Participants 2022
Rubato Lab.
AI Song Contest 2022 / participantsTEAM / Rubato Lab.
SONG / Beam of Seoul
TEAM MEMBERS / ChangJun, Yoo, Yookyung, Shin, JoongSun, Yean, KhwanHo, Kim, SeongHyu, Jeon, Dasol, Lee, Hyeshin, Joo, Dahye, Park
About the TEAM
We are a research group with un-unified backgrounds. Student, DJ, Office worker, Dancer, Media artist with different majors(CS, COMP, EE). The thing combining us is Passion about music. Engineers are weak at music code, but composers are weak at computer code. We tried our best to solve our passion of generating music in concert.
We are creative Media Artists. Rubato Lab was created in 2018, we would like to introduce the artificial intelligence piano model that was exhibited in 2019 which successfully held. Last year, we participated in the AI SONG CONTEST 2021. All the things are made with AI, but after showing our music, we also strongly agree with AI music and human creativity should walk together.
From that point of view, In 2022, we successfully recruited talented music composers to our team. More creativity with the best technology we already had, we produced conceptual music. Please enjoy.
About the SONG
Tech with music part: Music VAE, Multitrack VAE, DDSP and other latest AI technologies are blended in.
Music with tech part: We tried to deliver Cyberpunk of Korea. Tuned AI models to generate future house concepts. We captured a dark, cybernetic side of korea and bring it into film which can be blended with future house music.
About the HUMAN-AI PROCESS
Drum line:
Used the latest AI generative models to make drum MIDI. We choose drum set from VST we had. Considering our film, we developed our sounds with composers to bring its sound to the future hiphop sound.
Melody line:
Used the latest AI generation models to generate melodies. Multitrack model generates MIDI for the tracks respectively which is proper for its own instruments. We also chose VST sets and transferred its timbre(instruments’ own feature of sound) via an AI model(name DDSP).
Mixing:
Now, we have drum sounds(MIDI from AI * VST) and Melodies for each tracks(MIDI from AI * VST * timbre transfer from AI). All the sources are ready. AS you might expect all the sounds need to be adjusted. Our team also has great listeners. Composer, DJ, dancer with plenty of sound knowledge made it possible to adjust its sound. Might heard of reverbs, EQ, compressors and other FXs. We adjusted the tune carefully.
Mastering:
We used izotope Ozone 9. Fortunately, our AI group leader(born as a DJ) has an amazing room acoustic system in his room! We introduce our fine-tuned music with AI powered.
