Songle – The music player that analyzes what you’re hearing

By Kira Grunenberg

I’ve posted here before about cases where audio and visuals crossed paths. Jordan Rudess’s “Tachyon App” is a prominent example of such, with its ‘starlights;’ reactive to touch and pitch for making music.

Tachyon had a slightly complex, extra layer absent from many apps and music players –the incorporation of tonal keys, modulation, musical modes and complete control of them all. These elements, while allowing for freedom within Ruddes’s app, don’t exactly resonate with the whole of the music appreciating public. Not everyone necessarily wants to be bothered learning the minute aspects of music in general, nor the popular songs they enjoy listening to.

That said, it doesn’t mean ‘looking under the hood’ of what one is listening to isn’t without purpose or fascinating to some –and one music based web service is out to prove that very point.

Songle, (pronounced Song-LEE) is a new and radical type of online music player/streamer out of Japan that hits many strong points in uniqueness, separating it from being “just another service that plays music you like.” Just like the Tachyon app, visualizations accompanying the music are part of the initial appeal. Then on top of that, (as reported in Wired), for a service that only launched its first beta mode in February of this year, Songle packs a large capacity track library with “about 80,000 [searchable] songs so far.”

However, don’t think that Songle is out to just make the prettiest flying lights alongside that pop hit. The visuals users can view while listening are paired with information that can illuminate beat structure, melody and chord progressions.

As described on the “What is Songle?” portion of the website, the player uses what are described only as, “music understanding technologies” (when translated form Japanese to English) to analyze a song and deliver the primary aspects shown in playback. Of course, it’s not as if the creators of Songle at the Advanced Industrial Science and Technology Institute are providing the master key to ripping off anyone’s composition. In fact, although it has reached public mode, Songle is still in development and furthermore, functions in a very Wiki-type manner; encouraging users to input new information on tracks they hear if the given sets of data are wrong.

“…you can then correct the existing annotations. …Songle accumulates everyone’s contributions, including yours, to gradually refine the annotations. So even correcting only one or a few errors makes all the difference. Songle was designed to involve everyone to annotate and enjoy music together.

Other fun and intriguing aspects of Songle include the ability search for songs by chord progressions. Any song containing said sequence will come up in the results. (Punk music would be in trouble with this!) For those times demanding some spontaneity, the homepage also has a similarity graph that provides a branching collection of songs that have alike vocal timbres and any song can be put in the center to see what branches off of it.

When it comes to the legitimacy of this whole endeavor, Songle runs a methodology similar to Germany’s Musicplayr. No songs in the Songle library are cached and hosted directly, thereby not getting caught up in the fragile hazards of licensing. If a track is uploaded for analysis and then is later removed from the place of origin, the song respectively disappears from Songle’s database.

Currently, many of the tracks available are by Japanese artists, but for an example relatable to American audiences, No Doubt’s new track, “Push and Shove” makes an interesting rhythmic and melodic display, shown below, through a couple screenshots of two out of four available visual modes.

Kira is an old school music nerd with a love for all things creative; always searching for music’s common ground. She graduated with an M.A. in Performing Arts Administration from New York University. Drop her a tweet @shadowmelody1

Comments

comments