Essynth

Abstract: A new way to control sound spatial dispersion using the ESSynth Method is introduced here. The Interaural Time Difference (ITD) is used as genotype of an evolutionary control of sound spatialization. Sound intensity and the ITD azimuth angle are used to define spatial dispersion and spatial similarity. Experimental results where crossover and mutation rates were used to create spatial sonic trajectories are discussed.

Abstract: The creation of an artwork named RePartitura is discussed here under principles of Evolutionary Computation (EC) and the triadic model of thought: Abduction, Induction and Deduction, as conceived by Charles S. Peirce. RePartitura uses a custom-designed algorithm to map image features from a collection of drawings and an Evolutionary Sound Synthesis (ESSynth) computational model that dynamically creates sound objects. The output of this process is an immersive computer generated sonic landscape, i.e. a synthesized Soundscape. The computer generative paradigm used here comes from the EC methodology where the drawings are interpreted as a population of individuals as they all have in common the characteristic of being similar but never identical. The set of specific features of each drawing is named as genotype. Interaction between different genotypes and sound features produces a population of evolving sounds. The evolutionary behavior of this sonic process entails the self-organization of a Soundscape, made of a population of complex, never-repeating sound objects, in constant transformation, but always maintaining an overall perceptual self-similarity in order to keep its cognitive identity that can be recognize for any listener. In this article we present this generative and evolutionary system and describe the topics that permeates from its conceptual creation to its computational implementation. We underline the concept of self-organization in the generation of soundscapes and its relationship with computer evolutionary creation, abductive reasoning and musical meaning for the computational modeling of synthesized soundscapes.

Abstract: This article focuses on the interdisciplinary research involving Computer Music and Generative Visual Art. We describe the implementation of two interactive artistic systems based on principles of Gestural Data (WILSON, 2002) retrieval and self-organization (MORONI, 2003), to control an Evolutionary Sound Synthesis method (ESSynth). The first implementation uses, as gestural data, image mapping of handmade drawings. The second one uses gestural data from dynamic body movements of dance. The resulting computer output is generated by an interactive system implemented in Pure Data (PD). This system uses principles of Evolutionary Computation (EC), which yields the generation of a synthetic adaptive population of sound objects. Considering that music could be seen as "organized sound" the contribution of our study is to develop a system that aims to generate "self-organized sound" - a method that uses evolutionary computation to bridge between gesture, sound and music.