7 ms·

> Which music service's discovery features actually work? It depends on what you mean by "work". Discovery capabilities have certainly gotten vastly better. 1

> Which music service's discovery features actually work? It depends on what you mean by "work". Discovery capabilities have certainly gotten vastly better. 10+ years ago, the only decent one was the now (effectively) defunct Last.fm. These days, they're all pretty good. Spotify, pandora, google music, and now youtube music will do a good job of giving you recommendations based strictly on what you've been cue-ing up. But the recommendations from these services are the equivalent of going into a record store and getting advice from a dim-witted and disinterested employee. You'll get all the obvious stuff, maybe things you forgot about, and if you happen to like popular stuff the recommendations will work OK. But you won't get challenging, provocative recommendations that expand your taste. You'll get cloying recommendations that try to cater to your taste like it was a static attribute. Oh, yeah, and there's "the surveillance capitalism thing" which happens to be the centerpiece of all these services. Is that a problem? Yes. The best "discovery algorithm" is still HUMAN BEINGS. If your cool friends aren't available, then the next best thing is a mag like pitchfork (https://pitchfork.com/ https://pitchfork.com/), xlr8r (https://xlr8r.com/ https://xlr8r.com/) or in-depth reviews like Anthony Fantano's channel (https://www.youtube.com/user/theneedledrop https://www.youtube.com/user/theneedledrop).

greggman3 6y ago
> These days, they're all pretty good Your experience is vastly different than mine. Youtube music seems to be recommending nothing but what's popular. Justin Bieber is being recommending to me. I've never listed to him or anything remotely related. No good recommendations on any of the others. Maybe you have a different definition of "good" Good to me means "sounds similar and in the same genre as what I'm currently listening to". It does not mean "people who liked this song also liked that song"
- crispyambulance 6y ago
  > It does not mean "people who liked this song also liked that song" Actually, yes, it does (and many other things too), but to be fair "good" is a highly subjective judgment which is going to be different for everyone. I don't think, at this point in time, we have recommendation engines that can do much more than fling out recommendations based on an unknown convolution of your listening history combined with music meta-data combined with social network data and a mix of paid stuff courtesy of your surveillance capitalism purveyor. I know it's possible to capture some characteristics from the music track itself, like bpm (perhaps usable for EDM DJ's?). The "holy grail" would be to have a system that can truly assess the nature of a piece of music based on audio and use it make "interesting" and non-obvious recommendations. We are very far from doing that in software, but humans are still very good at it.
  - vidarh 6y ago
    The problem with "people who liked this song also liked that song" is that very often I don't want to listen to that song now even if it's something I really love. If I'm listening to ambient music I don't suddenly want to be ambushed by something uptempo. If I'm listening to e.g. Debussy, you might be excused for suggesting something vaguely new age in a similar mood and tempo, but certainly not rock. Another problem is that you can't just take raw overlap in tastes, because some people like "everything", and the fact their tastes overlap with mine does not mean I'll like everything else they like. I've yet to hear a recommendation system that chooses music I want to listen to reliably enough that I can generally stand to listen to them for more than a few songs at a time without it turning into an endless annoying sequence of skipping. Respecting genre (segues need to be gradual, if at all), respecting mood and tempo needs to come first. Then you can consider what others who likes the same songs within those constraints also likes within those constraints. Honestly if I have to choose between personalised recommendation and precise control of genre and mood/tempo, I'd take genre and mood/tempo over personalisation any day. Another pet peeve of mine is lack of visibility into how to teach a system what I want. E.g. if I dislike or skip a song, will it get that it doesn't fit my current mood or what I want to listen to now, or will it wrongly infer I don't like the song at all? Sometimes it feels as if the people designing these systems don't use them.
    - crispyambulance 6y ago
      > you can't just take raw overlap in tastes, because some people like "everything", and the fact their tastes overlap with mine does not mean I'll like everything else they like. I think that these recommendation systems, especially youtube's, are much more nuanced than you're suggesting. And I say this as someone who finds these recommendation systems lacking and of limited utility compared to discovery from reading Pitchfork, for example. They're trying to balance quite a lot of inputs and actually somehow make money from it. Also, "genre" and "mood" of a piece of music are not easy to define, let alone measure. You're asking for a lot and I don't think what you're asking for is realistic from a piece of software.
      - 6y ago
HeadsUpHigh 6y ago
Spotify has gotten worse. They don't want you to listen to too much music, they want just enough to keep you subbed.
tomc1985 6y ago
DJs should be in your list -- downloading tracklists from sets that I like is my #1 method for discovering new music Find a couple DJs you like, download everything you can from their soundcloud or mixcloud, then give it a listen. Keeping Shazam close by is also helpful

