Meet Eduardo Miranda’s singing robots. Miranda is a researcher at the Interdisciplinary Centre for Computer Music Research at the University of Plymouth. He built these two warbling toasters, gave them both a vocal range and the ability to adapt, then put them in a room together for two weeks. They went in singing radically different “tunes”—random sequences of six notes at a time—and came out crooning an oddly haunting duet consisting of about 20 sounds.
The robots have simplified vocal cords, and their voices are strangely human. They also have microphones for ears and cameras for eyes. Their song was a spontaneous, emergent order. One robot would sing its randomly generated sequence, then the other would follow suit with its own tune. If the first robot judged the sounds to be similar, it would nod and store the notes for future use. If the sounds were too different, it would shake its head and the tunes would be forgotten.
A shared musical vocabulary evolved. “The robots develop their own musical culture,” Miranda told New Scientist. “There are no preprogrammed musical rules.”
Katherine Mangu-Ward (kmw@reason.com) is an associate editor of reason. To hear the robots
The robots have simplified vocal cords, and their voices are strangely human. They also have microphones for ears and cameras for eyes. Their song was a spontaneous, emergent order. One robot would sing its randomly generated sequence, then the other would follow suit with its own tune. If the first robot judged the sounds to be similar, it would nod and store the notes for future use. If the sounds were too different, it would shake its head and the tunes would be forgotten.
A shared musical vocabulary evolved. “The robots develop their own musical culture,” Miranda told New Scientist. “There are no preprogrammed musical rules.”
Katherine Mangu-Ward (kmw@reason.com) is an associate editor of reason. To hear the robots
No comments:
Post a Comment