navidrome is a subsonic compatible musing streaming service that relies on metadata to catalog a music library
Manage metadata #
In order to work properly navidrome needs that tracks have metadata populated, to compile them there are a pletora of software, one of them being picard
Collection file structure #
To keep things ordered and clean files inside the collection folder are organized as follows
collection/
├── album 1
│ └── track 1
├── album 2
│ ├── track 1
│ └── track 2
├── playlists
│ ├── ....nsp
│ ├── ....nsp
│ └── ....nsp
.....
In order to divide tracks in album folder run the following oneliners
[!TIP] picard can move files based on metadata values inside folders so this 2 snippets aren’t needed
this will create folder based on ALBUM metadata
find . -type f | parallel 'ffmpeg -i {} -f ffmetadata' 2>&1 | grep 'ALBUM ' |awk -F':' '{$1=""; print $0}' | while read dir; do
if [[ -d "$dir" ]]; then
echo "$dir already exists"
else
mkdir -p "$dir";
fi
done
this will move tracks in ALBUM folder according to metadata
find . -type f | while read f; do ALBUM="$(ffmpeg -i "$f" -f ffmetadata 2>&1 | grep 'ALBUM ' | awk -F':' '{$1="";print $0}' | awk '{$1=$1;print}')";
if [[ -f "$ALBUM/$f" ]]; then
echo "$ALBUM/$f already exists";
else
mv "$f" "$ALBUM/";
fi
done
Creating smart playlists #
Smart playlists are ways of grouping tracks together based on metadata values, they are defined inside json files with the .nsp extension
file
last_played.nsp
{
"name": "Recently Played",
"comment": "Recently played tracks",
"all": [
{"inTheLast": {"lastPlayed": 30}}
],
"sort": "lastPlayed",
"order": "desc",
"limit": 100
}
This json files are composed by a sequence of statements in the form
{"operator": {"field": "value"}}
The tracks with metadata that matches the statement are included inside the smart playlist, here a complete list of fields and operators
Backup physical collection #
To backup the physical collection the process requires to acquire the audio tracks of a cd in a file format (typically wav) and then edit metadata, under archlinux the utility cdda2wav can do the job
from the arch wiki
cdda2wav -vall cddb=-1 speed=4 -paranoia paraopts=proof -B -D /dev/sr0
[!WARNING] Notes that paranoia parameters make the process slower, for cd that are in good conditions they can be omitted
An handy alias to avoid memorizing parameters
alias ripCD='mkdir temp && cd temp && cdda2wav -vall cddb=-1 speed=4 -paranoia paraopts=proof -B -D /dev/sr0'
