A reflection on the statistical method for trimbre synthesis

Abstract

There is relatively little effort placed into exploring potential expressive usage of natural synthesis techniques. This research aimed to investigate possible expressive applications of Josh McDermott’s natural synthesis method, which employs a statistical decomposition methodology and has shown promising findings in the field of sound texture psychoacoustics. The study modified the computational code of the original method to examine if it could be used as an instrument synthesizer and morpher between instrumental samples and sound textures. The findings demonstrate that the approach fails as an instrumental synthesizer due to an apparent inability to statistically deconstruct the harmonic and tonal qualities inherent in instrumental sounds, however, it performed well as a morpher with a high degree of parametrization. In its current state, the synthesis technique can be used as an expressive tool, as it adds distinct qualities of sound textures to instrumental sounds.

Date of Award	18 May 2022
Original language	English
Awarding Institution	Universidade Católica Portuguesa
Supervisor	Pedro Duarte Pestana (Supervisor)