I’m going to miss Soundcloud


It seems that Soundcloud is about to disappear: the sound-sharing website has only a few weeks, possibly months, of money left in the coffers and, if current reports are to be believed, once that has run out it will disappear. There are reports of backups being made, by archive.org and even individuals, but time will tell how accessible and/or user-friendly those might end up.

On NewMusicBox today, bassist Gahlord Dewald has posted an overview of why musicians might share their music, and what other services, besides Soundcloud, might serve people in the future. I really just want to add to that to say what value Soundcloud has had for me as a writer/investigator into new music over the last few years.

And that is: enormous. When I first noticed a few years ago that composers and performers were putting their works up on Soundcloud it was a tremendously exciting moment. Until then, it had been possible to access bootlegged new music, live recordings and so on through chatrooms and personal contacts; but from a research point of view it was a laborious process that required a certain amount of pleading. Now with Soundcloud – and for some reason this seemed to be the breakout platform everyone was using – I could search for things proactively, at my own pace and according to my needs. Asked to write a programme note about composer x? Chances were, if she was under, say, 45, I could find a bunch of her work on Soundcloud. Wanted to explore who was on this year’s Gaudeamus shortlist, or currently at Schloss Solitude, or making waves at Darmstadt – again, Soundcloud. When Riot Ensemble ran its most recent call for scores, the overwhelming majority of our 279 applicants had posted their portfolio works on Soundcloud. For a generation of composers, I got the sense that Soundcloud had become a default setting – and in that respect it was becoming a transformative technology for the visibility and reception of new music, and especially that by composers too young or too weird to have record deals or broadcasts. This was undoubtedly new, and very healthy. Sites like Soundcloud have made it easier to know what composers in their 30s are up to these days than composers in their 50s or 60s, who may be locked into more traditional modes of dissemination for their work.

Now, when Soundcloud is gone no doubt something will arrive in its place. Still more likely, though, several things will arrive at once. And some will already be here: Bandcamp is covering some of that territory, and I’ve even heard talk of retreating back to MySpace. And this will mean fragmentation across platforms, with all the inconsistencies, annoyances and break-ups of putative communities that that entails. You can’t follow a thread of likes between platforms, for example. You can’t easily curate a playlist of recommendations.

Soundcloud wasn’t perfect, and there wasn’t anything inherently special about its offering. (Although I did like its feature that tracks would continue to play even if you clicked to a new page. This seems so intuitive it always surprises me when it doesn’t happen on other sites.) But it had become something of a norm, a standard. And when those disappear everyone will be back to square one. I’m going to miss it.

Update: … If, that is, those reports can be believed. Since I posted this, a Facebook reader alerted me to this post on the Soundcloud blog, from 14 July and written by Soundcloud co-founder Alex Ljung, which claims that ‘Soundcloud is here to stay. … The music you love on SoundCloud isn’t going away, the music you shared or uploaded isn’t going away, because SoundCloud is not going away. Not in 50 days, not in 80 days or anytime in the foreseeable future. Your music is safe.’ I’m still wary, especially in the fragile world of Internet economics, that there’s rarely smoke without fire, but let’s hope this post’s claims are true.

Spotify just got a whole lot easier for classical listeners

One year ago, practically to the day, I posted this picture of what it looks like to search the complete Haydn symphonies on Spotify and lamented

“Please: we’ve had digital music for nearly two decades now. Can we start to get our act together on this?”


I mean, that list of results is basically useless. The legacy of a digital music tagging system that is designed for songs and albums, not works and movements.

Well, thanks to a tip-off from Ulyssestone (now on Spotify’s staff), whaddya know – today, the same page looks like this:


Good work everyone who made that happen. This is more like it.

The news gets better: slowly but surely, composer names are being added to the database too. Take a look at this image (from Ulysses’ blog):

Now we really are getting somewhere. It’s not ideal, sure – where there’s a second performer it’s not immediately clear which name is the composer and which is the performer. And the composer names thing only applies to Naxos-distributed labels so far – but that’s tens of thousands of albums already. However, this is definitely progress (from a very poor starting position), and it’s good to know that people are at least working on this stuff. Before long it will be possible to do a classical search on Spotify and reliably be able to find what you were after. Imagine.


