Software
for the Visually handicapped
LLF2SPCH : Utility to speak out a local language file.
This application can
be invoked from the command line to speak the contents of a file prepared
using the IITM editor. Reading the lines can be effected with some control
to help a visually handicapped person go through a file in a desired way.
When invoked, the
application gives out information about the number of lines in the file
and offers a choice of two modes. The contents can be read continuously
line after line or one line at a time. In the continuous mode, the user
can abort reading and resume from the same line or go a specific line.
In the line reading mode, a line can be read again and again if necessary.
The application will
read local language text as well as English. The two can be mixed in the
document. This is accomplished by automatically identifying the text and
switching the language for synthesis. The MBROLA speech engine permits
dynamic selection of the speech data base and this feature is effectively
utilized in the application.
Controls are provided
for the user to adjust the rate of speech.
The llf2spch application
can be run under Linux as well as Microsoft Windows 98/Me/2000/XP. The
application is invoked from the command line so as to help the visually
handicapped input the command from the keyboard.
shell prompt> llf2spch
<input file>
The MBROLA speech
engine is required to be installed in the system for the application to
work. The English Data base (EN1) is used for synthesizing English and
the Swedish Data base (SW1) for local language.
The application will
create a temporary text file containing the phonemes present in the synthesized
output. This can be saved and subsequently used directly with MBROLA to
generate examples of synthesized speech.
The llf2spch application
used adhoc prosody models for generating the speech. While this does not
guarantee natural sounding speech, the result is quite satisfactory for
the visually handicapped.
