Hacker News Reader

Comment by dankwizard | Hacker News Reader

This article fails to mention the absolute butchering of features that takes place moving from a typical music streaming subscription to a self hosted Jellyfin library.

A large part of my listening on YouTube Music is going to a particular song or band I like and clicking "Radio", which generates a playlist of similar sounding songs. You can then fine tune it with a filter i.e "Popular songs, deep cuts" or specific elements of the song "More emo", "Slow paced" etc. This exposes me to a lot of new music and keeps it fresh and if I'm lucky I'll discover a new artist or song to add to my rotations.

You lose that.

A lot of these services overtime build mixes which takes your listening habits and tries to categorize them into specific mixes made up of your existing library & new music.

I don't browse any music forums and so apart from my favourite bands, I have no idea on when artists I like release new albums and would not encounter them on a self hosted solution, etc.

View on HN

I would have agreed with you 3 years ago. But now not so much.

Spotify "Radio" feature just tends to want to give me music I've already listened to over new music. Whatever algorithm they are using has waaaay overfit to what I have already liked.

There used to be curated playlists done by humans, now almost everything is "made for you by Spotify" playlists which, have the exact same issue as the radio stations, suddenly it's all the same music you've already been listening to, very little new music. If you want new music, you need to find a playlist made by a user instead.

Spotify « radio » is the best reason to listen to real radios ! [honestly the DJs on most of the radios I listen to are insanely skilled !]

Btw, is there somewhere a search engine to know when a given [set of] track was played where, in the internet radio world?

This! I recently ditched Spotify and rediscovered radio in the past few weeks. There are so many great songs I've come across from bands I enjoy that I had never heard of because, as someone else said, Spotify's algorithm is way overfit.

It's also great sometimes to discover great music from genre you usually don't like, or... just be exposed to songs you don't like. This is what helps building a musical culture.

Please allow me to recommend FIP, as a human (it's a classic here but there's no such thing as recommending too much FIP) : https://www.radiofrance.fr/fip

Radio Paradise is a great alternative to FIP.

https://radioparadise.com

Both are classics. I like SomaFM-GrooveSalad too. Oh and BBC6, or course. [The podcast of Guy Garvey is an absolut must, in my sunday schedule]

France Inter Paris (FIP) it's awesome!

And remember you can always get the audio steam through HTTP Live Streaming (HLS), on its M3U format, or others with better quality. There are many Android apps like Transistor to enjoy the stream, and even VLC can open these, in order to avoid using a web browser.

Likewise, I prefer online radios than big tech algorithms that craft my music experience.

FIP and NTS are my goto's. The discovery features for shows on NTS and the "in focus" specials are great, so many good opportunities there for serendipitous listening. Will def check out radio paradise

FIP is broadcasting in FM in France, so no big news on that one, for me. But i will investigate NTS. I knew their radios streams, but it seems they also have some pretty niche podcasts !

Have you tried searching other user’s playlists instead? At least for me, that’s how I have been using Spotify’s to discover new music.

I don't know if its the best at this, but I've been listening to radio from around the world at https://radio.garden/

For a certain range of indie pop, KCRW's Eclectic24 is great: https://www.kcrw.com/music/shows/eclectic24

100%

No algorithm has been able to be able to be as weird but consistent as a community radio DJ.

The radio can still surprise and delight like little else. All the tech companies have been able to replicate is the disappointment.

Not to say never but people’s great advantage here is that they’re people.

I respectfully disagree. If you're into classic rock those stations are pivoting around here to 90s and 2000s rock since that's "classic" now. Then you're left with ButtRock stations that play mostly the same thing every day at the same times in the same order. The best radio we had in our area was a college station that has an hour or two of stuff I'm interested in or as close to a legal pirate radio you can get (100w tower) that shut down - THAT was amazing. Had a ton of DJs who played things they liked.

Outside of rock you're left with automated pop and country stations who have computerized playlists.

note - I'm in the Midwest US.

Community radio DJ. Community being the important part.

Most of the radio stations here in Columbus, Ohio are what you described, the clearchannel / IHeartMedia stations.

