From ‘All by myself’ to ‘Happy’, songs we decide to listen to often reflect our mood. Now Spotify thinks it can make that decision for us – by listening to our voices.
The streaming giant has been granted a patent for a technology that analyses your voice to determine your emotional state, age, gender and even the environment you are currently in – then suggest songs based on the results.
By listening out for background noises, such as birds chirping, printers or traffic, it can determine whether you are at work, out for a walk or even in an office.
As well as looking at your social setting the technology gets personal – it can listen out and decide whether you’re happy, angry or downright miserable, then select a song based on a combination of your social setting and mood.
No date has been set for when this technology might be included in Spotify, or even if it will be included – or how they will gather the voice data to make it work.
As well as looking at your social setting the technology gets personal – it can listen out and decide whether you’re happy, angry or downright miserable, then select a song based on a combination of your social setting and mood. Stock image
The firm wrote in the patent filing, granted on January 12, that current techniques for determining users music taste involve long, tedious form filing.
Instead of requiring someone to fill in a form about their age, gender and favourite bands, the technology uses speech recognition to gather the same information.
It is then able to take context clues from the voice including the tone, stress and rhythm and use that to determine a number of key ‘metadata characteristics’.
Spotify told Pitchfork it ‘has filed patent applications for hundreds of inventions’, adding they ‘regularly file new applications’, some of which become future products and others that don’t.
The technology takes in users previous song requests, listens for emotions and creates song suggestions based on the analysis
The audio input leads to content by analysing age, emotional state, gender and accent as well as background sounds via speech recognition technology
‘Our ambition is to create the best audio experience out there, but we don’t have any news to share at this time,’ the firm said in response to Pitchfork.
The patent outlines a few options for determining emotion state – including the very simple process of creating a list of emotions – happy, sad, lonely etc – then having songs associated with each emotion.
Other, more complex options, involve creating an emotion-tree and linking that to other information gathered about the user including their environment, age, gender and even their accent.
By listening out for background noises, such as birds chirping, printers or traffic, it can determine whether you are at work, out for a walk or even in an office
‘It should be understood that the above example metadata categories of emotions, gender, age and accent are merely examples, and numerous other characterizations and classifications can be used,’ the firm wrote in its filing.
Other information that could be gained by listening to the users voice includes background noises from ‘sounds of vehicles’ and ‘people talking’ to ‘birds chirping’.
A chirping bird could indicate a more natural environment, whereas the sound of printers printing could suggest someone is in an office or other workplace.
Combined with the users emotions, this could allow Spotify to automatically generate a playlist that reflects their current situation and state of mind.
The patent suggests this information could be used alongside other information provided by the listener including their previous requests, listening history and links to associated profiles of their friends and colleagues.
Exactly how the technology will be used isn’t clear – it could be to play a song based on the emotional state, or it could be to show a list of recommended tracks.
This is the latest in a long line of Spotify patents granted in recent months – including one for a karaoke feature that lets users overlay a music track with their own vocals.