Markov-Model-Based Interactive Computer Performance in PureData
The project aims at implementing a musical system (in PureData), which can learn from a previously played performance, using the concept of Markov model.The system divide its learning stages into three main musical components: harmony, melody, and drums, each one of which integrates Markov-models to assimilate from the musical input. To make the system more expressive, the time and pitch dimensions are learned independently.
This project has been developed with my classmate Davi Bedoya, and resulted in a musical performance, in which we both performed using MIDI keyboards in order to demonstrate the system capacity. The patch can be downloaded by clicking the button up-top.