They didn’t have much trouble finding musicians to compose tunes; Peter had deep connections to the New York music scene, and people like Jarret needed something to occupy their time.
With Jarrett and a small group of composed Broadway composers, arrangers and orchestras to write the music, Dynascore’s technical team stumbled upon the best case when it comes to in-house talent. Who better to write works that evoke visual drama than those who are already paid to do so every night?
A human problem
The golden ears of human composers are truly the key to Dynascore’s ultimate success. The historical problem with AI-based music or “algorithmically composed” music, according to Saatchi, was that he mostly tried to teach software instruments to write music from scratch, instead of reinterpreting works that had already been written.
“Creating music that actually resonates with people is a human problem, so you have to start with people,” says Saatchi. “AI is a charger for people.”
In the early days of Dynascore, Saatchi and his team worked to develop ways to break up original music and classics that are not copyrighted (you think Grieg’s “Mountain King’s Hall”) In segments called“ morphons ”. They would learn AI songs and then ask him to recompose something similar using the morphons of the original song as a guide. Afterwards, they would ask the musicians to criticize the composition of AI.
Achieving music to fit perfectly into a video is not as easy as remembering existing songs in predictable ways. Organic transitions require a more thorough understanding of key, rhythm, and intensity, among other musical markers. As such, morphons don’t just show the speed and key of a song. They also point to various other tonal and musical characteristics, all of which allow the AI to know which types of Lego music blocks are interconnected in which way.
After developing the morph system, the team would feed the AI songs and ask them to recompose them. It was some time before he became musically literate enough to make good choices.
“AI would compose a piece, and musicians would say, ‘That was bad,'” Saatchi says, grinning at the simplicity of the test. “He gets that feedback and he learns from it and gets to the point where he suddenly creates coherent compositions.”
AI soon became smart enough to fit into the transitions, fading, interruptions, and other time shifts dictated by users in every song he wrote. A version of Dynascore was born to which I witnessed it reshape the Moonlight Sonata.
Take off the load
Dynascore’s achievement represents a dramatic improvement on what has been a cumbersome workflow in the past.
“When you work as an editor or director, you spend so much time with music because you have to adjust it frame by frame,” says DiGiovanna, who has worked on everything from feature films to TV commercials, “With Dynascore you can do it on the go. “
A tool that enables dynamic music composition is especially useful when working on commercial projects where you may need to cut certain items. DiGiovanna gives an example of a director who needs to remove a purse from an ad he has created.
“You have to remove five seconds of video, and now the end of the song doesn’t work, the transition to the next song doesn’t work,” he says, “Then Dynascore will save me a lot of time.”