IEEE

Talk on Speech synthesis work at the Centre for Speech Technology Research

IEEE Signal Processing Society, Bangalore Chapter, IEEE Bangalore Section,
and
Department of Electrical Engineering, Indian Institute of Science

Invite you to a seminar by
Professor, Simon King
University of Edinburgh
on
Speech synthesis work at the Centre for Speech Technology Research

Time & Date:  3:00pm; Monday February 18, 2013
Venue: Multimedia Classroom (PE 217), Department of Electrical
Engineering, IISc.

Abstract
I will describe two projects that make use of "knowledge" of speech
production and perception to improve speech synthesis. The first project
uses measured articulatory data as a parallel data stream, alongside the
acoustic information, in a statistical parametric speech synthesizer. This
facilitates direct manipulation of the speech in terms of articulation –
such as creating vowel sounds that do not occur in the training data,
simply by moving the tongue position - whilst maintaining high-quality
speech based on textual input. In the next project, I will describe how a
computational model of speech perception can be employed to guide
modifications in synthetic speech, leading to a text-to-speech synthesizer
that automatically adjusts its output to maximize intelligibility in the
presence of additive noise.

Finally, I will describe a rather different approach to speech synthesis,
in which we replace traditional forms of linguistic knowledge used in most
TTS systems (e.g., the phoneme set, the lexicon, parts of speech) with
machine learning. This enables us to build systems for any language at low
cost and with limited expertise.

Biography of the Speaker
Simon King holds M.A. (Engg) and M.Phil. degrees from the University of
Cambridge and a Ph.D. from the University of Edinburgh. He has been with
the Centre for Speech Technology Research at the University of Edinburgh
since 1993, where he is now Professor of Speech Processing and the
director of the centre. His interests include speech synthesis,
recognition and signal processing and he has around 140 publications in
these areas. He has served on the ISCA SynSIG board and currently
co-organizes the Blizzard Challenge. He serves on the IEEE SLTC, has
served as an associate editor of IEEE Transactions on Audio, Speech and
Language Processing and is a current associate editor of Computer Speech
and Language.

Leave a Reply