Those here who are musicians? And among them, how many of them program? And that they use artificial intelligence tools to make art? Who is programming artificial intelligence with Neural Synthesis to make music?”
With this little questionnaire, the researcher CJ Carr, one of the gurus of this technology applied to music creation, took the pulse of the Sonar D audience yesterday, whose presence can only be justified at the advanced music festival, already that the 2023 edition celebrating three decades of history is largely dedicated to AI.
Sergio Caballero, for example, co-founder of Sónar and architect of the curious and influential image campaigns, has invited him to swallow each and every one of these works to turn them into a shocking installation, Processing.. 30 years of Sónar, where the images change for more than half an hour in a video creation in which Franc Aleu collaborates.
On the first day of Sónar, the issue of the climate emergency came to the fore at Stage D by Mediapro. The research group Forensic Architecture showed how it uses digital tools like those of a simple mobile phone to defend causes and injustices that could well be artistic material for a museum or evidence in a trial. And a practical example was the case of a shipwreck between Turkey and Cyprus: with GPS, it was shown that it happened in Greek waters. Jana Winderen also performed, who is dedicated to capturing sounds in Antarctica and is able to capture that of certain crustaceans, which are scandalous.
Daito Manabe, the Leonardo Da Vinci of Sónar, the Japanese technologist who goes to every festival, be it with interactive installations or shows with robots, will perform this edition of the festival every day: today, Friday, with AI.
If we go back to the SonarÀgora - where the Domestic Data Streamers from Barcelona will later be there with a conference on generative AI -, we find CJ Carr, who continues with the talk full of gags and ironies about his geekiness: "We made an emo AI with Silverstein 26 hours with a thousand songs..., which, of course, no one would listen to in three days, except me, who listened to them three times", he laughs.
The 35-year-old Bostonian who decided to make his own when he ran into a skeptical AI professor in college – “he told us it wouldn't work, that it wouldn't become a reality” – is the other half of the pair of music hackers and metal fanatics Dadabots. In addition, they have shamelessly experimented with the limits of electronics through algorithms. Therefore, the more rare and indigestible the sound that the machine vomits, the more his eyes shine with enthusiasm at the result. Because, as director of research at Harmonai, CJ Carr has explored how to apply Stable Diffusion technology to music (the one that mixes the art of a Dalí with that of a Van Gogh). With individuals like him, everything that was previously reserved for supercomputers is now for domestic use.
Carr combines technicalities with examples that suggest multiple questions. Like the crazy idea he had to get Frank Sinatra to sing a Britney Spears song. Will there be, like Blade Runner, ways to detect replicating music?
"Surely in five years we will see companies dedicated to distinguishing what is real and what is fake. But before that we'll see a lot of AI-generated music that won't come to light as such. It will be a problem, impossible to discover. And every time a fake detector emerges, there will be a way to avoid it, and the AI will be more and more ready", he says to La Vanguardia. And the culture of effort? "It will be replaced by the culture of the concept. Thinking of something and having it ready in a minute is a very powerful thing."
When asked if the machine is creative, his answer is immediate. "Of course - answers Carr with the patchwork pants and dynamic hair. Finding unusual solutions to problems is also being creative, and AI is very good at that. In addition, she has no experience of living in the world, but if we are able to describe a musical idea or concept to her and she is able to interpret it, it is a creative process."
The Project Area, on the extensive surface that connects Sónar D and Sónar Village, is a project exhibition space designed to be seen and touched. And, as explained by the director of Sónar D, Andrea Faroppa, this time it was by public call. During three weeks, they received 450 projects and "of an unexpected quality", he says. If the area was designed for 20, they have managed to contain 85 and have welcomed the entire scientific sector, including the National Museum of Science and Technology and a selection of universities: the Norwegian Polytechnic, the of Art in Vienna... and those in Barcelona. The Minister of Culture, Natàlia Garriga, listens to the explanations..., and also the mayoress of Barcelona, Ada Colau, wanders around the site. A forum co-designed with Hac Te has brought together representatives of the main European research centers, as well as museums and residences.
All kinds of inventions coincide, and it contrasts from the "order of plants" to the "chaos of technology", which aims to educate the eye to distinguish (or not) between reality, virtual reality and augmented reality From the stand of modular brands and synths for fans of music creation, there is the Barcelona Supercomputing Center stand, which has been at Sónar for ten years. And, as it could not be otherwise, they have tackled the AI and tried to see what is inside the GPT Chat, the assistant that the world faces without knowing what intelligence and skill it has. Supercomputing has visualized it and exposed it as the statistical model it is. And, taking advantage of the fact that it is the 30th anniversary of Sónar, they have loaded it with the lyrics of the songs of all the artists who have performed at the festival: 40,000 tracks, 10,000 artists..., which appear grouped according to language affinity, lyrics and styles of music, which shows how a kind of culture has been created.
Any conclusions from the analysis of 30 years at Sónar? What is the emotional tendency of the music? "We see that it has changed from electronics to urban and that nowadays urban has more weight - explains Fernando Cucchietti, from Supercomputing-. And, in the emotional sphere, we see a small increase in sadness..., despite the fact that joy has a brutal representation in terms of the neutral".
