Robot Musician is Programmed to “Jam”

A researcher at Queensland University in Australia has succeeded in creating a musical robot capable of improvising, or “jamming” along with any music it is exposed to. The robot, nicknamed Jambot, can improvise melodies and rhythms to accompany other musicians or backing soundtracks.

But to call Jambot a robot is really not very accurate. It is actually a software program which analyzes musical compositions and can provide an accompanying solo part, backing track or even drum and percussion solos to complement the arrangement.

The software is still autotype, but Toby Gifford, the Ph.D. candidate who invented Jambot, plans to develop a commercial version of the program next year. Gifford envisions a more user-friendly but powerful version of the software, which could be used to create a fuller sound for small bands, or improvise interesting melodies and counter melodies in the recording studio.

Jambot is said to be a significant step forward in the area of software audio perception. The program uses a unique algorithm to follow along with any music that it “listens to.” Jambot can also recognize pitch and tempo changes in real time, so it is not likely to get lost in the middle of a song, no matter how complicated the composition

But one of the most ingenious aspects of the software is a proprietary algorithm that analyzes incoming music and identifies the style, pitch, rhythm and tempo almost instantaneously. This audio detection algorithm potentially has lots of other uses, including voice to text, language translation and accessibility software for the disabled.

But the big question on many people’s mind is how does it sound? So far, there have been very few public performances by the Jambot, but musicians and music technology experts have praised the software for its accuracy in identifying what it hears, and “intuitively” adding to (or jamming along with) the composition.

While Gifford admits that the algorithm is not perfect (yet), he claims that the Jambot produces musically appropriate contributions, and sees great potential for the software as a learning, production or performance tool.

Among Jambot’s strong points is it’s ability to detect higher pitched and percussive instruments with a great degree of accuracy. This could potentially be very useful for DJs working with highly percussive dance music, or also in recording studios to accurately track and keep time while recording.

It will be interesting to see what types of musical uses are created for the Jambot. Who knows, perhaps soon we will see an entirely new genre of music developed using the jamming software?


Related Posts