Hacker News new | past | comments | ask | show | jobs | submit login
How streaming is changing songs (theverge.com)
73 points by tosh 24 days ago | hide | past | web | favorite | 52 comments

I'm basically streaming-only these days but I still find "singles only" artists tiresome to deal with. Maybe it's a shortcoming of Apple Music, but if I want to listen to an artist, I usually try to pick an album so there's something playing for at least the next half an hour without me managing my up next list.

For some artists there may be an "essentials" playlist or something, which helps, but that already requires a certain level of popularity. If there's nothing like that and no albums, I need to open a single, pick one track to play (because four remixes of the same song one after another isn't my idea of a good time), then the same thing for the next single, etc. Usually I just give up and go find something with less clicking.

Most people don't see that behind the scenes payola dictates everything you see on charts now...

Artists pay to promote their posts on Social media, so only the rich or well funded rise to the top.

Artists remix the same tracks over and over because they know doing the same track over and over means they can hopefully capitalize on profits already made rather than creating anything new, which takes more time and costs more.

Albums used to be high on filler, I think most people forget how only 2/16 songs made it to radio, while 10 of those tracks were often skipped, while you paid for all of them. now if you buy a single, you're only paying for what's good, even though $1-$2 is still quite high for a single music track if you ask me...

Music artists have always been hosed by the industry, good music is always suppressed by gatekeepers that want to influence what everyone gets to hear, but that one little band you like can greatly benefit, and will likely live longer together, by you telling people about them... Stop relying on other people to tell you what's good and let's let the listeners dictate what the charts say...

Remember singles? On records, CDs, and tapes? And then on iTunes? I’m not really sure there’s that much of a change here. A lot of artists have been always been releasing music this way, haven’t they?

The content of the article is an interview with a single artist, who just released an album.

It reeks of promotional material - and The Verge doesn’t have a great track record of being an expert on anything.

Oh lol. I reached out to Charli's team before I even knew her album was dropping. It almost seemed fortuitous that the album was happening by the time I talked to her in her house. Then after the fact she dropped way more singles than I was anticipating, and the whole feature was in flux until the end because she is so reactive. There was nothing planned about this - she represents exactly the topic I talk about. And it made it harder for me because I couldn't anticipate her next move. Not easier, which would have been the case if this were "promotional material."

Neir the end of the interview, she mentions some specific formulaic tricks to reduce skip rate. She says it's all about grabbing people in the first five seconds, putting all the good parts in the beginning of the song, avoid "weird self-indulgent intros", etc. Those are things that sound pretty specific to streaming.

Also, she talks about how artists stop focussing on albums (it's no longer single -> single -> album -> promo tour), and just release new music all the time (single -> single -> single -> ...)

And make shorter songs. Which is my main surprise. I thought that by freeing themselves from radio playlists (where the radio station mandates short songs) and the physical capacity of a CD, artists would get to make songs of any length, including longer. I am not sure I understand why streaming calls for short songs.

Perhaps compensation is higher for those with shorter songs, possibly due to being paid per play of the song?

This seems plausible. Recommendation Algorithms definitely could play a role as well. I believe that at some point YouTube started recommending videos based on "time watched" so videos slowly got longer over time, as those videos tended to be promoted more.

(Or vice versa, that they became shorter because the algorithm preferred that.)

Wasn't that like 10 minutes after which all videos became 10 minutes to explain what could be told in 30 seconds?

I believe the 10 minute mark has to do with being able to insert additional ads into the video, so if your video is close to 10 minutes it pays off to "round it up".

I predict that streaming, listener analytics and metrics will have a much bigger effect on music production by removing the human producers for mass consumption pop music in the next two decades. The writing is on the wall.

We're already at a point where the producers (who compose the songs and usually even create the audio using computer software) spend a lot of time analyzing the metrics they get from Spotify et al. They look for stats on how many seconds someone listens to a track before skipping to the next one, do they turn the volume up or down, which songs are paired to the same playlists, etc etc.

When reading interviews of musicians, they frequently bring this up and usually not in a positive light. You can also hear the results if you're exposed to modern pop music - when there's a hit single, there will be quite a few copycats that will shamelessly borrow the elements of the hit. This is not a new thing but the degree to which it's occuring is.

