Inspired by one of 82Mhz recent link-dumps, I dropped Navidrome onto my home server and gave it access to my vestigial music downloads collection. I have long used Plex which is great for video but I've never really got into it for music. I think the issue has always been the noise in my collection, which is full of missing tracks and duplicates that Plex does its damndest to gloss over but Navidrome being a simpler application doesn't hide the years of neglect from me. I decided it was time to sort this out.
First I needed to understand the problem a bit better so I rsynced everything onto my laptop to take a closer look, and it was worse than I imagined.
One problem is I keep copies of music in lossless .flac files but I also have copies of that same music in lossy formats for streaming for use on an iPod or whatever, so I decided to spin up a second instance of Navidrome dedicated for lossless music.
Next was to look at what formats the lossy music was in. With a quick bit of bash-foo I got a uniq -c (wiki) summary showing me the vast majority is in MP3 format but a fair bit is in M4A and even a handful of files were in WMA format. I immediately just deleted the WMA files and the future of M4A in my collection is debatable as I'd like to standardise on one format ideally and I'm already very close to that.
Digging deep into the MP3 files is a task still ahead of me, but as I started collecting MP3 files around the launch of Napster and possibly prior to that even, the quality of some of those early encodings is often very low. I'll do an analysis somehow and anything sub-192kbps is likely to be cut at some point too.
I also found that I have two whole "backups" of my CD collection. The most recent I remember doing but I clearly forgot I'd already done it about a decade earlier. I compared the two and deleted the older backup.
For the majority of my music, Navidrome presents no cover art. I understand from my research (reading the docs) is that it will pull cover art from a cover.jpg kept in the same directory as the music, or from the metadata tag itself. Clearly most of my music has neither, so I've begun a (very) long process of pushing everything through Musicbrainz Picard to fix that. So far this has uncovered how much music I have attributed to the wrong album and how some are missing tracks or even whole discs. In more severe cases I'll delete those files if I can't be bothered filling the gaps.
I also have long had the habit of organising my music by their origin, so if I purchased an MP3 from an online store, or it came with a vinyl LP, I would know that file is definitely legit. Probably unnecessary these days but back in the day when people were getting bankrupted for file sharing, it was a genuine - if slightly paranoid 😳 - concern that I couldn't account for some of the files I had. That said, there exists a subset of my music collection I can't account for and I'm now considering making another decisive decision on whether to just delete them and then pick up legitimate copies safe in the knowledge I have paid what is owed to the artists, or at least I can brag about an awesome CD I found in a charity shop bin. Some of these are classic albums I enjoy a lot and really would like to own a physical copy of, maybe on vinyl too.
At the end of all this, I don't imagine moving away from streaming, I have my views on that topic but I'll save that for another day. Right now I want to adjust how I engage with music and the people that make it by buying it properly and keeping my own well tended system for accessing it.