Create sounds with a stroke, play the piano with ease, produce music with a click. Among the operations allowed by the applications that Sony CS Lab presents to Maker Faire 2021, scheduled from 8 to 10 October. All characterized by the use and development of artificial intelligence.
Among the most interesting creations is the application Piano Inpainting, which allows anyone to compose complex pieces of music on the piano in a wide variety of styles. It is a virtual assistant capable of adding missing parts to a performance and making it possible to create new pieces of music quickly. Pia is available as an Ableton Live plug-in and this allows musicians to integrate it into workflows, within a professional digital audio workstation.
To approach the musical world in an original way here Notono, an open-source web interface that helps to generate and transform the sounds of musical instruments through operations similar to painting. It therefore becomes possible to “paint” a little guitar at the beginning of the piece for a rougher attack, or maybe insert the organ in the lower frequencies, up to draw the luminous textures of the trumpet in the mid and high frequencies at the end of the piece. . Just connect a Midi keyboard to your computer and use the resulting sound as a synthesizer to play chords and melodies. You can also keep tweaking the sound or letting it transform itself as you play.
Robots, artificial intelligence, prototypes: discovering new technologies at Maker Faire 2021
Thanks to artificial intelligence, it also becomes a musical instrument Flow machine, which functions as a plug-in to the digital audio workstation and features a machine learning model that analyzes musical data based on the stylistic palette, selected by the user, to match the genre and chord progression of the song they wish to create. Within Flow Machines, users can create their own original style palette or choose from various presets.
Maker Faire Rome, the best twenty projects of the event
In the end DrumNet, the interactive tool of artificial intelligence that is used to generate drum rhythms on existing music. The user can select a style by placing a point on a 2D plane. Available styles are learned from real music via machine learning. The rhythms automatically adapt to the music’s tempo and structure and to achieve consistent rhythms: the different types of percussion can be based not only on the incoming music, but also on the rhythms already generated within DrumNet. With DrumNet a user can also extract the rhythm style from an existing audio file and use it as a preset.