Now what's good at looking at metrics, doing tiny adjustments and gradient descending towards some local optimum and cheaper than hiring a musician? A machine learning algorithm of course.

While it's not common yet, there have been some fairly impressive demonstrations I've heard. Computer-generated music notation is easy to generate (read: number of bits in the output is rather low) and has been experimented with for a long time.

But this can go even a step further, rather than producing sheet music, an algorithm can generate the audio directly. As in the PCM audio waveforms (or maybe spectrograms, frequency domain might be easier for algorithms). I recall hearing AI-generated classical music audio (found through HN) which sounded fairly convincing to me (albeit with a white-ish noise on the background).

If you're a professional pop music producer, make the most out of your career while you can. You're closer to being replaced by robots than you think.

I'm not suggesting that human musicians will go away any time soon, I'm sure people will enjoy playing instruments with their own hands in the future. But the kind of producers that create music with a computer mimicking popular songs and tracking analytics from listeners will be, if not replaced, at least heavily assisted by algorithms more and more in the near future.

The tools evolve and that's a great thing.

Music, as painting, drawing, sculpting, writing, stays a mean for people to express themselves, to share their feelings & thoughts.

But it's also a mean to grab attention and manipulate it, too.

> I'm not suggesting that human musicians will go away any time soon

It already happened. Since records exist. See the quantity of shows where there's only a voice performer alone on the stage or on the street, with pre-recorded music.

Not that it is a bad thing - it is a _different one_ that does not remove anything, only adds. Where the purpose is not to perform, but to grab attention with less logistics.

Live musicians are still a (great) thing, but it's definitely not the same thing.

I agree with your main points, even though I'm unsure of the time scale.


> But this can go even a step further, rather than producing sheet music, an algorithm can generate the audio directly.

Even if this is a cool idea, I don't think it ever will be deployed large scale. Good quality beats/samples and increasingly even software instruments are becoming cheap. Working with raw audio seem like the wrong abstraction layer for music, almost like working with pixels instead of characters to synthesise text. While limiting - working with raw audio can possibly create completely new sounds - it will probably always be cheaper and yield better results to work with some representation of the music instead.

The big advantage of using raw audio is that there's an abundance of data (all the music that's ever been released pretty easily scraped from the internet), whereas "semantic" data is rare, expensive and of lower quality (as in: not as good/catchy). The best solution is probably to extract the semantic representation from raw/pcm form. (Incidentally, I'm working on this very problem. Anyone interested, feel free to contact me)

> Working with raw audio seem like the wrong abstraction layer for music

I do kinda agree with you, but I wish I could find the classical music audio clip I mentioned. It was generating a fairly convincing clip of audio complete with an orchestral backing and a piano solo track. Maybe that would change your mind a little.

Unlike text which is discrete characters, music is a continuum of pitch, tone and other characteristics.

Maybe it won't be generating the music, but adjusting the tone instead. There's quite a lot of effort in mixing and mastering music, which sounds like good territory for ML.

To return to your text analogy - it's more like generating convincing images of hand-written text. You might generate the discrete characters one by one, and then proceed to turn that into pixels with another algorithm.

Generative music is already a thing, Jean-Michel Jarre's "En Attendant Cousteau" piece was released close to 30 years ago and Brian Eno coined the term a few years later: https://en.wikipedia.org/wiki/Generative_music

You can find a bunch of video on YouTube where Eno explains the whole process and how it works.

Idle thought: if musicians are being slowly replaced by algorithms, the future might not look like streaming individual tracks, but like apps that generate an endless amount of music locally. I think there'd be demand for such a thing from people who just want something on in the background. Possibly currently supplied by the four-hour or longer sets of electronic/classical music one sees on YouTube.

It's nice they have this freedom, but for a while now, it's been harder and harder to find a truly coherent album. Something that you can leave on and feels like a single piece of work.

I would say the current situation is better than the 90s, in which the CD pushed acts to make albums far longer than they had any sense being (despite being an amazing period for hip hop, it's hard to sit through a lot of the very best acts albums). Acts now have the freedom to make things whatever length they feel suits. Minor releases can be album length, major releases can be EP length, it's all whatever people want to make it.

