The new feature of the Google Magenta project, created by an intern, allows users to create custom music rooms by mixing lo-fi, hip-hop tracks in a browser without any musical foundation. The Magenta project aims to create art and music using Google’s machine learning system, and Lo-Fi Player is an interesting example.
When you turn on Lo-Fi Player, you’ll be taken to a pixelated virtual “room” where you can change tracks, including bass lines and melodies, by clicking on different objects in the room– such as clocks, cats, or pianos.
Vibert Thio, founder of Lo-Fi Player, wrote in a blog post: “The view from outside the window is related to the background sound in the track, and you can change your vision and music by clicking on the window.”
Lo-Fi Player also has an interactive YouTube stream that users can think of as a “shared space” that invites friends to create in the same music room. But instead of clicking on the elements in the room, the player enters commands in the live chat window to rearrange the tracks.
Magenta, powered by Google’s open source TensorFlow system, demonstrates the role of machine learning as a tool in the creative process. Magenta’s other projects include Piano Genie, an artificial intelligence program that allows anyone to play the piano (similar to Guitar Hero), and NSynth, a machine learning algorithm that uses neural networks to learn and create new sounds.
Lo Fi Player is customizable; its source code can be found on GitHub, thio says, and the team has created a “Play, Magenta!” Tutorials that allow users to edit sounds and canvases live in their own browsers.