←back to thread

I ditched Spotify and set up my own music stack

(leshicodes.github.io)
275 points starkparker | 4 comments | 04 Sep 25 22:47 UTC | HN request time: 0s | source
Show context
sunrunner ◴[04 Sep 25 23:54 UTC] No.45133582[source]▶
I'm always glad to see people move away from Spotify's model and towards options that better support artists directly, and I definitely don't mean this to take anything away from the article despite how it sounds, but just seeing the system diagram reminds me that it's amazing the lengths that systems-minded people will go to to create their own Rube Goldberg-esque systems to 'optimise' the experience.

I counted thirteen separate components. If it works for the author then more power to them, but I personally want to spend less time futzing with technology when it comes to this kind of thing and more time actually just actively listening to new music.

I buy from Bandcamp or Apple, sync locally, and I'm done. Bandcamp's iOS app is better than Apple's Music at this point (though not a hard bar to reach). And I find new music organically from listener-supported streaming public radio.

I haven't mentioned analysis or recommendations, but honestly I so rarely seem to find anything through the typical algorithms and recommendation-type mechanisms that I genuinely like, and stumbling across something new just from having public radio on in the background still feels magical, organic, and overall such a good way to broaden your musical horizons.

Still, a good starting point for people wanting their own similar setup.

replies(2): >>45133643 #>>45135257 #
1. sfRattan ◴[05 Sep 25 00:04 UTC] No.45133643[source]▶
> stumbling across something new just from having public radio on in the background still feels magical, organic, and overall such a good way to broaden your musical horizons.

I've largely given up on algorithmic recommendations and gone back to human curation. There are humans out there writing about music, movies, and everything in culture. I've found the ones whose tastes I largely trust, and I follow them via RSS to read about the things I might like.

Are some of those critics probably using algorithms themselves? Sure. Let them dive into that swamp and pull out the gems. I'll stay on the shore, watch, and wait.

replies(1): >>45133776 #
2. sunrunner ◴[05 Sep 25 00:28 UTC] No.45133776[source]▶
> human curation

More and more I feel like recommendation algorithms for discovery of anything seem to just not actually work for finding things which are new and exciting, but perhaps that's by definition.

If information is surprise then the most interesting things are those which aren't like the things I already know. And the easiest way to find those things I find is to just tune in to something where you don't know what you'll hear, and simply wait. That's it. It might take a while, but I bet you'll find something that feels new, exciting and perhaps expands your tastes a bit. And what could be better?

replies(1): >>45133807 #
3. sfRattan ◴[05 Sep 25 00:34 UTC] No.45133807[source]▶
Absolutely. I've made several new Spotify and Pandora accounts over the years. Initially they offer good recommendations but eventually the algorithms always aggressively funnel down to the same 2-to-3 dozen similar-sounding songs (though its a different set of songs with each new account). Once trapped in that algorithmic tarpit, the only thing to do is start over, which is annoying. Now I let myself discover things via human critics or just in the course of life.
replies(1): >>45135177 #
4. prawn ◴[05 Sep 25 05:13 UTC] No.45135177{3}[source]▶
I use the song radio mode to find tracks relevant to distinct songs that I've enjoyed recently, and I find this incredibly effective at unearthing new material that I really like. For many years, I've had no trouble with this finding new artists and quite different songs. Then if I've listened to those a few times, that seems to populate Discover Weekly with a new angle.