2001 Conference Proceedings
Go to previous article
Go to next article
Return to 2001 Table of Contents
END TO END SOLUTION: DIGITAL TALKING BOOKS USING SYNTHETIC
SPEECH UNDER THE NISO/DAISY STANDARD FOR TALKING BOOKS.
Thomas Kjellberg Christensen and Christian Wallin
The Danish National Library for the Blind
Teglvaerksgade 37
2100 Copenhagen
Denmark
Tlf: 45 39 27 44 44
E-mail: tkc@dbb.dk
A complete solution
This paper will describe how an ASCII or a file from a common
word processor can be transformed into a structured Digital
Talking Book (DTB). The DTB will be recorded using synthetic
speech and will be structured in XML using the NISO/Daisy
standard for Talking Books. Human effort will be put in the
preparation of the source file - the rest is an automatic
process. The resulting product is especially usefull for
information where "the human touch" is not of great importance,
e.g. timetabels, dictionaries, manuals etc.
Starting from a text file - ASCII, Word or another wordprocessor
file - the first step is to generate an XML file using the
dtbook3.dtd (Published by the NISO 2001). For this purpose The
Danish National Library for the Blind has developed the XML/HTML
editor ELG (Event List Generator). The ELG features:
- Semi-automated structuring
- No need of XML knowledge
- Fast editing
The resulting XML file is now ready to be processed in EGUSS
(E-book Generating Using Synthetic Speech) - also developed at
Danish National Library for the Blind. This application operates
through the Speech Active Programming Interface (SAPI) which
facilitate the use of Speech Synthezisors in different languages
- even in the same book. Great for dictionaries! EGUSS will
extract information from the XML file and - using the NISO/Daisy
standard for Talking Books - convert it into a Talking Book
consisting of the full text of the publication in XML or HTML,
files that enable precise navigation in the book, audio files and
files for synchronisation (SMIL). All of it in 8 times the normal
recording speed!
At CSUN we will demonstrate the process and produce a short
Digital Talking Book using this method.
Go to previous article
Go to next article
Return to 2001 Table of Contents
Return to Table of
Proceedings
Reprinted with author(s) permission. Author(s) retain copyright.