Newsgroups: comp.speech
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!news.mathworks.com!news.alpha.net!uwm.edu!math.ohio-state.edu!howland.reston.ans.net!swrinde!pipex!uknet!bcc.ac.uk!pc13.phon.ucl.ac.uk!andrew
From: andrew@PROBLEM_WITH_INEWS_DOMAIN_FILE ()
Subject: Diphone Synthesis
Sender: news@ucl.ac.uk (Usenet News System)
Message-ID: <1995Feb17.101947.45328@ucl.ac.uk>
Date: Fri, 17 Feb 1995 10:19:47 GMT
Organization: I need to put my ORGANIZATION here.
X-Newsreader: TIN [version 1.2 PL2]
Lines: 28

We are working on techniques of enhancement of natural and synthetic
speech that overcome some of the limitations imposed on speech by
human perception and production systems.

To test the benefits of such manipulations we would like to use a
synthesiser that allows us access to the diphones/allophones used to
construct the final output.  After enhancing information-rich regions
of the diphones, we hope to show an improvement in the intelligibility
of the resultant synthetic speech, shown by a reduction in cognitive
load or by a greater resistance to noise.

Does anyone have a synthesiser we could use, or know of one that
produces reasonable output?

We would not need to know the secrets of the conjoining mechanism,
but just have access to the diphones so that we could enhance them.

Regards,

--

Andrew Simpson                      |
Phonetics and Linguistics           |  Tel. +44 71 380 7777 x5003
University College London           |  Fax  +44 71 383 0752
Wolfson House, 4 Stephenson Way     |  E-Mail: andrew@phon.ucl.ac.uk
London NW1 2HE                      |

 
