Newsgroups: comp.speech
Path: pavo.csi.cam.ac.uk!pipex!uunet!haven.umd.edu!darwin.sura.net!jvnc.net!princeton!phoenix.Princeton.EDU!lseltzer
From: lseltzer@phoenix.Princeton.EDU (Linda Ann Seltzer)
Subject: Re: Phoneme prog/data for SGI Indigo?
Message-ID: <1992Oct7.145716.20508@Princeton.EDU>
Originator: news@nimaster
Sender: news@Princeton.EDU (USENET News System)
Nntp-Posting-Host: phoenix.princeton.edu
Organization: Princeton University
References: <1992Oct5.230617.20423@news.iastate.edu>
Date: Wed, 7 Oct 1992 14:57:16 GMT
Lines: 17

>Does anybody have a Phoneme-based or just plain
>ANY speech generation program for the Indigo?

There are two parts to a speech synthesis system: (1) the program
that synthesizes the samples and (2) the rules which provide the
proper information to the synthesizer program.  Dennis Klatt's
1980 JASA paper contains Fortran code that you can translate to
C for generating the sound samples.  There are errors in the
paper and in the Fortran program, though.  As for the rule program,
I don't know of any publicly available software.  It takes so many
years for a group of linguists to develop it, that it's not the
sort of thing one obtains for free.  You can buy speech synthesis from
Centigram in San Jose.  As long as the synthesis program writes samples
into a file in a format the Ingido knows about, then any rule program
in a language compilable on the Indigo should work - your synthesis
program just has to read the output.

