Spotify’s Voice Analysis Technology May Soon Suggest Music Based on your Mood
Spotify has been granted a patent for new technology that can suggest songs to users based on the analysis of their voice data. The global audio and music streaming giant had first applied for this patent in February 2018, and it was granted by the United States Patent and Trademark Office (USPTO) this month.
According to NDTV Gadgets 360, the patent outlines tech that will be able to identify diverse personal traits like “emotional state” of the users, and suggest content based on this model. Spotify was earlier granted another rather unsettling patent that involved curating a user’s Spotify experience based on their personality traits.
The tech website reports that the patent titled “Identification of taste attributes from an audio signal” describes a method for processing an audio signal input from users – including speech content and background noise – to identify playable content based on the input.
In the patent description that is available on the USPTO website, the Swedish audio and music streaming giant explains that “it is common for a media streaming application to include features that provide personalized media recommendations to a user”. A common approach that Spotify employs to identify a listener’s music taste is by using basic information like age or gender.
The new patent takes it a step further and uses voice recognition technology to “collect the taste attributes of a user”. The patented technology aims to detect the “intonation, stress, rhythm, and the likes of units of speech” and combine it with “acoustic information within a hidden Markov model architecture” so that Spotify with be able to categorize the user’s mood as “happy, angry, sad, or neutral”.
Spotify’s patent details that the emotional state, gender, age, and accent of the speaker providing the voice will also be analyzed. The metadata collected by the technology will also indicate the number of people in the environment in which the audio signal is input.
All this metadata is gathered and analyzed along with a Spotify listener’s “previous requests,” “listening and rating history,” “links to associated profiles such as those of the user’s friends or colleagues”, and “the user’s existing music collection and/or library”.
According to the Spotify patent filing, the output will be the audio output of music corresponding to the preferences, or a display of recommended next music tracks corresponding to the preferences. The output could also be to simply provide a recommendation on a visual display.
While the new Spotify patent doesn’t come as a surprise, the effectiveness or the problematic factors that the new technology entails will only be known to the users if it hits the market.