MusicBox is written in Java 6, and uses Processingfor drawing the visuals. Its dependencies are
written in Python (for creating the cached soundsieve thumbnails) and Perl (for uploading and
retrieving analysis result files from The Echo Nest). Audio analysis depends on The Echo Nest
Analyze API [56] and Rhythm Histogram Matlab code from the Vienna University of
Technology . Jonathan Hilliker created the icons used for the playlist playback buttons.
The Echo Nest (the.echonest.com) seems to be the classification engine that makes it possible.
beats per minute, length, tone, and a whole bunch of other semi-tangible quantities
Not once have I ever gone through my collection looking for a song fitting any of the above particular qualities. Maybe some have, but when Pink floyd, the Clash, and Beethoven can occasionally aggree on BPM, it becomes a distinctly useless metric.
The point is that those are metrics which the computer uses to classify the music. You can then find similar music which may not be obvious at first. (Just look at the video to see examples.)
MusicBox is written in Java 6, and uses Processingfor drawing the visuals. Its dependencies are written in Python (for creating the cached soundsieve thumbnails) and Perl (for uploading and retrieving analysis result files from The Echo Nest). Audio analysis depends on The Echo Nest Analyze API [56] and Rhythm Histogram Matlab code from the Vienna University of Technology . Jonathan Hilliker created the icons used for the playlist playback buttons.
The Echo Nest (the.echonest.com) seems to be the classification engine that makes it possible.