Analyzing the case against “Blurred Lines”

Update (Mar 13 9:15 AM): This article is now somewhat out of date; with the verdict came reports that these melodies were not the only things compared at trial. Stay tuned. In the mean time, this analysis is still accurate. 

Update (Mar 13 4:30 PM): There is now a part 2.

Court has been called back into session in the Blurred Lines plagiarism lawsuit, and as far as I can tell, the plaintiffs are currently making their case. Gaye’s camp does not have much to go on, since the judge ruled that the full recordings of the original songs could not be played in court, but their case looks stronger than I thought it would. According to Kate Conger at Ratter, the musicologist hired by the Gaye family presented the following slide in court last week:


With this slide the musicologist puts both melodies in the same context; and proves that they share the following elements:

  • a: a melodic fragment of three repeated notes (but not on the same scale degree)
  • b: a melodic fragment consisting of the 5th, 6th, and 1st scale degree in upwards motion in eighth notes
  • c: a rhythm of six repeated eighth notes preceded by a rest
  • d: a melisma (a phrase where one vowel sound is continued over more than one note) that contains a descending fourth (the second b)

Although this is a much more compelling case than I was expecting, these elements are far from novel, and I can’t see how this case could be won by the Gaye family without opening the door to a wave of lawsuits and copyright trolling. Take those a and b melody fragments of Got To Get It Up, for example, notated here as “5-5-5” and “5-6-1.” Together they are also the melody of the opening phrase of the chorus to Janet Jackson’s song Runaway:


They’re also basically the opening piano melody to Smokey Robinson’s You Really Got a Hold On Me:


(*In this case phrase a is reduced to one note on the 5th scale degree, however phrase b continues onto the 2nd and then 1st scale degree, as is the case of Got To Give It Up.)

That eighth note rhythm indicated as is also extremely common,which makes me wonder why it’s even being considered as part of the lawsuit. Here it is in Beauty and The Beast:


Here it is in Old Time Rock and Roll:


And in the opening phrase of Beethoven’s Waldstein Sonata:


So even though the musical material labeled a, b and c is all relevant to the case, I don’t think there is enough here to warrant a verdict in favor of Gaye’s family. These are common melodic and rhythmic ideas. I have only been working at this for about an hour and have already came up with one example that is an exact match for the Gaye a, b and c; one match for b and c that contains a fragment of a, and three rhythmic matches for c. This can be contrasted with the Sam Smith / Tom Petty plagiarism case, where my $180 prize for a third exact match that predates those songs has yet to be claimed despite a response from my friends that would lead me to estimate at least 15 man-hours have been spent looking for one.

You may be asking “what about the musical material labeled d and the second b?” The fragments that the musicologist is labeling d and the second b are irrelevant, I think; melismas are frequently improvised, and besides, they are different melismas. You can’t copyright the broad idea of including a melisma that contains a descending fourth.

Bright Tunes Music v. Harrisongs involved two recordings that had the same three- and five-note melodic fragments. In his ruling the judge in that case decided that using the same three- and five-note fragments on their own was not novel enough to be copyright infringement, but that “the four repetitions of A, followed by four repetitions of B, is a highly unique pattern.” So the copyright was violated not because a few notes were the same but because each of those fragments were repeated in the same way in each song for a total of twenty-seven shared notes in a row between the two songs. This is clearly not happening with Blurred Lines; there are only three (arguably five) notes that match each other exactly, and they are not repeated.

All things being said, this case is a lot closer than I thought it was going to be, and the Gaye family’s chances depend on on the jury’s perception of all of this. We are in fairly murky waters, but based on my understanding of our copyright laws and the precedent set by Bright Music v. Harrisongs, this is not a case of plagiarism. If the jury does rule in favor of the Gaye family, things are about to get very interesting. The Gayes’ musicologist has chopped the music up into pieces that are so small that it would fundamentally change the process of songwriting if they won. If three common notes within a single similarly-shaped phrase is all that is needed to successfully sue someone, then the floodgates of litigation are about to swing wide open. Stay tuned.

