Header

Synesthesia

Music visualization for the modern audiophile

What's New?

Visualize your music.

The purpose of this application is to visualize music in a meaningful and aesthetically pleasing way. The visualization encodes information about pitch, loudness, rhythm, and timbre.

Improving Music Visualization

Listening to digital music has never been easier. With but a few clicks, ordinary internet users can have the world's music library at their fingertips. Popular digital music players, such as Windows Media Player and Apple iTunes, provide quick access to local and internet music libraries, while providing many additional features. One of the common features of these music players is the ability to display music visualizations during song playback. While these visualizations can be aesthetically pleasing (indeed, iTunes's new Magnetosphere visualization was originally developed by a third-party design company using Processing and OpenGL), it is difficult (perhaps impossible) to extract meaningful information about the music from viewing the visualization.

Wma
Windows Media Player's "alchemy" visualization on John Coltrane's "Moment's Notice"

Itunes
Apple iTunes's "magnetosphere" visualization on Beethoven's "Moonlight Sonata"

While aesthetics are an essential component of any great music visualization, the perfect music visualization can also convey information about the music. There are plenty of types of data to be extracted from music, such as pitch, tempo, rhythm, timbre, and loudness. Hence the goal of Synesthesia is to improve upon these existing visualizations by creating a platform for visually encoding music data in an aesthetically pleasing manner.

Motivating Questions

With this overall goal in mind, we used these three questions to motivate the design and implementation of Synesthesia:

  1. How can we visualize music in a way that is both informative and aesthetically pleasing?
  2. How can we visually distinguish elements of songs, such as loudness, pitch, timbre, and rhythm?
  3. How can we visually distinguish different types of songs?
Question 1 is a restatement of our original task, as motivated by the lack of informativeness in existing music visualizations. Question 2 guides the specifics of our design and motivates our choice of visual encoding for our data set. Question 3 is hopeful; with an ideal visualization, we might be able to distinguish songs from each other simply by looking at their visualizations.