Entropy sonification

Abstract

We present entropy sonification, a technique to bring interesting data to the foreground of a sonification and push uninteresting data into the background. To achieve this, the data is modeled as an in- formation source, and the underlying sonification is converted into sound grains. Information-theoretic attributes of the data are used to determine the amplitude envelope and duration of the grains. The information source model adds an additional avenue for con- trol. By altering the information source model, one can focus on different aspects of the data via entropy zooming.