Aaron Montoya-Moraga, Corbin Ordel

Machine Learning will be demonstrated as we teach a piano to score films and tv shows according to the colors displayed on screen


Piano Vision: Dark Side of The Moon is a machine learning and music composition project. Simply put, this is a piano that watches a television screen playing movies and composes and original score. Using MaxMSP the patch will reads all of the RGB values from the screen and playsmusic according to the intensity of each color. For example, if there is fire (more red) it plays fast, water (more blue) plays slower, trees (more green) it plays somewhere in the middle. Using a machine learning program call Wekinator, this project attempts to teach a computer different musical notes that correspond with color values in order to trigger a varied response. The responses are then recorded or *learned* and then used to score different films. The instrument is a upright piano with the keyboard and hammers removed, exposing the internal *guts* of the piano. The playing of the piano will be the achieved with solenoids to hit the strings and servos to pluck the strings. This project will be the orchestra to all of your movies, haunting and serene, glitchy and humorous. This iteration will have our Piano watch the trailer for the movie Die Hard. Searching for explosions and fire, the piano will play accordingly to accurately play compositions familiar to the action movie genre. The intention of this piece is to associate the absurdity of the destruction constantly displayed in our consumable entertainment and the incredible technology used to make these sequences come to life. When we pair a half destroyed piano with a film conveying over the top destruction, can we sense the connections between the two? Have we become so desensitized to what we watch that cities falling to the ground no longer bothers us? Does this not bother you? Does a broken piano trying to play music bother you? Both a piano using machine learning and a big budget Hollywood movie utilize incredible technologies, but in the end, what do they make us feel, if anything? We want to demonstrate the connection – the connection that our ability to use these incredible technologies is a beautiful dance between art, creativity, and imagination.


Machine Learning for Artists, Readymades