Voice puppetry: exploring dramatic performance to develop speech synthesis

Matthew P. Aylett, David A. Braude, Christopher J. Pidcock, Blaise Potard

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Technology and innovation is often inspired by nature. However, when technology enters the social domain, such as creating human-like voices or having human-like conversations, mimicry can become an objective rather than an inspiration. In this paper we argue that performance and acting can offer a radically different design agenda to the mimicry objective. We compare a human mimic’s vocal performance (Alec Baldwin) of a target voice (Donald Trump) with the synthesis and copy resynthesis of a cloned synthetic voice. We show the conversational speaking style of natural performance is still a challenge to recreate with modern synthesis methods, and that resynthesis is hampered by current limitations in speech alignment approaches. We conclude by discussing how voice puppetry where a human voice is used to drive a synthesis engine - could be used to advance the state-of-the-art and the challenges involved in developing a voice puppetry system.
Original languageEnglish
Title of host publicationProceedings of 10th ISCA Speech Synthesis Workshop
PublisherISCA
Pages117-120
Number of pages4
DOIs
Publication statusPublished - 2019

Fingerprint

Dive into the research topics of 'Voice puppetry: exploring dramatic performance to develop speech synthesis'. Together they form a unique fingerprint.

Cite this