However, there is an independent radio station and it's so great. They play Democracy Now! during the daytime and they have a rotating list of shows for the evening. I've heard some really great music during the evening shows.

https://www.wcrsfm.org/

If you're in Columbus, tune in to 98.3 or 92.7 FM!

Thanks for putting it so well.

Commercial radio is the devil.

Classic rock exists to sell lawn mowers to middle aged men

reminds me that i should donate to somafm again :-)

Good idea.

To those who aren't in the know already, it's this:

https://somafm.com/player/

https://play.google.com/store/apps/details?id=com.dgmltn.rad...

https://onlineradiobox.com has the data but doesn't search on it it seems.

Disclaimer I make https://www.radio-addict.com but only retrieve the played song data on demand (never tried to probe all 80k+ radio streams at the same time on my small server, could be fun), but searching on it could be a new feature (it's stored in Elixir genservers :D)

Wasn't TuneIn providing this search feature before?

"Listen to a random radio" is the summum of serendipity, man ! #kudos

Thanks.

YouTube is much better than Spotify for this in my experience.

My experience with YouTube is that I start with an obscure song/artist and it will gradually bring me to the mainstream. Maybe that is just me... I feel like ideal algorithms died with last.fm era.

Does noone use the Spotify "daylist" playlist, that cycles between genres you have listened to previously?

This regular plays music I have never heard (both old and new).

Do you mean daily mix playlists? From the first 20 songs 17 are something which I added to my library, or listen them regularly. The rest of the 3 songs? 2 of them are from artists whom I listen to regularly. 1 clearly new song.

I have very similar rate with “daytime mix”.

So which one do you mean? “Discover weekly” and ”release radar” have new songs, yeah. But “radios” are like the previously mentioned playlists.

No none of those. Is a special dynamic playlist, starting to wonder if its not a standard playlist everyone has..

sort of like the DJ mix without the annoying voice. https://open.spotify.com/playlist/37i9dQZF1EP6YuccBxUcC1

Not available for everybody.

> As of today, daylist is available to both Free and Premium users across the U.S., Canada, the U.K., Australia, New Zealand, and Ireland at spotify.com/daylist.

https://newsroom.spotify.com/2023-09-12/ever-changing-playli...

(note: old post, but still accurate?)

I’m in Southeast Asia, daylist has been around for years in my side of the world.

> suddenly it's all the same music you've already been listening to, very little new music.

However, if you expose the gods of the algorithm to a new artist, suddenly all the auto-generated feeds will try to include that band regardless of fit. Weird how these "social graph" systems tend to form and perpetuate bubbles.

On top of that, there are some weird shenanigans with meta-data. Listening to "foreign" bands may very easily taint the weekly mix with songs in a language you don't even understand and probably don't care about. An anecdata of course, I just looked at my "daily mix x", which appears to be in my local language, but with styles all over the place. Another mix contains mostly correctly turn of the century romantic pop.

I suspect the algorithm biases heavily on metadata so that it could be easily fed "albums/artists that publisher x paid to promote".

> However, if you expose the gods of the algorithm to a new artist, suddenly all the auto-generated feeds will try to include that band regardless of fit.

cf YouTube when you watch one video on X that's outside of your normal viewing and RIP your homepage for the next few days until you've clicked "do not recommend" on enough videos to stop the flood of X and X-adjacent content.

Spotify radio regularly makes me angry, and makes me want to press a "dislike" button really hard. But of course, that button is missing ...

This.

My guess/conspiracy theory is that Spotify has cut deals with record companies that pay less on subsequent listens to a track so the repetitive radio algorithms are more profitable.

Anecdata incoming, but to offer an alternative view, I would really love to not use Spotify anymore since they change things constantly in ways I don’t like, but their music recommendations are fantastic for me.

Their generated playlists are great, and they do a good job recommending playlists I’d like from other users as well. And while I hate the format, their music shorts actually give me consistently good music. I just hate that it’s in the TikTok swipe style.

Spotify has not viewed itself as a music company for longer than that. It's a platform for audio. And, while there are still music first people at the company, they are not in the power positions that they used to be.

The transition didn't start when they laid off Glenn MacDonald, but that sort of cemented it. They had already gutted curation before that and by this time you were far more likely to find people talking about AI in the halls than music. If you've never heard of Glenn, check out his book: "You Have Not Heard Your Favorite Song: How Streaming Changes Music." Or his old online projects at https://everynoise.com/.

