Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

The lie of music discovery algorithms

zeynepevecen.dev

Post reply on HN
Print view View original

151–160 of 198 posts

Re: The lie of music discovery algorithms

#151

Post by digging »

aljgz wrote: post #17

In the age of internet, engagement optimization and recommendation algorithms create a new way that we are affected by the behaviour of others. That annoying dark pattern on a piece of software you use? Because there are people who fall to it, clicking on an ad or "engaging" more. That stupid show that keeps being recommended to you? Because a lot of people just sit on the couch, watching something on the list that d…

> I'm not interested in things that are very similar to this one.

Are you sure this is an accurate description of your taste? Or do you mean "I'm not interested in things that are similar, but lower quality"?

I'm very much a weird-music enjoyer, and I often have the latter problem where "similar" songs actually just don't capture the same vibe as the truly engaging new song I just heard. But that's not because the algorithm is choosing music that's too similar; it's the opposite. It's trying to choose something similar but can't, so it just picks the next-best thing which I actually don't like.

Re: The lie of music discovery algorithms

#152

Post by fantasybuilder »

DiggyJohnson wrote: post #129
fantasybuilder wrote: post #99

Earlier quoted context omitted.

In my experience literally anything beats that one friend who is a DJ. My friends - professional DJs in Berlin - haven't even heard of Can, to my absolute shock.

Why is it weird that a DJ friend hasn't heard of a artist that you are a fan of?

German beat-oriented DJs haven't heard of one of the most influential German bands that had a critical impact on beat-oriented music? Just an odd thing for people who dedicated their life to music.

Re: The lie of music discovery algorithms

#153

Post by gs17 »

InitialLastName wrote: post #131
gs17 wrote: post #116

Earlier quoted context omitted.

Release Radar is consistently the worst feature of Spotify. It misses entire new albums from artists I listen to regularly, and seems to have a quota of songs to fill so after the first two or three it's no longer aligned with my interests. I can forgive it not being coherent since it's supposed to include multiple genres together, but I can't forgive it going way off from what I like just to hit 30 songs.

Not even Release Radar, but the "New Releases for You" list should probably have new releases by the artists I follow (as a basic minimum).

Huh, I don't even have that section on my Spotify. I have a "New music you need to hear this week" at the very bottom (none of it is anything I need to hear this week), but it's just generic "new music in X genre" playlists.

Re: The lie of music discovery algorithms

#154

Post by entropicdrifter »

fantasybuilder wrote: post #97
entropicdrifter wrote: post #83

Earlier quoted context omitted.

I wish I had a music recommendation service built on Pandora's immense dataset of music tags that could build me a playlist that I could link back to whichever music service I happen to be using at the time. I could have it do things like require at least 3 tags in common between adjacent tracks such that it could jump around between 2 dozen genres but the transaction between any 2 given tracks isn't too jarring. It'…

The primary advantage of Pandora's algorithm is the human-labelled Music Genome database. I haven't seen any other company do music discovery as well as Pandora, and don't expect that to change any time soon.

Right? I feel like it might be worth licensing access to the Music Genome db and building a small business off of that

Re: The lie of music discovery algorithms

#155

Post by MrDrMcCoy »

bsder wrote: post #150
MrDrMcCoy wrote: post #119

Way back when Last.FM has its own radio service, I could throw a few random genres and/or artists at it, and it would recommend me pretty much exactly want I wanted every time. I gladly enabled scrobbling in my music players, and it tended to recommend me good stuff every time. Ever since its radio feature got killed, its database has been getting worse and worse. Just now, I tried to search for some things it used t…

> Whatever happened to the good databases and their algorithms? They definitely used to exist. It was called a record store employee, and they no longer exist. There's an AI search engine I'd like to see ... Even the "market segmentation" of pop music still doesn't work for crap. Even something as basic as "Gee, I like 80s New Wave, how about recommending some artists born roughly in the 21st century who would fit?"…

My tastes have always been niche enough that most record store employees would give me the "deer in the headlights" look or condescend to me when I asked for a particular artist or similar that I found on Last.fm.

If the database quality of current Last.fm were similar to its state back when it had radio, I would think that an AI trained on their data would be pretty good. With the current state of it... It would have to be crap. Heck, even if an AI model could be trained on the play counts of every song of every user on every streaming service, I'm not sure it could approach the curated relational algorithm that Last.fm had at its peak. Would definitely love to see an attempt, though.

Re: The lie of music discovery algorithms

#156

Post by MrDrMcCoy »

Triphibian wrote: post #140

One of the things I find frustrating about music suggestions is that the app/algorithm doesn't care why you like a certain band. A long time ago I asked Pandora to make David Bowie station and it rolled me a generic classic rock station -- Zeppelin and the Stones. I was hoping for old school glam, maybe T-Rex and Eno. There's no way to communicate that desire to our music players. To say, "please don't think me a bas…

At its peak, Last.fm was pretty good at this. It had enough data on "people that like these songs also tend to like these songs" that it would generally know what to recommend. I miss those days...

Re: The lie of music discovery algorithms

#157

Post by cmgriffing »

frakt0x90 wrote: post #143

There's a VST that creates embeddings of all the sound samples you have in a directory and then projects it down to 2d so you can visually explore and find sounds you're looking for. Similar sounds are near each other. I always thought doing that with the entire spotify library would be amazing. Give me a low dimensional space to explore the library. Even cooler if the embeddings have similar geometric properties of…

That sounds awesome. Link please?

Re: The lie of music discovery algorithms

#158

Post by karaterobot »

cush wrote: post #82

So I've been using Tidal for 5 years now, and feel they run circles around Spotify in terms of curation. Their algorithms and curated tracks are better, and they steer away from the social/gamification features and lean in to artist-centric features. For example, they've had a "credits" feature since day one - you can look up the producer, guitarist, oboeist, etc of any song, and see what other work they've done. In…

I would naively assume that both Spotify and Tidal pull that information from the same sources. Hopefully there is a web service. Not to take anything away from the first person who decided to, you know, actually tell us this information. At this point it should be table stakes for a music streaming service-indeed, it should be a legal requirement if you ask me.

Re: The lie of music discovery algorithms

#159

Post by alexhutcheson »

janalsncm wrote: post #112
vhiremath4 wrote: post #108

Earlier quoted context omitted.

So collaborative filtering?

Collaborative filtering is similar but for huge recommender systems they’re not going to create a huge MxN matrix where M is users and N is items. I think what they’re referring to would be called a “two tower” model where you have a learned vector for the user, a learned vector for the song, and the cosine similarity is their affinity. It’s pretty performant because you can cache the song vectors.

Google has a great free online course on Recommendation Systems that goes through the various common approaches, with working code in Colab notebooks: https://developers.google.com/machine-learning/recommendatio...

[Disclosure: Work at Google, but not on that. Just thought that course was particularly well-designed.]

Re: The lie of music discovery algorithms

#160

Post by msla »

zeynepevecen wrote: post #147
astrobe_ wrote: post #130

Earlier quoted context omitted.

Oh, my bad! It's a "landing page"! Totally normal that it needs to 200 Kb of JS to display 1 Kb of text!

There’s an email subscription field thats why i need js hahah you’re so fond of yourself

What in the world are you talking about?
Post reply on HN