UPDF AI

VQ-Synth: Development and Perceptual Evaluation of a System for Voice Quality Modification

Isabel S. Schiller,Alexander Schnapka,2 Authors,Simon Stone

2022 · DOI: 10.61782/fa.2023.0250
1 Citations

TLDR

It is concluded that VQ-Synth meets its intended purpose to increase perceived breathiness in vowels – not only based on unknown voices of varying quality but also on listeners ’ own voices.

Abstract

  • This paper introduces VQ-Synth , a prototype system for the voice-quality modification, designed to increase breathiness in sustained vowels. The system is currently operating offline but shall be used for real-time auditory feedback alteration in the near future. Here, we describe VQ-Synth 's architecture and operating principles and evaluate its efficacy through two listening experiments. In experiment 1, we examined the impact of different VQ-Synth settings on listeners' breathiness and naturalness perception of an unknown speaker ’ s normal and hoarse voice recordings. In experiment 2, an extension of experiment 1, we tested the same but based on listeners ’ own-voice recordings. Both experiments confirmed our hypothesis that, with stronger voice manipulation, vowels were perceived as increasingly breathy. At the same time, perceived naturalness declined. We conclude that VQ-Synth meets its intended purpose to increase perceived breathiness in vowels – not only based on unknown voices of varying quality but also on listeners ’ own voices. Further investigation into optimal parameter settings is needed, especially in regards to naturalness. With refinements and an implemented real-time mode, VQ-Synth may soon be used to study vocal motor control in healthy speakers and those with voice disorders.