Spotify patents speech monitoring technology to recommend users new music – here’s what you need to know
28.01.2021

Spotify patents speech monitoring technology to recommend users new music – here’s what you need to know

Words by Will Brewster

The technology also aims to suggest songs based on your gender, age, accent or emotions.

Streaming giants Spotify have secured a patent for a technology to record the background noise and speech of its users in order to recommend them new music, a report by Music Business Worldwide has found.

Summary

  • Spotify has secured a patent to extract the ‘intonation, stress, rhythm and likes of units of speech’ from listeners.
  • The streaming giants filed for the patent in February 2018 and received approval on January 12, 2021.
  • The service have claimed that ‘our ambition is to create the best audio experience out there’ and that the patent may not be used in the future.

Never miss an update – sign up to our newsletter for all the latest news, reviews, features and giveaways. 

What does the patent involve exactly?

The new patent, which is titled ‘Identification of taste attributes from an audio signal’, seeks to use speech recognition technology to collect information based on the ’emotional state, gender, age, or accent’ of Spotify users in order to recommend them content, including music and podcasts.

The technology aims to do this by combining factors such as gender, age and location with audio metadata obtained by listening to the pitch, tract length and intensity of a user’s voice patterns as well as the background sounds of their immediate environment, such as ‘sounds from vehicles on a street, other people talking, birds chirping, printers printing, and so on’.

By proxy, the technology could also utilise this audio metadata to gauge insights into what kind of listening environment the user is in, including whether or not the listener is isolated or with other people.

Why do Spotify want this patent?

A closer look at the patent reveals an overview of how Spotify currently use a decision tree in order to tailor its content recommendation algorithm to each user, with the streaming giants stating that by eliminating human input, the algorithm will function far more effectively.

‘What is needed is an entirely different approach to collecting taste attributes of a user, particularly one that is rooted in technology so that the above-described human activity (e.g., requiring a user to provide input) is at least partially eliminated and performed more efficiently,’ a portion of the patent reads.

The patent also claims that content will be recommended to the listener after metadata has been obtained and analysed alongside the user’s listening history and pre-existing music library, with the recommendations either coming in the form of a visual display or a random song election.

So, what comes next?

Straight off the bat, it’s important to note that audio monitoring isn’t anything out of the ordinary for tech companies – Amazon, Facebook, Apple and Google have all been stung for secretly listening to your conversations in the past.

Spotify even touched upon this themselves while submitting their patent for approval back in 2018, stating that ‘It is common for a media streaming application to include features that provide personalized media recommendations to a user’.

It’s also worth highlighting that just because Spotify have received the patent, it doesn’t mean that the streaming service will integrate it into future products, as the service have noted in a statement to Pitchfork today.

Considering some of the other patents the service has had approved in recent months, it’s fair to say that Spotify are truly aiming to ‘create the best audio experience out there’ – however, it seems that whether it’s the best audio experience for us or them is yet to be determined.

Read the patent in full here via Music Business Worldwide.