Hacker News new | past | comments | ask | show | jobs | submit login
The economics of streaming is changing pop songs (economist.com)
51 points by RachelF 14 days ago | hide | past | web | favorite | 14 comments

The article describes what used to be called a 'radio mix' or 'radio edit': https://en.wikipedia.org/wiki/Radio_edit

The only difference is that today it is less common to have several mixes of the same song (which mostly existed to fill singles) and people will often stream the album version, so everything is a radio mix today.

I used to notice radio songs would be played slightly faster than the CD or an MP3 ripped from the CD I downloaded. Sometimes I woundnt like the slower tempo of certain songs. Now that I steam a lot I don’t notice it. Maybe the streaming services play the radio edits or, as you said, it’s much less common.

No, radio stations would pitch up songs (play them faster/make them more "upbeat") to get more ad time: http://pulsemusic.proboards.com/thread/58007

You'll know the difference between a radio edit and the "original" because one will be ~3-4 minutes and the other can be somewhere from 4-9 minutes for the DJs to mix with.

https://www.youtube.com/watch?v=pUjE9H8QlA4 - 3:51

https://www.youtube.com/watch?v=vN2Q62uMew0 - 7:22

I think you're talking about slightly different things because of the different genres. You can think of there being three flavors of a track:

* A "dance mix" is specifically designed to be played by a DJ at a club. It's designed for mixing — overlapping with the previous and next songs so that dancers can continuously dance through multiple songs. It has a long intro (often only drums and nothing tonal to avoid clashes between different keys) and a long outro.

* An "album mix" is designed for personal listeners. It can be as long or short as the artist wants because they can assume the listener wants to hear this exact song because they chose it.

* A "radio mix" is designed to be played by radio DJs. Artists want to get played as much as possible and radio stations want to have plenty of time for ads and avoid boring dead air, so the incentive is to make this really short and punchy.

Electronic dance music often only has a "dance mix", especially in the niche subgenres. If a track is popular enough, it might have a "radio mix" that's actually closer to an album version — its intended for personal listening with less need of a mixable intro/outro.

Pop and rock used to often have both an album and a radio mix. Now that radio is dying and randomized streaming is growing, those two are converging (which is what the article hints at).

Many of these, especially upbeat ones will have longer "dance remixes", often many, often radically different from the original. A lot of times all that's kept from the original song is some of the vocals or the hook. These are usually created well after the original produced version (hence "remix").

So the "original" version of a song depends a lot on the genre. Most original versions of rock and pop songs are either the shorter or middle-length one and the longer dance mixes and shorter radio edit (if there is one) come later. Most EDM songs are longer in their original for-DJs version and the shorter edit comes out later if the track gets big enough to get popular adoption.

Thanks. Between my screen cut in half for cookies accepting (with no decline) and an immediate paywall, I was going to skip this.

“User-centric system” is when I listen only to a small indie band whole month, they get all of my $9.99. As opposed to the current state where the chart toppers get 9.98 and my indie favourite gets 0.01. But

> Spotify’s former director of economics thinks it neglected to account for higher administrative costs

I don’t get it. How are admin costs higher if you simply change the formula? To a simpler one IMO.

I don't know if they're higher administrative costs, but it requires more trust in Spotify.

In the constant stream revenue model, you take the total revenue Spotify is paying out to artists, and divide it by the total number of streams, and that's the cost per stream. Multiply that by the number of streams an artist had, and that's how much money they get. You only have to trust three numbers from Spotify, and two of them are global numbers, so other people will look into them.

In the stream cost per user revenue model, you do that _for each user_, and then sum up each artist's revenue. That's a whole lot of math to be done. Now, computers can do that, but auditing this will be far more complicated.

Thank you, makes sense. I was wrong that the per-user is simpler. Yet, it’s only fair. I wouldn’t buy a Rihanna album, in the old physical model, why would she get get my money in the new one? And it’s probably very likely they already have the userId -> songId logs for recommendations and such

That 40 second intro to Where The Streets Have No Name that the article mentions was probably only on the album version and not the single, I'll wager. Songs that wanted to be played on the radio a lot also had to get to the point quickly.

And finish before people lost interest too.

By my count, 11 of the 27 songs on the "Beatles 1" album start with the chorus. Granted, in those days a short and sweet single was the goal so you had to cut to the chase. Seems like history is merely repeating itself

The article surprised me. I was forming an opinion that streaming has the opposite effect. When I work at my coworking office, they play a continuous stream of indie rock aimed at adults. Muted guitars, raspy whispers, for well adjusted Nirvana fans. It's like an algorithm found the balance between muzak and the sex pistols.

Reminds me of how long songs used to be trimmed to fit on a 45rpm record.

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