Full Program »

A Fluid Chord Voicing Generator

We present an interface and methodology for composing a jazz piano accompaniment. Rather than restricting to voicings that are drawn from a preset vocabulary or, if necessary, generated by a fixed algorithm, our system supports a broad set of parametric specifications. A random element permits variety in the generated voicing sequences, as well as providing for aural experimentation. Generated voicings can be saved to augment a vocabulary if desired. The probabilistic method selects notes based on a series of weightings and multipliers. The general framework is designed to mimic human pianist hands. This method of voicing generation thus provides a viable alternative to jazz accompaniment for the Impro-Visor program, expanding the range of available sounds.

Author(s):

Daniel Scanteianu    
Computer Science
Stony Brook University
United States

Errick Jackson    
Computer Science
Harvey Mudd College
United States

Robert Keller    
Computer Science
Harvey Mudd College
United States