A Neural Network Based Framework for Archetypical Sound Synthesis
This work addresses sound synthesis for creative applications, but it appears incremental as it builds on existing neural network methods for perceptual modeling.
The paper tackles the problem of synthesizing sounds with specific human-perceived chaos/order levels by using a neural network to predict and generate timbres, reporting on the accuracy achieved.
This paper describes a preliminary approach to algorithmically reproduce the archetypical structure adopted by humans to classify sounds. In particular, we propose an approach to predict the human perceived chaos/order level in a sound and synthesize new timbres that present the desired amount of this feature. We adopted a Neural Network based method, in order to exploit its inner predisposition to model perceptive and abstract features. We finally discuss the obtained accuracy and possible implications in creative contexts.