Yeah I have no idea why music recommendation algorithms have gotten so bad. Rdio had the pinnacle of music recommendations in 2013. I wonder what happened to that tech after they shut down in 2015, and why no one has been able to reproduce it.

just look at google search in 2015 and now for the answers… :)

I never got on with Spotify radio.

In fact, things have never been as good as last.fm used to be when it hosted its own music.

It's a "YMMV" situation, because...

I don't want that. At all. It's algorithmic and there's nothing stopping artists and labels paying for placement in there. I don't want that.

I am a musician, and a DJ, and I've been digging deep through artist and label catalogues on my own for decades. The process of discovery via my preexisting routes is far more fruitful, enjoyable and rewarding than lazily letting an algorithm do the work.

But I like doing that. This works for me, not for others.

I miss last.fm and audioscrobbler so much. Yeah it's still there, but it's not like the old days, which was far better for feeling like I was in touch with other users, finding tons of new music, and not hearing yet another Home Depot ad.

I half wonder if Spotify US is one of those harbingers of doom for the old web in the same way I think of Twitter and iPhone/SPAs/The Stream UI.

Fun fact since I already ranted about last.fm being the bees knees when it integrated with Spotify - my wife looked up on lastfm music compatibility after we first met and her and a Russian lady are the only 3 above like 75% similar taste in music. Small world but also wild that it's not so small.

Big streamers cater to most people's interests of music and are probably satisfactory for them but for us weirdos in music last.fm and it's music genome thing were amazing.

ListenBrainz has filled that void for me. My profile here, https://listenbrainz.org/user/iPodSmypod/ Blog post on the topic: https://theonlyblogever.com/blog/2025/scrobbling-libre.html

I had a peripheral awareness of listenbrainz for scribbling, and think that's neat. However, last.fm's radio and song/artist/genre/tag search and suggestion capabilities were on an otherworldly level. Do the *brainz services have anything that really compares?

A slower speed on the hedonic treadmill is a feature of self-hosting, not a bug.

Leaving music streaming services has been a great excuse for me to rediscover music blogs like Gorilla vs. Bear and Stereogum, or even local culture magazines.

Another great way for discovering music I've found is just perusing Bandcamp, which is where I buy most of my music anyway. Love finding local artists, so I just put in some genre filters and the location filter. Found multiple great bands this way.

As for keeping abreast of new releases, Bandcamp is pretty good for that too. You can just follow artists and you get emails when new releases or merch or tours come around.

I've never seen this work. Either it plays the stuff I've listened in the past in a loop or shove some random things I really dislike (maybe hidden promotional stuff?). Personally it's the reason I've cancelled subscriptions each time I've tried, I always ended up listening to radio instead as the value brought by Spotify etc... was really poor.

I use YouTube Music and it definitely does work but yeah it weighs songs you've already said you like way too heavily and generally seems way worse at discovering similar music than Last.fm or Pandora from over a decade ago. (If anything I remember Pandora being too good at finding similar songs - the playlist would end up almost monotonic because it found such similar music.)

Fortunately they do have a "Discover playlist" that completely excludes music you've heard before. Unfortunately that's all you get. No way to e.g. say "play me reggae I haven't heard before", and it's only updated once a week.

So yeah... kinda shit. But still better than the alternative which as far as I remember from the 90s is to only listen to extremely well-known bands and find good news music like once a year.

Plexamp is really good for this.

The styles information that Plexamp has works really well and in my experience, as long as your library is large enough, works better than modern Spotify.

It was Spotify's degradation of their radio service and terrible "AI DJ" that finally got me off Spotify. Punishing them for platforming Joe Rogan was just icing.

I find Plexamps Radio features, and DJ's to work great against my library. It also helps if you, as an individual, have a diverse and large music library to supply for the sonic analysis.

I'll argue music algorithmic recommendation on these platforms is a bad thing anyway.

First, the algorithm is opaque, so it can push stuff to you because the platform decide it has to get the spotlights. Maybe the label/producer/musician paid for it or whatever you want to imagine that is even worse. It is a well-known phenomenon that if some music is pushed to your ears, you'll end up appreciate it most often than not. This is how hits have been and are still made.

