}\n")
Step 4: Encode to M4B with embedded chapters
ffmpeg -i combined.wav -i chapters.txt \
-map_metadata 1 \
-c:a aac -b:a 64k \
-movflags +faststart \
-metadata title="Book Title" \
-metadata artist="Author Name" \
-metadata album="Book Title" \
-metadata genre="Audiobook" \
-metadata date="2026" \
output.m4b
The -map_metadata 1 flag pulls chapter info from the second input (chapters.txt). The result is an M4B with proper chapter navigation.
For embedding cover art:
ffmpeg -i input.m4b -i cover.jpg \
-map 0 -map 1 \
-c copy -c:v:1 mjpeg \
-disposition:v:1 attached_pic \
output_with_cover.m4b
Audacity Workflow
For producers who don't want command-line work, Audacity can produce M4B-compatible output:
- Import all chapter audio
- Tools > Macros > Build the audiobook with chapter labels at boundaries
- File > Export > Export Multiple > One file per label
- Use a separate tool (like MP3Tag or m4baker) to combine into M4B
Pure Audacity doesn't ship M4B encoder; you need either FFmpeg or a dedicated tool for the final container step.
ACX Mastering Targets
For ACX submission, the audio needs:
- Peak: -3 dBFS or below (not -1 dBTP, the older spec)
- RMS: -23 dB to -18 dB
- Noise floor: -60 dB or lower
These targets are different from streaming music. ACX wants conservative levels with consistent volume across chapters.
In your DAW master bus:
- Linear Phase EQ (gentle high-pass at 80 Hz to remove rumble)
- De-Esser (tame sibilance)
- Compressor (3-4 dB GR on busy phrases, fast attack)
- Limiter (-3 dBFS ceiling)
- Loudness meter showing RMS in -23 to -18 dB range
For voice quality issues (mouth noise, breaths, plosives), iZotope RX is the industry standard. Less expensive: Adobe's built-in audio tools or Audacity's Noise Reduction effect.
Pro Tip: ACX's QA team listens with calibrated headphones at fixed volume. Mastering at -23 dB RMS sounds "right" on their system. Mastering at -16 dB RMS (typical music) sounds "shouting" and gets rejected.
Chapter Title Conventions
Chapter titles in M4B should be:
- Short (under 50 characters)
- Descriptive without spoilers
- Numbered consistently
Example pattern:
Introduction
Chapter 1: The Setup
Chapter 2: First Conflict
...
Chapter 12: Resolution
Acknowledgments
About the Author
Avoid character names that change later. Avoid plot details. Player apps display these prominently.
Cover Art Specs
Audiobook cover art is square (different from book covers, which are usually portrait):
| Spec | Value |
|---|---|
| Format | JPG or PNG |
| Dimensions | 2400×2400 minimum, 3000×3000 ideal |
| Aspect | 1:1 square |
| Color profile | sRGB |
| Maximum file size | 4 MB |
Embed in the M4B (see FFmpeg step above) and submit separately to ACX.
For converting book cover (portrait) to audiobook cover (square): re-design rather than crop. Cropping removes important parts of the cover. The audiobook cover is its own design.
For image conversion, see our image converter.
Common Issues
ACX rejects with "noise floor too high": room noise audible in pauses. Use noise reduction at moderate strength on the master bus, or re-record in a quieter environment.
Chapter markers don't show in iPhone Books: Apple Books reads M4B chapters from a specific embedding format. Verify with ffprobe -show_chapters input.m4b that chapters are present.
File too large: 64 kbps AAC is the production target for spoken word. 128 kbps is overkill (no audible benefit). Monitor file size; a 10-hour audiobook should be 280-300 MB.
Volume varies between chapters: chapters mastered separately at different times. Render all chapters in one mastering session, or apply the same chain to each.
Audible playback drifts out of sync: rare encoding bug with VBR M4B. Use CBR (constant bitrate) for audiobook delivery to avoid.
For background on audio bitrate, see Audio Bitrate Quality Guide.
Frequently Asked Questions
Why M4B instead of MP3?
M4B is recognized as an audiobook by player apps. They apply audiobook-specific behaviors (resume position, chapter navigation, sleep timer integration). MP3 doesn't get these treatments.
Can I convert MP3 to M4B?
Yes, with FFmpeg as shown above. The chapter markers need to be added separately; MP3 doesn't carry M4B-style chapter metadata in most cases.
What's the bitrate sweet spot for audiobooks?
64 kbps mono AAC is the production default. 96 kbps is acceptable for productions with significant non-voice elements (sound effects, music interludes). 128 kbps is overkill for pure narration.
Can I distribute outside Audible/ACX?
Yes. Findaway Voices distributes to libraries (OverDrive, Hoopla). AuthorsRepublic distributes to multiple platforms. Direct sales via Bookfunnel or your own website use M4B files directly.
Should I deliver the multi-file MP3 set or single M4B?
ACX wants multi-file MP3. Other platforms vary; some accept M4B, some require multi-file MP3. Check each platform's submission specs.
How long can a single M4B file be?
Technical limit is 4 GB (32-bit file size). At 64 kbps, that's about 132 hours. For practical purposes: keep files under 12 hours each. Longer files struggle with player seek performance.
Related Reading
Bottom Line
For audiobook M4B delivery: encode at 64 kbps mono AAC, master to -23 to -18 dB RMS with -3 dBFS peak, embed chapters via FFmpeg with FFMETADATA format, attach square cover art at 2400×2400, deliver multi-file MP3 to ACX or single-file M4B elsewhere. Our audio converter handles MP3-to-M4B conversion if your starting point is multi-file MP3 chapters.



