A method to watch a directory for newly added audiobooks, which triggers a script that converts the contents of that new folder into an m4b, and saves it to mp3tag's default directory ready for post processing.
I use multiple OSes for this particular automation. As such, this particular method may be impractical to some, but there's nothing special about these steps that requires multiple computers, and everything could be done solely on windows.
Please consider contributing a Windows Only method if you write a similar windows script.
This is how my system is set up. I have a fileserver/nas that uses SMB to share my folders with all my computers.
| Computer | OS | Noteable Installs |
|---|---|---|
| Fileserver VM | OMV (debian linux) | SMB network share, MergerFS+SnapRAID |
| Docker Server VM | Ubuntu 20.04 | M4b-tool docker and auto-m4b-tool.sh script |
| HTPC | Windows | Mp3tag, Dropit |
/original- Folder where I keep my untagged/unmodified original copies/temp/mp3merge- Folder where I copy recently added mp3 audiobook folders/temp/untagged- Folder where I copy m4b files from/original, also where I optput the m4b file created from/mp3merge. This is the folder I set mp3tag to open by default./temp/delete- Purely a temp folder, used as a lazy way to delete the mp3 audiobook folder copied to/mp3mergeafter conversion to m4b./audiobooks- Folder where I keep properly tagged and organized audiobooks. This is what Plex/Booksonic looks at.
- Newly acquired audiobooks are put in
/original - Auto Copy new books to appropriate folder based on filetype
- If book is already an m4b, then copy to
/untagged/Book1.m4b - If book is mp3, then copy to
/mp3merge/Book1/*.mp3
- If book is already an m4b, then copy to
- Every 5 min the
auto-m4b-tool.shscript checks/mp3mergefor new folders, when found creates a single chapterized M4b - This newly created m4b file is saved to
/untagged /mp3merge/Book1folder is moved to/deleteand the contents of/deleteis deleted- Open mp3tag, all books that need processing will be loaded
- Use mp3tag audible websource script to tag Keyboard Shortcut =
Ctrl+shift+i - Use mp3tag action script to rename/relocate to
/audiobooksKeyboard Shortcut =Alt+a 1
First let's prepare the Linux machine (Docker Server VM) . We will be installing the m4b-tool docker, configuring a docker run command, and creating the automation script.
Docker is by far the easiest way to install and use m4b-tool. Other methods will not be covered in this guide. Run the following 5 commands.
# Install FFMPEG
sudo apt install ffmpeg -y
# clone m4b-tool repository
git clone https://github.com/sandreas/m4b-tool.git
# change directory
cd m4b-tool
# build docker image - this will take a while
docker build . -t m4b-tool
# testing the command
docker run -it --rm -u $(id -u):$(id -g) -v "$(pwd)":/mnt m4b-tool --versionFor other methods of installing m4b-tool see https://github.com/sandreas/m4b-tool#installation
The docker run command is the heart of this operation. There are two sets of variables we need to define that correspond to a Docker portion and a m4b-tool portion of this command. The docker portion requires us to set the paths (-v volumes) that we will be working with. The m4b-tool portion will define how to encode and combine the mp3 files.
Example:
docker run -it --rm -u $(id -u):$(id -g) -v /path/to/temp/mp3merge:/mnt -v /path/to/temp/untagged:/untagged m4b-tool merge "$file" -n -q --audio-bitrate="$bit" --skip-cover --use-filenames-as-chapters --audio-codec=libfdk_aac --jobs=4 --output-file="$string3" --logfile="$string5"
Options Explained:
-v /path/to/temp/mp3merge:/mnt- MP3 Source folder (mapped to /mnt inside the docker)-v /path/to/temp/untagged:/untagged- M4B Destination folder (mapped to /untagged inside the docker)-n- No Interruptions-q- Quiet--audio-bitrate="$bit"- Bitrate used is based on the bitrate of the mp3 file--skip-cover- We will add a cover in mp3tag, no need to add it here it will be written over.--use-filenames-as-chapters- Depends on your source material, If your mp3 has chapter names (TITLE) tagged to your liking then do not use--audio-codec=libfdk_aac- High quality codec--jobs=4- How many CPU Cores to use, do not set higher than available- More m4b-tool options https://github.com/sandreas/m4b-tool#reference
All you need to do is update the two
-vpaths to your own
This is a Linux Bash script. You will need to update lines 23, 24, 25, and 31 of the script to your specific directories everywhere you see /path/to/....