Spotify? Not much better than piracy.

Spotify claims that it is better than piracy and that it generates revenue for recording artists. But is it? And does it?

If you are a frequent Spotify user, you’ve probably heard the ad that says “Piracy is so last year. Every time you listen to music on Spotify, you make money for the rightsholders and artists.” If you haven’t, you will;  it is aired about once every two hours.

I am a frequent Spotify user, and I think it’s a fantastic service, offering free access to an ocean of recorded music, including recordings that are hard to find in record stores. It has revolutionized the way that people consume music and will likely be around for a long time.

But Spotify is a young product and a young company, and Spotify’s claim that it generates revenue for artists deserves thorough scrutiny. Does listening to music on Spotify really help support your favorite artists? Is Spotify really a better distribution channel than the best illegal filesharing services?

The short answer to both of these questions: Sort of, but only barely, at least right now.

Until Spotify came around, the best digital music distribution service had been Napster,  the short-lived filesharing service that let users share their entire digital music library – for free – with everyone who was logged in at the same time. This provided users with a virtually limitless music library. It was wild. At one point, I had 22 days worth of music downloaded to my college computer, including rare Beach Boys outtakes, out-of-print jazz records, and new music that was not available at the mainstream record stores. But Napster did not generate any revenue for rightsholders on its own, and was successfully sued by the Recording Industry Association of America and shut down in 2001.

Napster’s longest-lasting impact on the music business is that it showed that online distribution of digital music was the future of the industry. Several online services came about in Napster’s wake, but it wasn’t until Apple introduced the iTunes Store that consumers had a stable, safe, industry-sanctioned online music distribution service that rivalled Napster in its ease of use and breadth of content. Like Napster, iTunes Store users had hundreds of thousands of individual tracks they could search for and download. But unlike Napster, a user had to pay about a buck for each download. The success of iTunes’ pay-per download model created a swamp of copycat services including Rhapsody, eMusic, and a pay-service run by Roxio that was given the Napster brand after Roxio aquired it in bankruptcy liquidation.

Spotify operates as a bridge between the two models. Users can have free, ad-supported access to stream the entire Spotify catalogue, or they can pay about $10 a month to be able to download an unlimited number tracks from Spotify to their computer. Tracks downloaded through Spotify can’t leave the Spotify application, so there’s no risk of them being distributed any further, and rightsholders are paid a nominal rate every time someone plays one of their tracks using the Spotify service. This rate is exponentially less than what would be received for one iTunes download of the same track, but Spotify still has the full support of the Recording Industry Association of America, meaning it’s legal and it’s probably here to stay.

To give you an idea of Spotify’s payout versus iTunes and similar services, here’s the first revenue report for Don’t Punch Your Friend in the Head, a song my band Ramforinkus added to several internet music services last April.

The report above shows the different revenue that 28 sales and 104 streams generated across five different services. Not a great start, for sure, but not too bad considering we hadn’t really promoted the song yet. Since the band doesn’t have a record label or any other stakeholder taking a cut of the revenue, we receive the entire payout from each service:

iTunes (UK/EU)     ~$1.22 / sale

iTunes (US)               ~$.76 / sale

eMusic                            $.40 / sale

Rhapsody                      $.01 / stream

Spotify                            $.00378 / stream

So if you live in the US and purchased a copy of Friend in the Head on iTunes for 99 cents, we would receive about 77% of the sale. This could add up pretty quickly; if we end up selling 9,900 copies of our song on iTunes, we’ll receive about $7600. That would finance a short tour for sure. 

Spotify has a completely different business model than iTunes in that it is a streaming music service, not an online store. As such, it pays rightsholders based on individual plays, not downloads, and its current rate is a little less than four-tenths of one cent per stream. That means an individual user would have to stream our song two hundred times in order for my band to receive the same 76-cent payout that would come from a single purchase on iTunes. This won’t add up in any meaningful way until our song receives hundreds of thousands of streams, and my guess is it is unlikely that someone would listen to any individual track by any independent artist on Spotify enough times to match the revenue of a single iTunes purchase.

