I ditched most streaming services well over a year ago now, but Spotify has clung on because I have a playlist of around 2000 songs. I’ve set up Navidrome but now need to transfer all my music in the highest quality possible as efficiently as possible.

I tried lidarr some time ago, but it seemed to be based more around artists than individual songs and my indexer failed to find most of my library.

I’ve seen a couple of apps that will look at a playlist and then try to yt-dlp the song from YouTube but I’m worried about having a lower quality or different version. I’ve wondered if automating an “analog hole” type approach where I just pipe the audio of each song to a file and leave it playing overnight for a couple of weeks might actually be the best approach but that does seem a bit insane at this scale.

  • paequ2@lemmy.today
    link
    fedilink
    English
    arrow-up
    5
    ·
    13 hours ago

    So what do they actually offer, on top of the work of creative people making music? Not much.

    OK, so I left Spotify for Navidrome a while back BUT. What Spotify sells isn’t music. Spotify sells curation and recommendations. Most people aren’t music lovers that want to hunt for cool new music. They just want a pre-generated list of songs that they’ll more or less like. That’s actually kinda huge.

    A recommendation engine is something I wish fediverse or open source would tackle. I’m on Navidrome now, but I’m definitely listening to way less music now—access isn’t an issue—I just haven’t had time to hunt around for new music. Investigating new bands takes time. On Spotify, you do it without even really thinking about it.

    • Fizz@lemmy.nz
      link
      fedilink
      English
      arrow-up
      3
      ·
      10 hours ago

      Listenbrainz is trying to tackle open source recommendations. Its not to bad.

    • damnthefilibuster@lemmy.world
      link
      fedilink
      English
      arrow-up
      2
      ·
      12 hours ago

      I reckon the problem with that is… what’s the source for the recommendations and then what’s the sink?

      Like, first, how do you get all that information about music, type of music, musicians, year of release etc?

      Then where do you store it? Then you come to the problem of building a robust recommendation engine. Sure that’s one step that seems solvable with open source. Not easy. Solvable.

      Then, what does a person do with the recommendations? So you have to build ways to export to YouTube, Spotify, Apple Music, FOSS music solutions. Perhaps plugins are the way to solve this.

      Not saying it’s not doable. Just difficult.

      Though I also believe someone would have tried to tackle it in their capacity in the FOSS world. Don’t know how Fedi plays into this. Maybe an online radio station?