Pex is the industry expert at identifying modified audio, and has released multiple data reports on the rampant proliferation of modified audio on UGC platforms and DSPs. Billboard also shined a light on modified audio, reinforcing what Pex had unearthed: hundreds of millions of modified audio tracks have been released as remixes. What does this mean? Modified audio rarely (if ever) ties back to a song’s proper rightsholders, label or publisher, and so when these tracks are distributed and released on DSPs, any royalties they generate end up being paid to the wrong parties, and the proper rightsholders get nothing. This lost revenue is fraud. Because of the DSP pay models, including Spotify’s recent changes that now require a threshold before paying royalties on sound recordings, modified audio is able to steal significant streams and divert revenue from the proper rightsholders.
These models make it critical to count all listens, even when they are modified. To illustrate the revenue impact of modified audio, and why platforms and distributors need to protect artists from this issue, here’s a breakdown of Spotify’s pay structure and how the most recent changes impact modified audio streams.
Streamshare model vs. royalty-based payouts
In a traditional royalty-based system, payments are made based on a fixed rate per play or per sale. Spotify and other DSPs, however, use a streamshare model. This means that the total payout to rightsholders (such as record labels and independent artists) is determined by their share of total streams on the platform.
Spotify’s royalty pool and payment structure
Spotify generates revenue from two primary sources: Spotify Premium subscribers and advertisers on Spotify’s Free tier. Nearly 70% of this revenue is paid out to music rightsholders, which is referred to as the “royalty pool”. The distribution of this royalty pool is based on each rightsholder’s streamshare on Spotify. This means that the total number of streams in a given month determines what proportion of those streams were for music owned or controlled by a particular rightsholder.
Factors influencing payouts
Several factors influence the payout for sound recordings on Spotify:
- User Subscription tier: Streams from Premium subscribers are worth more than streams from Free tier users due to the difference in revenue generated from these two groups.
- Geographical Location: The value of streams can vary depending on the country, as subscription costs and advertising revenues differ across regions.
Spotify’s payouts are not a simple multiplication of streams by a set royalty rate. Instead, Spotify calculates the payout by dividing the total revenue generated from subscriptions and advertisements by the total number of streams in a given period. Then, they distribute this amount based on an artist’s or track’s proportion of total streams.
The 1,000 stream count threshold
In December 2023, Spotify announced that in the coming year it was introducing a new model for how it calculates recorded royalties next year. Starting in January 2024, Spotify introduced a new policy where tracks must have reached a threshold of at least 1,000 streams in the previous 12 months to be included in the recorded music royalty pool calculation. This policy aims to prevent manipulation by ensuring that only tracks with a significant number of streams contribute to the royalty calculations. However, the new model does not account for modified audio or counterfeit recordings pulling legitimate streams from the legitimate authorized releases, making identifying and removing these counterfeits even more crucial.
Counterfeit sound recordings can create royalty dilution
Counterfeit sound recordings, which are modified versions of audio tracks, dilute the royalty pool by adding low-quality or fraudulent content to the platform. These tracks take up a portion of the royalty pool, without contributing to the genuine music ecosystem, and divert royalties from the rightful owner. When a counterfeit sound recording Is distributed it can replace the authorized release of the same song in listener activity, artificially deflating the streaming numbers of the legitimate songs, and fraudulently diluting the revenue pool. In the new model, that means to qualify for any royalties, a sound recording is in competition with itself to reach the minimum stream threshold. This makes it more important than even before to find these bootlegs and root them out.
When counterfeit tracks generate streams, they take away from the royalties that would otherwise be allocated to the legitimate artists and rightsholders of those songs. In the new payout system, counterfeit tracks of legitimate songs can negatively impact an artist if they do not meet this threshold, as they end up competing with their own works for the streams needed to reach the threshold. If they lose out to their own counterfeiters, they would not be eligible for payouts, reducing the legitimate revenue generated from the song.
A band from Illinois, Old Capital Square Dance Club, released an album in 2019. Recently, the album gained unexpected attention due to a fraudulent scheme involving artificial intelligence. A band member discovered that another unrelated artist, “Marico Charlotte” had bootlegged their music, and released it as his own The fake tracks, such as “The Loose Change,” (barely retitled from the band’s song “Loose Change”) were nearly identical to Old Capital’s songs but had been slightly altered to evade plagiarism detection. This manipulation involved speeding up the tempo and pitch-shifting the vocals.
Further investigation revealed that the entire “Marico Charlotte” catalog consisted of similarly altered tracks from various artists. Another musician and recording engineer from St. Louis found that these tracks were part of playlists created by a user named “Nathaniel Yee,” a likely pseudonym. These playlists contained music from artists with names that seemed generated by a chatbot and featured numerous pitch-shifted, sped-up tracks. The engineer identified six similar playlists, each with about 30 artists, suggesting a widespread scheme.
Despite Spotify’s efforts to combat such fraudulent activities, the platform’s algorithms were fooled by these minor alterations. This has allowed the perpetrators to divert significant revenue from legitimate streams, which is likely lost forever. The incident, not an isolated case, underscores the challenges music streaming services face in detecting and preventing copyright infringement and fraudulent activities.
Pex Search for modified audio
At Pex, our AI-powered content identification technology is superior at identifying modified audio, especially sped up remixes that dominate social media and DSPs. With Pex Search, anyone can leverage our technology to identify modified audio in real time. DSPs or music distributors can plug in to detect modified tracks before they are distributed or rack up streams. Rightsholders can drop in their recordings to see all known uses of the song, including modified tracks on DSPs, so they can combat lost revenue or pursue licenses. Check out Pex Search here to see how our audio and melody identification technologies find modified audio in seconds.