The biggest issue right now is that gameifying streaming is leading to similar issues (Chris Brown's 5 disc album being an extreme example) but Kanye West's series of 20 minutes albums points towards another direction that could be a much better fit for the modern musician (making an album that's so concise that it results in the listener playing the whole thing over and over instead of just adding their favourite song to a playlist.

Depends on who you're listening to. Most big names rely on streaming plays so you're probably not going to get it there but you can kinda drop the shitty songs and understand what the album could have been. It difficult to make a coherent album with 20 songs but that is what major labels are requiring these days.

I think Bandana by Freddie Gibbs & Madlib is a great example of a full and coherent album I've loved this year. Lots of indie rockers, metal, jazz and experimental EDM musicians are still making great full albums

Bandana is probably my fave of the year so far. Lots of great albums this year:

- JPEGMAFIA "All My Heroes Are Cornballs"

- IDK "Is He Real?"

- Earthgang "Mirrorland"

- Lewis Capaldi "Divinely Uninspired to a Hellish Extent"

- slowthai "Nothing Great About Britain"


- Benny the Butcher "The Plugs I Met"

- Cate le Bon "Reward"

- Denzel Curry "ZUU"

- Flume "Hi This Is Flume"

- Gesaffelstein "Hyperion"

- Kota the Friend "FOTO"

and there are several upcoming releases I'm quite hyped for.

Agree to all of this. These artists, besides maybe Flume, are exactly the size / popularity of what I was talking about in my post.

Sure, but I think "the album" was a side effect of LP/CD era constraints.

Those days are over, and it's a new modern world. As always :)

Sure, but those "constraints" could be compared to switching from a piece of letter paper to a canvas. The modern "liberations" would be having a box of a million tiny artistic legos which are getting tinier and tinier.

In all situations you want the best piece of art, it's not entirely clear to me that a million hand-painted legos is superior to an artist's canvas painting. Sure, it's a lot easier to grab all the blue legos and have 30 hours of deep house, but it's very difficult to end up with starry night, with bits of euphoric light trance and depressing shadow dub melded together into your blue canvas, unless you go back to your 1980s flirting skills and spend 20 hours making a thoughtful mixtape.

I would even argue that the memory of "the album" is some instance of the Survivorship Bias. We fondly remember the great albums, forgetting the large amounts of bad or average albums from the past.

Yeah, the "extended single" albums, that had the one good hit song and 11 crappy tracks, will not be missed.

I'd say that's more of a genre problem than a streaming problem. Metal and less-commercial rock styles still have the album model as king. I hear hiphop tends to follow that too.

I'm not sure there were many of these to begin with. I love albums like that, but I think most seemingly coherent "pop" albums of the past followed basic rules of upbeat/downbeat flow with a very loose "common" theme. That can be accomplished with a simple mix tape, doesn't really matter who does each track.

> it's been harder and harder to find a truly coherent album

Not really. Just look through stuff from the 1970's.

I guess the album is dead

Not at all true, many of the new rap scene has some very cohesive albums. Brockhampton saturation 1-3 have very clear beginning and ends, JPEGMAFIA's new album is a clean 42 minute experience. The album is not dead. Its just the audience listening to it

Edit: This expands past these two bands, I'm giving two small examples in a large field that I listen to. I think artists are leaning toward either singles that get tons of airplay or albums that get listen time drawn out because the album is cohesive. Its two different styles of listening,and both are being catered. You can have a one off single that one group of people love, and another group that can't listen to that single outside of the album. Streaming services have done a good job at allowing this to happen.

Major label guy here. Not quite. While this isn’t as important for a lot of artists, on the top (like megastar-level: Drake, Post Malone, etc.), it’s still a huge money maker. More tracks means more streams which means more revenue. The waterfall is effective, but the album is still important. Especially in the release week.

What about cohesion in the album? I guess that prog-rock/indie/hiphop themed albums thing isn't very common in general these days, I'd be surprised to see one today from any new artist.

The intros and skits would never fit into streaming properly, especially with the ranking algorithms. Nothing worse than getting some random intro skit come up on a recommended playlist!