But even if the algorithm was not gamed at all, I still think it is a bad thing. It is not going to push you out of your comfort zone. Listening to new stuff is usually not pleasant at first. You will only "discover" things that are very similar to what you know and already enjoy.

If these recommendation algorithms were about food, they would "reason" like this: "Hey, you've really enjoyed this whole pack of M&M's, I'm sure you'll like this Kit-Kat bar now! Oh and you've had a glass of wine, what about trying out meth, it's pretty good too.". Do we really want our computers to reinforce such behavior?

Go to concerts, buy merch, buy albums on bandcamp (it has not enshittified too much yet apparently), donate money to artists; discover music through your friends and other humans recommending it. Recommend what you like to your friends. Cancel your Spotify subscription, none of that money is going to artists anyway. And use soulseek.

What are the musical equivalents to Kit-Kat and meth?

Equivalence is too strong a word, but content produced by spotify where musicians (or AI prompters) are mere contractors comes to mind.

Getting back to "I don't even want virtuous algorithmic recommendation"… I like jazz rock/fusion, especially when it has a touch of bluesy/blues rock influence. There is probably a lifetime of listening time of that genre, and it takes no effort for me to appreciate anything that resemble this. Long guitar solos by a jazz-educated guitarist who happens to like Jimi Hendrix, sign me up.

But I do think there is value in getting out of my comfort zone, and listen to something drastically new, from time to time. It requires effort though. My first reflex when I hear synthetic drums or autotune, for instance, is to press "next". But it is through other humans being recommendation, that I sometimes make that effort, and actually learn to appreciate something else.

Call me an elitist prick, but I hate to think of music as a commodity for us consumers to consume. It is art. Art is not always pleasant. It sometimes becomes pleasant after overcoming an initial disgust.

It was like this in the past, now it's crappy. The algorithmic optimization started eating its own tail. And it's a problem on all platforms, from Spotify to YouTube.

Let's take YT. In very simple terms, instead of taking a bold move and suggesting a few outliers (similar to differentiating the population as it's done in evolutionary algorithms), it takes an easy shot and, if I'm identified as male, suggests some videos with females with big breasts and other generic junk many people just click on autopilot. It works well for them because most people click and click and spend their days uselessly hooked and feel bad, but in my particular case I lose what I had earlier, i.e. suggestions of interesting bands (they still do happen but the selection is of much lower quality).

This hasn't been my experience at all. Not sure why Youtube would suggest big breasted female artists when I never, ever, ever watch the video when I'm playing music.

> This hasn't been my experience at all.

No wonder, I'm pretty sure they're doing A/B testing all the time.

> Not sure why Youtube would suggest big breasted female artists when I never, ever, ever watch the video when I'm playing music.

They don't care. I sometimes open YT on new machines and it's always the same generic junk.

My favorite songs gravitate heavily towards 2 very different genres. This seems to confuse the hell out of Spotify. The "discover weekly" is comically bad no matter how hard I try to prime my library.

I've discovered so many niche bands and subgenres since I got Spotify.

> I have no idea on when artists I like release new albums and would not encounter them on a self hosted solution, etc.

Depending on what you like, bandcamp makes it easy. You can follow any artist (which is also offered whenever you buy), and from then on get release notifications. But of course, what’s available differs by genre. For metal, most bands are on BC, except most Japanese artists and major label stuff.

I buy, download, and put the flacs on my Jellyfin server.

There are, of course, also piracy solutions for that, pretty sure the *arr stuff has automatic downloading per artist.

By moving away from streaming services, you can once again own what you bought and paid for. Algorithmic playlists are nothing, nothing at all compared to the loss of ability to use your own player, edit your files, back them up, or not be nickel-and-dimed to get around artificial restrictions. Not to mention that with streaming services, music can be taken away from you after the purchase.

One can have both: use the streaming services to discover new music, buy physical media of the bands you like most. Not everyone has this option, but it is an alternative.

My experience with Youtube Music is that the recommendations are quite poor. So I wouldn't miss that. But it's hard to replicate the breadth of coverage of YT music (even though sometimes songs just vanish from my playlists). But I have started buying a couple of albums every now and then and slowly I am building my owned music library.