Does Spotify pay? Another look at the numbers

Despite this blog’s basic remit to cover contemporary classical music, one of its most popular posts has been ‘How much do musicians make on online?‘, a quick analysis of a graph published by Information is Beautiful about the relative remunerations of different ways of selling music.

That graph is widely-known, but it’s also four years old now. And since it was produced, Spotify have opened up a lot about how much they pay artists. It turns out that the Information is Beautiful graph was wrong on how much Spotify streams pay by around a factor of 30.

As Spotify themselves now point out, thinking of their royalty rates on a per-stream basis is a bit misleading anyway, since they make their calculations based on a percentage of the total revenue pie. So the more subscribers there are, the larger the effective royalty rate. But back in July 2013, when they started to release this information, they suggested that a figure of between $0.006 and $0.0084 per stream was not a bad basis for calculations. As their subscriber base grows, so that figure will go up.

Anyway, apropos of not much, I thought I would use these numbers to do a quick tot-up of how much money two famous Spotify objectors – Radiohead and Metallica – might have made from the service. I took the figures for number of plays given for the top ten songs on the artist’s Spotify page, and multiplied them by both the low and high estimations of what Spotify says it pays out on average per stream.

Bear in mind that these numbers are not externally verified – they’re what Spotify tell us is going on – but they are worth considering in the context of some recent debates over the long-term viability of paid-for streaming.


Thom Yorke has called Spotify ‘the last desperate fart of a dying corpse‘, and in October last year pulled his Atoms for Peace album, made with Radiohead producer Nigel Godrich, from the site. Yet here’s an idea of how Radiohead are actually doing out of Spotify:


That makes 107,302,714 plays (as of Friday 28 March). Based on Spotify’s estimated per-stream pay out, that’s somewhere between $643,816 and $901,342. OK, computer.


Despite dragging himself through the mud for a decade over Napster, Metallica’s Lars Ulrich made peace with online music in December 2012 and Metallica’s back catalogue came on stream. In fact, in the same month that Yorke and Godrich made their comments about Spotify, Ulrich claimed that ‘Spotify is working right now‘. The delay in joining probably accounts for Metallica’s fewer plays, but the numbers seem to bear him out:


That makes 67,151,066 plays (as of Friday 28 March). Based on Spotify’s estimated per-stream pay out, that’s somewhere between $402,906 and $564,068. Sad but true.

None of this is a definitive answer to the ongoing future of music debate, but I think it’s useful to see figures like this while that debate is being had.

One day the classical music industry will wake up to ID3 tags and its mind will blow

I just tried looking for Haydn’s Symphony no.73, ‘La chasse’ on Spotify.

Being a contrary type, and knowing that searches for symphony + no. very rarely narrow the field, even for a number as high as 73, I thought I’d just pull up Haydn and flick through a few album covers instead.

Oh look, here’s Antal Doráti and the Philharmonia Hungarica performing the complete symphonies on Decca. Perfect – just click and scroll until I hit no.73.



There are 425 lines of this. Around four times that number if you include all the ‘additional tracks’ that Spotify lists – similarly without any identifying features.

It’s word soup: an endless stream of tempo indications, with barely anything to attach them to one particular symphony. I’m listening to a minuet and trio at the moment – God knows which one. The information is completely meaningless, completely unusable. And, if I had searched for “Symphony no.73”, or even “La chasse”, unfindable too.

This a particularly shaming example, but this kind of metadata wastage happens all over digital classical music. Even in new music, where pieces tend to have unique titles, it can be almost impossible to find things that you know are there using first-time search terms. There’s one album I know where each track is simply identified by the surname of its composer. No titles at all. And there are too many others where the composer’s name doesn’t appear anywhere in the tagging. Who browses anyway, right?

Please: we’ve had digital music for nearly two decades now. Can we start to get our act together on this?

(NB I have just found that “La chasse” – on disc 21 of 33 of the Decca set – is actually one of few that can be found by searching by its title. But I think my point still stands.)