Split a M4B audiobook into chapters.

m4b-splitter.sh

#!/bin/bash

# Description: Split a M4B audiobook into chapters.
# Requires: ffmpeg, ffprobe, jq, yq
# Author: Hasan Arous
# Improved by: Ewilan Rivière
# License: MIT
# https://unix.stackexchange.com/questions/499179/using-ffmpeg-to-split-an-audible-audio-book-into-chapters

# Usage: ./m4b-splitter.sh <input_file.m4b> [--no-convert | -n]
# Enable script execution: `chmod +x ./m4b-splitter.sh`
# Run the script: `./m4b-splitter.sh mybook.m4b`
# This script splits an M4B audiobook file into chapters, extracts metadata to YAML,
# and extracts the cover image if available. It can also convert chapters to MP3 format.

# To skip MP3 conversion and keep chapters as M4B:
# `./m4b-splitter.sh input.m4b --no-convert`
# or
# `./m4b-splitter.sh input.m4b -n`

# Based on `Audiobook Builder 2` audiobook construction https://apps.apple.com/fr/app/audiobook-builder-2/id1437681957

set -euo pipefail

show_help() {
  cat <<EOF

Split an audiobook file into chapters, export metadata.yaml, extract cover image.

Requirements:
  - ffmpeg, ffprobe, jq, yq

Options:
  --no-convert, -n   Skip conversion of chapters to MP3 (keep as M4B)

Usage:
  $(basename "$0") <input_file.m4b> [output_extension]

Example:
  $(basename "$0") mybook.m4b

EOF
}

if [[ $# -eq 0 || "$1" == "-h" || "$1" == "--help" ]]; then
  show_help
  exit 0
fi

input="$1"
input_base=$(basename "$input")
input_name="${input_base%.*}"

shift || true # Remove first arg (input file)

convert_mp3=true

# Parse optional flags
while (($#)); do
  case "$1" in
  --no-convert | -n)
    convert_mp3=false
    shift
    ;;
  *)
    echo "Unknown option: $1" >&2
    exit 1
    ;;
  esac
done

echo "Convert chapters to MP3: ${convert_mp3}"

for cmd in ffmpeg ffprobe jq yq; do
  if ! command -v "$cmd" &>/dev/null; then
    echo "Error: '$cmd' is not installed or not in PATH." >&2
    exit 1
  fi
done

output_dir="${input_name}_chapters"

# Delete output directory if it exists
if [[ -d "$output_dir" ]]; then
  echo "⚠️ Output directory '$output_dir' exists, deleting it..."
  rm -rf "$output_dir"
fi

mkdir -p "$output_dir"

# --- Extract chapters and split ---
chapters=$(ffprobe -v error -print_format json -show_chapters "$input" | jq -c '.chapters[]')

count=1
echo "📂 Splitting chapters into: $output_dir"
echo

ffprobe -v error -print_format json -show_chapters "$input" | jq -c '.chapters[]' | while read -r chapter; do
  start=$(echo "$chapter" | jq -r '.start_time')
  end=$(echo "$chapter" | jq -r '.end_time')
  title=$(echo "$chapter" | jq -r '.tags.title // "Chapter"')

  # safe_title=$(echo "$title" | tr -cd '[:alnum:] _-' | tr ' ' '_' | tr -s '_')
  safe_title=$title
  output_file="$output_dir/${count}_${safe_title}.m4b"

  echo "🎵 Chapter $count: $safe_title"
  ffmpeg -loglevel error -y -i "$input" -ss "$start" -to "$end" -c copy "$output_file"

  ((count++))
done

# --- Export all metadata to YAML (full) ---
echo "💾 Exporting full metadata to YAML..."
ffprobe -v error -print_format json -show_format -show_streams -show_chapters "$input" >"$output_dir/metadata.json"

# --- Extract tags only (format tags + chapter titles) for tags.yaml ---
echo "💾 Extracting tags to tags.yaml..."

# Extract general format tags as JSON object
format_tags_json=$(jq '.format.tags // {}' "$output_dir/metadata.json")

# Extract chapter titles array, key = Chapter_<number>, value = title or default
chapter_titles_json=$(jq '
  .chapters
  | map(
      { ("Chapter_" + (.id|tostring)): (.tags.title // "Untitled") }
    )
  | add
' "$output_dir/metadata.json")

# Merge format tags and chapter titles into one JSON object
merged_tags_json=$(jq -n --argjson f "$format_tags_json" --argjson c "$chapter_titles_json" '$f + $c')

# Convert merged JSON tags to YAML
echo "$merged_tags_json" | yq -P e - >"$output_dir/tags.yaml"

# --- Extract cover image (if any) ---
echo
echo "🖼️ Detecting cover image stream..."
cover_stream_index=$(jq -r '.streams[] | select(.disposition.attached_pic == 1) | .index' "$output_dir/metadata.json" | head -n1 || echo "")

if [[ -n "$cover_stream_index" ]]; then
  echo "Extracting cover image from stream $cover_stream_index to $output_dir/folder.jpg"
  ffmpeg -hide_banner -loglevel quiet -y -i "$input" -map "0:$cover_stream_index" -c copy "$output_dir/folder.jpg"
  echo "✅ Cover image extracted successfully."
else
  echo "⚠️ No cover image stream found, skipping cover extraction."
fi

echo
echo "Converting in parallel..."

# Conditional MP3 conversion
if [ "$convert_mp3" = true ]; then
  echo
  echo "🔄 Converting all chapter .m4b files to .mp3 in parallel (fast encoding) and deleting originals..."

  max_jobs=4
  job_count=0
  quality=2 # Adjust quality as needed (1 = best, 9 = worst)

  find "$output_dir" -name '*.m4b' | while IFS= read -r m4bfile; do
    (
      mp3file="${m4bfile%.m4b}.mp3"
      echo "Converting '$m4bfile' → '$mp3file'"
      ffmpeg -loglevel error -y -i "$m4bfile" -acodec libmp3lame -qscale:a "$quality" "$mp3file" && rm "$m4bfile"
    ) &

    ((job_count++))
    if ((job_count >= max_jobs)); then
      wait
      job_count=0
    fi
  done

  wait
  echo "✅ Conversion and cleanup complete."
else
  echo
  echo "⚠️ MP3 conversion skipped (--no-convert or -n). Output files are .m4b chapters."
fi

echo
echo "✅ Done! All files saved to: $output_dir"

Enable script execution:

sudo chmod +x ./m4b-splitter.sh

Usage:

./m4b-splitter.sh input.m4b

Usage without conversion to MP3:

./m4b-splitter.sh input.m4b --no-convert # or
./m4b-splitter.sh input.m4b -n