There is a little (understandable) confusion in some of the comments here.
The issue here is not that Spotify is infringing copyright (so it’s nothing to do with BitTorrent etc) but rather that Wixen says that it has not been paid correctly for the mechanical exploitation of the copyrights it administers. This is complex for several reasons.
First of all: the recordings in question have without doubt been legitimately supplied to Spotify by the record labels of the artists in question.
Secondly - a record label controls only one of the two copyrights in a song - the recording copyright.
It is impossible to record a song without using the other fundamental copyright in music - which is the song itself. The literal words and music.
This is the copyright that the publisher controls - and which Wixen claims has not been properly accounted.
Third point: When the words-and-music copyright is reproduced by mechanical means for commercial gain the publisher is due a royalty - this royalty is known as “mechanicals”.
In the “old” music business when a label pressed a record they paid out the “mechanicals” to the publisher directly - for every record pressed they owed a set amount. So this was easy to track and it meant that the money flowed directly from label to publisher based on volume of records manufactured and sold.
Point number four: Spotify’s licensing regime is complex. They need to pay a royalty for the use of the sound recording. This generally goes direct to the label or artist if they self release often via their distributor).
The mechanical reproduction of the words-and-music copyright is much less straight forward in terms of how the money flows.
The reason for this is that songwriters and the performers of a song are not always the same person - though they may also be the same person. This is where publishers come into play. If I write a song, I may not be able to record and perform it sufficiently to generate good revenue from that work. So if I can persuade a high profile artist to perform and record it I will probably make much more money from my words and music. A publisher’s job is to maximise the commercial exploitation of my words and music. This can be done in several ways - first of all, they go and shop my song around to labels in the hope that an A&R at a label (basically someone a bit like a product manager in a tech company) will see a good fit between my song and an artist that they work with. The artist records the song and I generate money in two ways - first of all the mechanicals previously discussed, but also from “performance” royalties - which is where my song is performed live to an audience. So a prominent artist performs my song on a tour - the audience in the arena have paid good money for tickets, and the artist would not have any material without my song and other songs. This also needs to be compensated, and this is done through a “performance royalty” which is generally a split of Ickes revenues.
There are other revenue streams as well, and a publisher’s job is to manage and administer these revenues as effectively as possible.
Now: I mentioned Spotify has a complex licensing model. When you play a song through Spotify it is being mechanically reproduced (the data that makes the audio waveform is being transmitted from one machine to another) but the audio is also being performed by the Spotify application on your computer. So Spotify needs to pay out a mechanical royalty and a performance royalty.
It used to be that the mechanical royalty was paid out by the label - but as the label is no longer pressing discs the “replication” of the music has passed to Spotify, and so Spotify is liable.
In the “old” music business once you bought the record or CD the revenue flow ended: your CD player manufacturer was not continuing to benefit each time you played a disc. So the “performance” of the music did not have a revenue stream attached to it, assuming that you were enjoying the music in the privacy of your own home.
But Spotify can be seen to benefit each time you play a song, because you pay an ongoing subscription for access to the music.
So a performance royalty is also due.
Now: further complexity. In many territories songwriters and publishers assign the management of performance and mechanical royalties to third parties. In the U.K. for example a songwriter will join PRS - the performing right society, which administers performance copyright for the words-and-music copyright in a song. They should also join the MCPS - the Mechanical Copyright Protection Society - to collect mechanical royalties. These organisations and others like them in other territories are known as Collective Management Organisations or “CMOs”.
Spotify has deals with these two organisations (PRS/MCPS) - and generally also with the equivalent CMOs around the world. Spotify says “this song was streamed X times and so we owe Y for mechanical royalties and Z for performance royalties. This money is then paid trough the CMOs to the publishers/songwriters. You cannot - as a publisher or an artist - get this money directly from Spotify.
This works on one level because it means that so long as a work is correctly identified in the database of the CMOs and that whoever is performing or mechanically reproducing a song submits accurate data showing how and when hat song was reproduced and what revenue is attached then it all works perfectly. More critically the CMO mechanism makes it possible for services like Spotify to exist (and even for live concerts to happen) because otherwise anyone who wanted to put on a concert would need to obtain the permission of every songwriter individually before their songs could be performed.
However - you need to keep in mind that this system was invented when sheet music and clockwork pianos where the main way that words-and-music copyright was mechanically reproduced, and performance was limited to someone standing up and singing the song in a concert hall.
So far, so complex.
To add a further few layers of complexity:
There is no authoratitive database that says “this words-and-music is the copyright of Tom Petty”. Additionally, when a record label ingests music into the Spotify catalogue there is no requirement to specify who wrote the words-and-music.
The way it works is that (in essence) Spotify sends the CMOs a list of everything that has been played and the CMOs say “I represent that words-and-music copyright in X territory - so you owe me $X representing Y streams.”
It’s a pretty clunky system - but as yet, no one has really come up with a better solution. Technology moved faster than copyright admin.
Add into the mix that some artists may not register with the appropriate CMOs; some artists may not have a publisher; some labels may be releasing cover versions of songs where they have not directly obtained the permission of the copyright holder - but this is ok because in theory the words-and-music royalties flow from Spotify through the CMOs to the copyright holders.
So it’s a very complex case and hinges on whether the relevant CMOs had been delegated authority by songwriters/copyright holders/publishers to exploit those works and whether Spotify was accurately accounting and paying out for the use of the copyrights.
So it’s not about BitTorrent or Spotify having pirated music and is everything to do with how technology, the exploitation of copyright and how money flows through music are all a bit out of sync with each other.