Organize audiobooks
This script provides tools for manipulating digital audiobooks. It accepts Audible (.aax) files, as well as generic audio files (mp3, m4b, wav, etc). By default it will concatenate input files into one output. It preserves embedded chapter metadata from single files, or generates chapter metadata based on file breaks if multiple input files are given.
It has the ability to de-DRM Audible books if you have your user's "activation bytes". I should point out here that I am not encouraging copyright infringement; this functionality is intended to provide a means of compressing, and archiving audiobooks only. I think it is important to have choice when it comes to how you listen to audiobooks and DRM infringes on that ability.
It can also transcribe audiobooks to plain text files with CMU's PocketSphinx speech-to-text engine. This is useful for finding your place in an audiobook if you lost your bookmark. The speech recognition is mediocre at best, but it does enable you to do crude searches through your audio files with embedded ~1 minute timestamps.
abtools
NAME
abtools -- Archive audio books from Audible or elsewhere
SYNOPSIS
Usage: abtools [ OPTIONS ] input_file(s)
OPTIONS
--no-encode Do not produce encoded audiobook.
--no-cover Do not download cover image.
--no-transcript Do not transcribe audiobook.
--key,-k KEY Audible key
--sphinx exe PocketSphinx executable
--hmm HMM Hidden Markov Model
--lm LM Language Model
EXAMPLES
abtools AudioBookUnabridged.aax
abtools --no-transcript AudioBook1_librivox.m4b
abtools ./book-chapter*.mp3
The script works out of the box, but these variables are useful.
Define them in ~/.config/abtools.rc or on the command line
(command line options override config file):
KEY='AUDIBLE KEY ("Actvation bytes")'
DEST='Default output directory' (defaults to $HOME)
POCKETSPHINX='PocketSphinx executable'
POCKETSPHINX_POSTPROCESS='pipe pocketsphinx output'
(Defaults to 'python ./pocketsphinx_filter.py')
HMM='Hidden Markov Model'
LM='Language Model'
The script requires python, curl, and ffmpeg (compiled with --enable-libopus, as is standard in many package managers) to be in your PATH. All other dependencies should be standard on any POSIX system.
Optional:
For transcribing you need CMU's PocketSphinx.
https://github.com/cmusphinx/pocketsphinx.git
PocketSphinx depends on SphinxBase: https://github.com/cmusphinx/sphinxbase.git
It also requires a Hidden Markov Model, and a Language Model, which you can find here.
$ wget https://archive.org/download/walden_librivox/WaldenPart1_librivox.m4b
$ abtools.sh WaldenPart1_librivox.m4b
---------------------------
;FFMETADATA1
title=Walden
album=Walden
artist=thoreau_henry_david
album_artist=thoreau_henry_david
genre=Audiobook
CHAPTER01=00:00:0.000
CHAPTER01NAME=walden_c01 part 1
CHAPTER02=00:30:16.008
CHAPTER02NAME=walden_c01 part 2
CHAPTER03=01:08:52.003
CHAPTER03NAME=walden_c01 part 3
CHAPTER04=02:07:50.020
CHAPTER04NAME=walden_c01 part 4
CHAPTER05=02:54:16.017
CHAPTER05NAME=walden_c01 part 5
CHAPTER06=03:17:42.020
CHAPTER06NAME=walden_c02 part 1
CHAPTER07=03:45:1.019
CHAPTER07NAME=walden_c02 part 2
CHAPTER08=04:10:51.022
CHAPTER08NAME=walden_c03
CHAPTER09=04:46:45.006
CHAPTER09NAME=walden_c04
CHAPTER10=05:37:9.007
CHAPTER10NAME=walden_c05
CHAPTER11=06:09:7.004
CHAPTER11NAME=walden_c06
CHAPTER12=06:42:33.003
CHAPTER12NAME=walden_c07
---------------------------
Edit this metadata? [y/N] Y [ENTER]
[EDIT METADATA AS A FILE]
---------------------------
;FFMETADATA1
title=Walden
album=Walden
artist=Henry David Thoreau
album_artist=Henry David Thoreau
genre=Audiobook
CHAPTER01=00:00:0.000
CHAPTER01NAME=Chapter 01 Part 1
CHAPTER02=00:30:16.008
CHAPTER02NAME=Chapter 01 Part 2
CHAPTER03=01:08:52.003
CHAPTER03NAME=Chapter 01 Part 3
CHAPTER04=02:07:50.020
CHAPTER04NAME=Chapter 01 Part 4
CHAPTER05=02:54:16.017
CHAPTER05NAME=Chapter 01 Part 5
CHAPTER06=03:17:42.020
CHAPTER06NAME=Chapter 02 Part 1
CHAPTER07=03:45:1.019
CHAPTER07NAME=Chapter 02 Part 2
CHAPTER08=04:10:51.022
CHAPTER08NAME=Chapter 03
CHAPTER09=04:46:45.006
CHAPTER09NAME=Chapter 04
CHAPTER10=05:37:9.007
CHAPTER10NAME=Chapter 05
CHAPTER11=06:09:7.004
CHAPTER11NAME=Chapter 06
CHAPTER12=06:42:33.003
CHAPTER12NAME=Chapter 07
---------------------------
Edit this metadata? [y/N] N [ENTER]
Outputs:
/home/user/Audiobooks/Henry David Thoreau - Walden/Walden.ogg
/home/user/Audiobooks/Henry David Thoreau - Walden/cover.jpg
/home/user/Audiobooks/Henry David Thoreau - Walden/transcript.txt
Modify outputs ("No" will initiate processing) [y/N] N [ENTER]
Cover URL: http://archive.org/download/walden_librivox/Walden_1105.jpg
... diagnostics ...