EarGram : an Application for Interactive Exploration of Large Databases of Audio Snippets for Creative Purposes

Abstract

This paper outlines the creative and technical considerations behind earGram, an application built as a Pure Data patch for real-time concatenative sound synthesis. The system encompasses four generative strategies that automatically re-arrange and explore a database of descriptor-analyzed sound snippets (corpus) by rules other than its original temporal order into musically coherent outputs. Of notice are the system’s machine-learning capabilities that reveal musical patterns and temporal organizations, as well as several visualization tools that assist the user in making decisions during performance.