I'll admit it, I have a fairly narrow range of music I like so the following works for me on this basis: I don't like Spotify and other music streaming services as they never are consistent with their licensing or good with their recommendations. And the adverts are obnoxious. What I like is radios like Radio Paradise: https://radioparadise.com/player or regular radios available through online streams (such as the French radio FIP: https://www.radiofrance.fr/fip). There is enough to discover on either and they are still mostly in the range of what I would/could listen should they not have existed.

IMO using a streaming service’s recommendations is a way to filter out bands that labels aren’t promoting. The services have to be getting paid for pushing - right?

If everyone is this lazy about music discovery, then music suffers. I am not using “lazy” as a pejorative. There are people who just couldn’t be bothered and that’s fine. Music just isn’t that important to you. But if the people who deeply love music are corrupted by the ease and dopamine, it will deeply wound music as a whole.

My problem isn’t discovering new music, it is “discovering” my massive library. I love AM, but the fact that 3 of the five large icons taking up precious screen real estate are devoted to discovering music that Apple is paid to promote is infuriating.

> IMO using a streaming service’s recommendations is a way to filter out bands that labels aren’t promoting. The services have to be getting paid for pushing - right?

I keep hearing this, I also read Mood Machine that says the same thing over and over again.

Yet, every week for years, thanks to Spotify I'm discovering artists that don't have the means to pay to be promoted in any way.

I'm not saying that I'm special or that I hacked the algo or that it's false that Spotify promotes artists that are paying to be promoted.

I'm just saying that, depending on how you consume (I'm using this verb on purpose) music on streaming platforms, you'll be more or less targeted by money-making thongs.

If you listen to the main playlists, they are basically 100% promotion based. But if you listen to a non-promoted artist's radio, you'll have far less promoted content.

Seriously. I've discovered artists with less than 1000 monthly listens thanks to the algo.

Same, a few just a few hundred monthly as well. And they were good hits, ended up going to several concerts based on Spotify's random playlist.

This isn't an article about replacing all-knowing streaming service content recommendations, it's about finding a solution for playing local music files you own in a convenient way.

I'd like to shoutout PG Vogt (from Reply All podcasting fame) for this episode of his new show:

https://pjvogt.substack.com/p/how-am-i-supposed-to-find-new-...

> I don't browse any music forums and so apart from my favourite bands, I have no idea on when artists I like release new albums and would not encounter them on a self hosted solution,

Music Brainz provides this at https://test.listenbrainz.org/explore/fresh-releases/

There's also Music Butler: https://www.musicbutler.io/

I think for some people the goal shifts from discovery to ownership - knowing your library, building it intentionally, and not being nudged by what the algorithm thinks you should be into

PlexAmp has DJs which allow you to get the song/playlist based radios.

LMS (Logitech now Lyrion) also has something similar in MusicIP (not as good as PlexAmp).

its hard to beat the convenience of being able to right click/radio to get new recommendations but there has to be other options that arent that much more effort either?

i think you can add plugins to jellyfin. maybe there is a last.fm plugin? i know of some other last.fm alternatives like maloja or libre.fm but i cant comment on how good they are

Sure, but then what - A script to pirate the song? Or would your script generate the playlist and automatically buy the tracks from bandcamp?

Those features can, and should, be made completely separate from the system that hosts the media. In fact, they used to be, with great success.

How are LLMs for music recommendation?

Napster / audio galaxy... I mean your own legal burned music with AI generating a radio playlist.

LLMs can’t hold the context of my library in memory. They come up with playlists constructed of music they are aware of, not necessarily music I have.

I imagine not great. Almost certainly heavy bias to popular songs instead of "niche but relevant"

Plexamp has this feature, i use it all the time.

That was cool 10, 12 years ago. Now that everything has been exploited and enshittified, I want to go back to the pre-infinite-streaming experience. The only issue is cost vs being a pirate...

I mean right at the top he says hes just trying to listen to his own music. I dont get how this is a downside if you wanna discover new shit you can always just go to youtube.

Frankly 99.9 of my music listening is stuff I already know and enjoy. But I still like to listen to new stuff often. So this kinda thing is perfect for me 99% of the time.

Reply on news.ycombinator.com