Participants 2023
Luminos
AI Song Contest 2023 / participantsTEAM / Luminos
SONG / Night Serenade
TEAM MEMBERS / Sofía Rueda, Laura Tamath, Martín España
About the TEAM
We are three students of Computer Science at Universidad del Valle Guatemala. We are studying a game development specialization and taking a computational music course.
"My name is Laura Tamath, I was born in Guatemala. I like to dance, discover new places and go out to have fun with my friends and family."
"My name is Sofia and I was born in Colombia. I like to watch movies, draw and listen to indie music."
"My name is Martin and I was born in Guatemala, I love playing videogames and playing guitar. I enjoy listenting all kinds of music, from classical to hard rock."
About the SONG
We used some inspiration on Muse and other rock bands to create this retro indie-rock single. This song consists of 7 parts: Intro (0:00), Verse 1 (0:11), Bridge (0:38), Chorus (1:03), Verse 2 (1:33), Bridge (2:00), and Chorus (2:25). The intro only uses piano and the drums to showcase the main melody that will repeat throughout the song. The Verse 1 uses piano, drums, bass and overdriven electric guitar. The Bridge consists of a set of chords played with both the piano and the guitar. The chorus consists of a melody that is first played by the guitar and then by the piano. The Verse 2 is the same as Verse 1 with some offstet for the guitar. There is a little break between Chorus and Verse 2 that uses an acoustic guitar. And the last Chorus ends with a violin playing the main melody.
About the HUMAN-AI PROCESS
We used ChatGPT (https://chat.openai.com/) to create some melodies specifying in the prompts to use Muse and other rock bands as inspiration. Then we used Python programming language (https://www.python.org/) with Jython Music Library (https://jythonmusic.me/) to create a .mid file that would have all the basic structure of the song. Then we imported this file to Band Lab (https://www.bandlab.com/) and did some modifications regarding the instruments, and some measuring things. Finally we exported the .mp3 file as the final result.