greggman3 6y ago

> These days, they're all pretty good Your experience is vastly different than mine. Youtube music seems to be recommending nothing but what's popular. Justin Bieber is being recommending to me. I've never listed to him or anything remotely related. No good recommendations on any of the others. Maybe you have a different definition of "good" Good to me means "sounds similar and in the same genre as what I'm currently listening to". It does not mean "people who liked this song also liked that song"

crispyambulance 6y ago
> It does not mean "people who liked this song also liked that song" Actually, yes, it does (and many other things too), but to be fair "good" is a highly subjective judgment which is going to be different for everyone. I don't think, at this point in time, we have recommendation engines that can do much more than fling out recommendations based on an unknown convolution of your listening history combined with music meta-data combined with social network data and a mix of paid stuff courtesy of your surveillance capitalism purveyor. I know it's possible to capture some characteristics from the music track itself, like bpm (perhaps usable for EDM DJ's?). The "holy grail" would be to have a system that can truly assess the nature of a piece of music based on audio and use it make "interesting" and non-obvious recommendations. We are very far from doing that in software, but humans are still very good at it.
- vidarh 6y ago
  The problem with "people who liked this song also liked that song" is that very often I don't want to listen to that song now even if it's something I really love. If I'm listening to ambient music I don't suddenly want to be ambushed by something uptempo. If I'm listening to e.g. Debussy, you might be excused for suggesting something vaguely new age in a similar mood and tempo, but certainly not rock. Another problem is that you can't just take raw overlap in tastes, because some people like "everything", and the fact their tastes overlap with mine does not mean I'll like everything else they like. I've yet to hear a recommendation system that chooses music I want to listen to reliably enough that I can generally stand to listen to them for more than a few songs at a time without it turning into an endless annoying sequence of skipping. Respecting genre (segues need to be gradual, if at all), respecting mood and tempo needs to come first. Then you can consider what others who likes the same songs within those constraints also likes within those constraints. Honestly if I have to choose between personalised recommendation and precise control of genre and mood/tempo, I'd take genre and mood/tempo over personalisation any day. Another pet peeve of mine is lack of visibility into how to teach a system what I want. E.g. if I dislike or skip a song, will it get that it doesn't fit my current mood or what I want to listen to now, or will it wrongly infer I don't like the song at all? Sometimes it feels as if the people designing these systems don't use them.
  - crispyambulance 6y ago
    > you can't just take raw overlap in tastes, because some people like "everything", and the fact their tastes overlap with mine does not mean I'll like everything else they like. I think that these recommendation systems, especially youtube's, are much more nuanced than you're suggesting. And I say this as someone who finds these recommendation systems lacking and of limited utility compared to discovery from reading Pitchfork, for example. They're trying to balance quite a lot of inputs and actually somehow make money from it. Also, "genre" and "mood" of a piece of music are not easy to define, let alone measure. You're asking for a lot and I don't think what you're asking for is realistic from a piece of software.
    - 6y ago

HeadsUpHigh 6y ago

Spotify has gotten worse. They don't want you to listen to too much music, they want just enough to keep you subbed.

tomc1985 6y ago

DJs should be in your list -- downloading tracklists from sets that I like is my #1 method for discovering new music Find a couple DJs you like, download everything you can from their soundcloud or mixcloud, then give it a listen. Keeping Shazam close by is also helpful