[ad_1]
They had no problem finding musicians to compose tunes. Peter has a deep connection with the New York music scene, and people like Jarret need something to pass their time.
With Jarrett writing music with a small group of Broadway composers, arrangers and coordinators on vacation, Dynascore’s technical team stumbled upon the best situation for internal talent. Who can write works that evoke visual drama better than those who already pay every night?
Human problem
The golden ears of human composers are indeed the key to Dynascore’s ultimate success. According to Saatchi, the historical problem with music based on artificial intelligence or “algorithm creation” is that it mainly attempts to teach software instruments to create music from scratch, rather than reinterpreting works that have already been created.
“Making music that really resonates with people is a person’s problem, so you have to start with people,” Saatchi said. “Artificial intelligence is a supercharger for humans.”
In the early days of Dynascore, Saatchi and his team worked to develop a way to decompose original music and classics other than copyright (think Grieg’s “Temple of the Mountain King“) into segments they call “morphones.” They will teach AI a song and then ask it to use the original song’s morphine as a guide to recreate something similar. After that, they will ask musicians to comment on AI’s work.
Getting music to fit into a video perfectly is not as simple as cutting existing songs in a predictable way. Organic transition requires a more thorough understanding of tone, rhythm and intensity, and other musical markers. Therefore, morphones are more than just the speed and tune of the song. They also indicate various other tones and musical characteristics, all of which let the AI know which types of Lego bricks are put together in which way.
After they develop the morphine system, the team will input AI songs and have it re-arrange. It took a while to have enough musical literacy to make the right choice.
“Artificial intelligence will create a song, and the musician will say,’This is too bad,'” Saatchi said, laughing at the simplicity of the test. “It gets feedback, learns from it, and then suddenly creates a coherent composition.”
Artificial intelligence quickly becomes smart enough to adapt to transitions, fades, interruptions, and other user-specified time changes in each song it composes. I have witnessed the birth of the Dynascore version of the reinvented Moonlight Sonata.
Reduce the burden on
Dynascore’s achievements represent a huge improvement over the cumbersome workflow of the past.
“When you work as an editor or film producer, you spend a lot of time in music because you have to adapt it frame by frame,” said DiGiovanna, who has been involved in everything from feature films to TV commercials. “Dynascore you can do it instantly.”
Check out a demo of how Dynascore handles scene transitions.
Tools that enable dynamic music creation are particularly useful when dealing with commercial projects that may need to cut certain items. DiGiovanna gave an example of a director who needed to take out a wallet from an advertisement he produced.
“You have to delete the five-second video. Now the end of this song doesn’t work, and the transition to the next song doesn’t work,” he said. “Dynascore will save me a lot of time then.”
[ad_2]
Source link