For the record: Four-tenths of a cent per stream is a marked improvement over the way artists were initially compensated by Spotify; Lady Gaga famously received only $167 after her song “Poker Face” was the first track to hit 1 million streams on Spotify in 2009, which amounted to 1.67 thousandths of a cent per stream.

This figure – .38 cents per stream – shocks a lot of people at first because Spotify’s entire business model rests on the fact that it’s supposed to generate revenue for rightsholders, and they vigorously refute the idea that they are anything other than the next great revenue source for artists. In a recent interview, a Spotify spokesperson said “Spotify is now generating serious revenues for rights holders; since our launch just three years ago, we have paid over $100 million to labels and publishers, who, in turn, pass this on to the artists, composers and authors they represent. Indeed, a top Swedish music executive was recently quoted as saying that Spotify is currently the biggest single revenue source for the music industry in Scandinavia.”

That claim is certainly impressive, but there’s a serious disconnect here between what Spotify says and what artists end up receiving per-play on the service. Most major labels keep a considerable cut of that less-than-four-tenths-a-cent payout, and independent artists who release their own material generally don’t attract the number of listeners it would take to make Spotify a better revenue generator than iTunes.

In fact, I might argue that fans who pirated Don’t Punch Your Friend in the Head could do more for us than had they used Spotify to listen to it. If 103 people had pirated Friend in the Head instead of streaming it through Spotify, those 103 people would possess a digital copy of the song, which means there are dozens of extra ways they could pass it along to others. They could burn it to a mix CD; They could listen to it using any audio player, in any setting, and could easily transfer it between devices; They could convert it to any format, they could add it to a video, they could sample it, they could make it their ringtone, they could play it as house music at a theater or club, and they could broadcast it on their radio show. By listening to Friend in the Head in Spotify instead of pirating it, those 103 people could have shared the tune with other Spotify users and could have added it to their Spotify playlists, but there’s not much else they could have done that would help the song reach new people since music on Spotify can’t leave the Spotify application, and you can’t take the Spotify application with you without paying $10 a month for it. So as an artist in several independent bands, I don’t see how Spotify is really better for us than an equally accessible piracy channel would be.

I don’t mean to complain about Spotify as a service. Spotify represents the future of music consumption. We are a culture of convenience, and nothing is more convenient than immediate, cost-free access to every song you’ve ever wanted to hear.  And I was always okay with small-scale music piracy, whether it meant taping songs off the radio, ripping a library CD to your computer, or making mix CDs for friends, so I am not writing this article to complain about music consumers who do not pay for all of the music they listen to. I am not only happy that Spotify is an option for my own music but I also use it several hours a day, both for fun and for work. The problem, as I see it, is that Spotify boasts that they have surpassed piracy and that you, the Spotify user, are supporting artists by using their service. I find this hard to swallow because they’ve released no data or projections to support these claims and their defense of their payout model is always presented in general terms. Spotify should not be advertising that their listeners generate revenue for artists; this could lead to someone choosing to listen to a song on Spotify instead of purchasing it on iTunes with the idea that they were supporting the artist either way. In reality, it takes nearly 300 Spotify streams for one member of an independent band to be able to buy a cup of coffee at a bodega, and probably many many more times if an artist has a record label.

If Spotify wants to change my mind, they’ll need to publicly release their payout model, and I’ll need to see concrete, verifiable data that can confirm that Spotify users listen to individual songs at a rate that will eventually surpass iTunes revenue. Until that happens, I will continue to look at Spotify as a free music service, like Napster and Grooveshark before it, and it should be known that the best way to support your favorite artists is to purchase their music at full-price – especially their self-released material – and to go to their shows when they’re in town.

Update (12/2012): In response to a few comments, I’ve expanded this article and have clarified some of the language.