MUSICA: MUSical Improvising Collaborative Agent
Abstract
Major Goals: The original goal of the MUSICA project is to develop a real-time Jazz improvisation system capable of interacting meaningfully with human jazz musicians -- this is referred to as the Trading Fours (T4) use case scenario within the Communicating with Computers Program. In order for this to be possible, MUSICA must be able to do the following: (1) Take as input a digital music representation of music (whether originating from a quantization facility for real-time audio or from a representation of a score) (2) Analyze the music excerpt to identify patterns recognizable by a human musician as significant (3) Formulate a response the both (a) would be recognizable by a human musician as referencing musical patterns/features in the input score and/or from previous interactions while also (b) contributing novel musicalstructure. (4) Output that response in both a score and audio form that a human musician can recognize/read. Step (3) is the key step to achieving musical communication between a human and machine: MUSICA will need to demonstrate that it "hears" what the human musician is expressing while also contributing novel musical structure to the interaction. In Mid 2016, a second use case for MUSICA was introduced: musical composition by conversation (CbC). In composition by conversation, MUSICA interacts with human musicians through a natural language interface while the human and Musica collaborate on the shared task of creating a musical score. The goal is to develop the natural language interface capable of parsing text into a representation that MUSICA can then interpret as expressions about a score, with both the human and MUSICA contributing to edits made to the score.
Document Details
- Document Type
- Technical Report
- Publication Date
- Apr 30, 2017
- Accession Number
- AD1113960
Entities
People
- Clayton T. Morrison
Organizations
- University of Arizona