#!/bin/bash
# set n to 1
n=1
# continue until $n 5
while [ $n -ge 0 ]
do
if ls -d */ 2> /dev/null
then
echo Folder Detected
string1="/untagged/"
string2=".m4b"
string4=".log"
for file in *; do
if [ -d "$file" ]; then
mpthree=$(find . -maxdepth 2 -type f -name "*.mp3" | head -n 1)
string3=$string1$file$string2
string5=$string1$file$string4
echo Sampling $mpthree
bit=$(ffprobe -hide_banner -loglevel 0 -of flat -i "$mpthree" -select_streams a -show_entries format=bit_rate -of default=noprint_wrappers=1:nokey=1)
echo Bitrate = $bit
echo The folder "$file" will be merged to "$string3"
echo Starting Conversion
docker run -it --rm -u $(id -u):$(id -g) -v /path/to/temp/mp3merge:/mnt -v /path/to/temp/untagged:/untagged m4b-tool merge "$file" -n -q --audio-bitrate="$bit" --skip-cover --use-filenames-as-chapters --audio-codec=libfdk_aac --jobs=4 --output-file="$string3" --logfile="$string5"
mv /path/to/temp/mp3merge/"$file" /path/to/temp/delete/
mv /path/to/temp/untagged/"$file".chapters.txt /path/to/temp/untagged/chapters
echo Finished Converting
echo Deleting duplicate mp3 audiobook folder
fi
done
else
rm -r /path/to/temp/delete/* 2> /dev/null
echo No folders detected, next run 5min...
sleep 5m
fi
done- Change to the
/mp3mergedirectory
cd /path/to/temp/mp3merge- Create auto-m4b-tool script
nano auto-m4b-tool.sh- Copy and paste the above code, and update lines 23, 24, 25, and 31 with your own paths everywhere you see
/path/to/...,- Save
ctrl-s - Exit
ctrl-x
- Save
- Change to the
/mp3mergedirectory
cd /path/to/temp/mp3merge - Run the script
./auto-m4b-tool.sh
This will run the script in a terminal window. To exit the script type ctrl-c. You can get fancy and run the script as a service, or set a cron job to start this script when the computer starts, but for my pursposes having an extra terminal window open with this running helps to see what's actually happening and keep tabs on the conversion.
So far we have a script that watches a directory (/mp3merge) for new folders, converts the contents of that new folder into an m4b, and saves it to mp3tag's default directory (/untagged) ready for post processing.
Now we need to configure the Windows portion of this process to monitor your /original folder for recent additions and copy mp3 book folders to /mp3merge. I use a program called Dropit to handle this.
Dropit is a very configurable, lightweight, windows utility, that will help us monitor /original for changes and copy the files and folders to the right locations.
- Download, Install, and Run Dropit
- In the System Tray: Right-Click
Dropit>Profiles>Customize
- Create a new Profile
- In the System Tray: Right-Click
Dropit>AssociationsCreate the two associations shown below with your specific folder
- The first entry moves .m4b files directly to
/untagged - The second entry moves .mp3 book folders to
/mp3merge/Book1where ourauto-m4b-tool.shscript will take over - Click the Check button to Save
- The first entry moves .m4b files directly to
- In the System Tray: Right-Click
Dropit>Options>MonitoringCheckEnable scan of monitored folders- Select
immediate on-changefrom the drop down menu - Update the path to your
/original - Select the
Mp3mergeprofile from the drop down menu - Click
SaveandOK
Return to Guide to configure Mp3tag. Make sure to set /temp/untagged as mp3tag's default folder.
MP3TAG: To verify what chapter titles were set, and if the chapter order is correct, Open Mp3tag, go to Tools > Options > Advanced > CHECK "List chapters as separate files". You can close mp3tag, then reopen, and you should see all of the chapters as if they were separate files.
WARNING: If you are using the option of seeing all chapters as files, saving the tags to file will take a very long time. Once you verified that every thing looks good, best to switch back to only seeing the m4b as a single file (UNCHECK, close, and reopen).
- This is only for newly aquired books, and does not address any mp3 books already tagged and organized in
/audiobooks. - This uses the original filenames as Chapter names, and the original mp3 files as chapters. Depending on your source material his may not be ideal but it works well enough.
- Would be nice if the chapter.txt file that is generated could automatically end up in the
/audiobooks/author/book/folder. - Would be nice for the script to compare the total length of the source mp3 files vs m4b, send an error message with the time difference.
- Let me know what else could make this better, and consider contributing! Thank you!
There are many ways to customize this workflow. I'd love to hear what you've come up with.
The script will sample one of the mp3 files and dynamically set the bitrate based on the source material. Eg. book1.mp3 is 64k then book1.m4b is 64k, book2.mp3 = 128k, book2.m4b = 128k, etc. The output AAC will be variable bitrate (VBR), no way to force CBR sandreas/m4b-tool#55
Huge shout-out to sandreas who created the amazing m4b-tool and to tylerdotdo for sharing the original auto-m4b-tool.sh script!
