MusicBox

• Live: view application
• People Involved: Anita Lillie
• Created On: December 16, 2008

About

Anita's Master’s thesis is focused on the problem of finding music you like, both inside your own collection (to match a particular mood, for example), or outside of your own collection. She didn’t like the hugely text-dependent ways of searching for music in traditional media players like iTunes and Windows Media Player. Her thesis work will depart from that interface, and will present your music collection in a space, organized by similarity. She will be calculating similarity both from audio features (straight from the audio signal) and textual descriptions (gathered from tags on the web). Ultimately, her goal was to create a picture of what she listens to, what her friends listen to, and to see how those pictures sit in the broader context of music as a whole.

From the Author

I made MusicBox because I was frustrated with mainstream, text-based interfaces for music libraries. Since MusicBox makes music maps based on songs' acoustic characteristics, it relies on tools to analyze each song. The Echo Nest API was integral to doing this audio analysis.