Play a tune, dictate the mood. That’s what a good DJ does. But three students at the Copenhagen Institute of Interaction Design wanted to flip the script and use music to keep the existing vibe going without having to scour your library for the perfect mix.
Enter Classyfier, a table that listens to its surroundings and chooses music accordingly. Designed by Benedict Hubener, Stephanie Lee, and Kelvyn Marte, Classyfier is embedded with a microphone that picks up ambient noise and the routes inputs it to a computer to process the sound. The students trained a machine learning algorithm to listen for sounds associated with three beverages that often go hand-in-hand with a specific mood, then plays it through Bluetooth speakers. The sound of a spoon stirring a cup of tea triggers mellow coffeehouse jazz. Clinking stemware activates classy Flamenco music, and the hiss of a beer can opening prompts heavy metal. Knocking the table skips songs.
This is a limited application of the technology. But imagine how it could change streaming services if it were to be deployed more broadly. Platforms like Spotify already use algorithms to find music you’ll like based on your past listening habits. But what those algorithms don’t have is context as to why you listen to what you do. Are there playlists and artists you gravitate toward while you’re at work, or hosting dinner parties, or blasting music in your car? If sensors could understand your environment better, algorithms could then find just the right set of tunes.
We’ve all had that moment where a playlist on shuffle serves up a mood killer. Maybe in the future, an AI trained on the sounds of your activities will assure that none of your classy book club guests discover your secret love of Van Halen.DB