Spotify wants to suggest music based on your mood
Spotify is working on a feature that will detect a user’s mood and recommend appropriate music accordingly.
The music streaming service filed a patent for the technology with the United States Patent and Trademark Office (USPTO), Phone Arena has reported.
According to Spotify’s description of the feature in the patent, it will use speech recognition algorithms and machine learning to assess the intonation, stress, and rhythm of a user’s voice.
This would allow the app to automatically determine whether a user’s emotional state is happy, angry, sad or neutral.
In addition, Spotify would retrieve information from the user’s environment – including a category for their physical location – such as home, school, work, bus, or car.
The social setting would also be factored into the calculation, with Spotify taking into account whether you are alone, in a small group, or at a party.
To further optimise recommendations, these variables will be used in conjunction with other parameters like the user’s listening history and their friends’ music taste.
The app can already recommend and filter playlists based on a user’s emotional state, but to do this requires the user to manually select their mood.
It is not clear whether the feature will be included in a future update of the app, however, as it is still only a patent.
Mood detection technologies in action
There are existing mood detection technologies which indicate the idea may not be as far-fetched as some would believe.
One such technology is the Tone feature available in Amazon’s Halo band.
It uses the gadget’s built-in microphone to analyse how a user speaks and provides feedback to help strengthen their communication.
The feature requires the user to read several quotes and sentences in order to recognise their standard voice – which it uses as a baseline for comparing conversations.
Tone can then be set to track conversations throughout the day, or you can use the Live mode and track how you sound in real-time.
Loading ...