Spotify’s latest patent could let users mix their favourite songs
Spotify may soon let fans create seamless mashups with AI, a new patent hints at smart tools that blend songs together automatically.
Spotify has secured a patent that could reshape how we experience music mashups. The filing, first unearthed by Music Business Worldwide, describes a system capable of analysing songs and combining them into brand-new mashups tailored for individual listeners.
The technique works by separating songs into their component parts (vocals, instrumentals) and then layering vocals from one track onto the backing track of another, provided the two are deemed “musically compatible.” According to the patent, Spotify’s algorithm would measure compatibility via factors like key, tempo, harmony and beat alignment.
In one example, you might select a “base track” (an instrumental or full song) and the system would suggest other tracks whose vocal components could overlay cleanly on that base. The patent calls these evaluations “vertical mashability” and “horizontal mashability,” referencing how well tracks stack up (vocal over accompaniment) and flow sequentially.
The patent also mentions generating custom titles for the mashups (by combining words from original song titles) and even bespoke album art that blends imagery from both tracks.
Instead of recommending existing songs, the idea is to create new content tailored to a user’s tastes, potentially unlocking new degrees of engagement. The mashup tech aligns with wider trends in generative audio, customization, and fans wanting more uniqueness in their listening.
However, there are big questions, particularly around rights. Even Spotify’s own spokesperson stated the patent is “unrelated to” its recently announced “artist-first” AI initiative, and that “some patents become part of a future offering, while others don’t.” That suggests this might remain a technical asset rather than a consumer feature, at least for now.
Ultimately, Spotify is staking a claim in the next frontier of music tech, not just recommending songs, but crafting them.