TU Berlin

Quality and Usability LabReviewed Conference Papers

Page Content

to Navigation

Reviewed Conference Papers

go back to overview

A framework to incorporate aspects of social perception in synthetic voices
Citation key rallabandi2020a
Author Rallabandi, Sai Sirisha
Title of Book Proceedings of 6th Docotal Consortium, Student event of Interspeech 2020 (ISCA-SAC)
Pages 1–3
Year 2020
Location Virtual
Address ISCA
Month nov
Note electronic, online
Publisher ISCA
Series Interspeech
How Published Abstract
Abstract In my dissertation, I investigate the social speaker characteristics: warmth and competence. I refer to them as the global aspects of social perception and analyse them in synthetic speech. Specifically, I hypothesize two things: (a) there are certain vocal cues responsible for evoking these characteristics and (b) it is feasible to incorporate these characteristics in the present day speech generation mechanisms. To validate my hypotheses, I propose a two stage framework: In the first stage, I perform extensive subjective evaluations in the form of listening tests to identify which synthetic voices demonstrate characteristics of warmth and competence. I also perform analysis on the voices to identify the vocal cues corresponding to these characteristics. In the second stage of my dissertation, I propose to incorporate the identified vocal cues into the speech generation mechanism. Employing two target application scenarios - Customer Service and Health care, I propose to show that incorporating social speaker characteristics can not only improve user satisfaction but also user trust.
Link to publication Link to original publication Download Bibtex entry

go back to overview

Navigation

Quick Access

Schnellnavigation zur Seite über Nummerneingabe