Spotify should run some detection algorithms on songs to not recommend the intros/skits, or the uploader can explicitly mark them for exclusion, which would help boost their singles and not push unhelpful songs

I'm a big fan of artists releasing each song as it's made, with each having cover art and maybe a backup song. They . can still package them up in a full album for charting and full album sales.

Spotify uses peoples manual playlists to drive some recommendations. If two songs are on a playlist together they’re probably similar. In the same way, if a song isn’t on any playlists, like you’d expect a skit not to be, then they don’t recommend it.

Free human curation of stuff that goes well on playlists.

Perhaps, but I still think we're a long way away from, say, something like "Dark Side of the Moon" being a viable approach anymore. I don't miss albums so much as cohesive, long-form genuinely artistic statements... (Fancy way of saying concept albums).

Only if you listen to Top 50. The metal world is still going strong with album oriented albums

Metal and rock both put out some great albums still. And I still look forward to those releases to just experience half an hour of (mostly) new content all at once.

OTOH, Taylor Swift also just released a new album. So there's definitely still albums being made in the "top 50" as well.

With pop music, it arguably has been for some time.

Could we see much longer album-esque songs in the future?

AFAIK all streaming services pay per play (of more than 30s). So no, they incentivise the exact opposite of "longer album-esque songs". A single 30mn song would earn the same as a 2:30 song, and would be unlikely to have the early catchy hook which ensures it passes the 30s mark for a payout.

The most corrosive part of this that affects the quality of the music is this:

...songs are becoming shorter, and artists are front-loading all the catchy bits to keep a song’s skip rate as low as possible

As music consumers become more single-oriented and running an attention deficit, everthing is getting hook fronted in the first 30 seconds of the song to keep people from skipping. Now, that might be interesting for a while if it makes artists experiment with verse-chorus-verse pop song structure but I'll bet the long term effect will be a race to the bottom in terms of lack of complexity and a lot of pop music is going to sound samey very fast.

Nothing new. Korn mentions just that in their “Y’all want a single” [0] - back in 2003.

The text at 2:32 says “90% of singles get to the ’hook’ within 30 seconds”. Then: “98% of #1 hits are less than 3:30 long” and “does this sound like a formula to you?”

[0] https://www.youtube.com/watch?v=AyFlW_bOg-Q

And what's funny, is that originally, this 3/4 minutes formula is likely more tied to physical constraints than marketing/human psychology: https://www.vox.com/2014/8/18/6003271/why-are-songs-3-minute...

Which reminds a bit of http://www.astrodigital.org/space/stshorse.html as well.

> Which reminds a bit of http://www.astrodigital.org/space/stshorse.html as well.

TL;DR - the legend that the width of the shuttle SRBs is derived from a design decision about cart width that goes back to the Romans.

But it's actually a bit more complicated than that...


Excellent, thanks for the link!

Wow that last link is amazing. I love that kind of stuff!

Yep, funny thing is the song itself follows all the rules for a single and it's a banger because... well, it's a full fledged single.

oh i guess i sorta took this for granted, this is how I've been releasing music for the last decade or so. just make a song, upload it, and then compile conceptually related songs into an album later when enough material has been accrued. granted, I've only released about 7 albums so far but I have 8 more planned to be released in the near future when I can get around to making album art for all of them

Well, Melanie Martinez has just released an entire concept album (and a movie based on this album) with no singles all at once. There will always be artists who care about their art and not streaming numbers.

I love this line

> Plus, she says, streaming means listeners get to pick who’s worth hearing. “It’s not a bunch of white males at radio stations and record labels deciding what the general public should listen to.”

This article is mostly about the release schedule of music (TLDR: one song at a time, once every few months), saying very little about how the actual music has changed. For more on this, music journo Liz Pelly has written a lot about what she calls "Spotify-core", a new-ish genre of music which works as both background and foreground music. https://thebaffler.com/downstream/streambait-pop-pelly

My circle of experimental music artists were doing this in the early/mid-2000s.We weren't streaming so much as releasing MP3s as they were done.

I expect (somewhat) that song length may also morph, if it hasn't already. There's no need to be concerned about how much time fits on physical media. There are just attention spans.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact