Abstract
In this paper, we present a systematic attempt at identifying parameters of synthesized voices that can be adjusted in order to index the robot to which the voice belongs. We carried out an experimental study for which we first generated stimuli by synthesizing the robot voices using a free text-to-speech synthesis system, which were then manipulated according to the parameters f0 level, f0 range and formant dispersion, associated with melodic and physical dominance. These stimuli were matched with videos of robots of different sizes and appearance. We embedded the videos in an online questionnaire, where participants’ tasks were to indicate the impressions evoked by the robots and to identify the robot to which the voice was the best fit.
Originalsprog | Engelsk |
---|---|
Titel | Proceedings 1st International Seminar on the Foundations of Speech : Breathing, Pausing, and the Voice |
Forlag | Syddansk Universitet |
Publikationsdato | 2. dec. 2019 |
Sider | 88-90 |
Status | Udgivet - 2. dec. 2019 |
Begivenhed | 1st International Seminar on the Foundations of Speech: Pausing, Breathing and Voice - University of Southern Denmark - Campus Sønderborg, Sønderbrog, Danmark Varighed: 1. dec. 2019 → 3. dec. 2019 |
Seminar
Seminar | 1st International Seminar on the Foundations of Speech |
---|---|
Lokation | University of Southern Denmark - Campus Sønderborg |
Land/Område | Danmark |
By | Sønderbrog |
Periode | 01/12/2019 → 03/12/2019 |