Can robots sing opera? Not yet, but ‘Pavarobotti’ is trying

Robots will eventually take over more or less of human jobs, but some wonder just how many things will machines be able to do in the future. So when the University of Sussex saw the performance of an opera sung by robots more than one eyebrow was raised.

Two Nao robots recently took centre stage at a symposium bringing together research into music and artificial intelligence. It was only after weeks of ‘intensive singing lessons’ by programming experts.

The project was led by Dr Evelyn Ficarra, Lecturer in Music at the University of Sussex and Assistant Director of the Centre for Research in Opera and Music Theatre (CROMT).

“If, in the near future, we are expecting to see robots used as care workers or teaching assistants, then we need to teach them to understand and respond appropriately to humans. The virtues of the musician – listening, co-operation, group creativity – are transferable skills that could apply in all kinds of human situations. Opera requires all of these, plus vocal expression, acting skills, movement and the ability to respond to other performers. So, in addition to being a fascinating exploration of ‘post human’ performance, the work could have interesting implications for research in artificial intelligence and social robots,” Dr Ficarra said.

She was the one who composed one opera for two robots and cello, ‘O, One’, which is partly sung in binary code. Dr Alice Eldridge, cellist and Research Fellow in Digital Performance, accompanied the robots. A second opera, ‘Opposite of Familiarity,’ was composed by Professor Ed Hughes, Head of Music at the University of Sussex with librettist Eleanor Knight.

They collaborated closely with Dr Ron Chrisley, Reader in Philosophy and Director of the Centre for Cognitive Science (COGS), who has been programming the robots, and with Tim Hopkins, an opera director and researcher at CROMT, who staged the operas.

The challenge was in not using recorded human voices and having those sound files played by the robots but by having the robots ‘sing’ with their own synthesised voices.

“The Nao was not designed to sing: the hardware and software supplied with the Nao robots only allow for a limited variety of interventions into the robot’s speech sounds. It’s been an interesting task subverting and re-purposing the Nao’s speech engine, grafting the composers’ intentions onto pre-programmed vocal algorithms, and finding ways to express pitches and rhythms,” Dr Chrisley explained.

Robot Opera, a mini symposium, took place on the University’s campus, and also included wider talks and discussions about the philosophy and potential impact of artificial intelligence and the arts.