Linux Today: Linux News On Internet Time.

Festival Speech Synthesis System 1.4.0

Jun 26, 1999, 22:11 (4 Talkback[s])
        The Festival Speech Synthesis System version 1.4.0
         and Edinburgh Speech Tools Library version 1.2.0
                            20th June 1999

Festival offers a general framework for building speech synthesis
systems as well as including examples of various modules.  As a whole
it offers full text to speech through a number APIs: from shell level,
though a Scheme command interpreter, as a C++ library, from Java, and
an Emacs interface.  Festival is multi-lingual (currently English
(British and American), and Spanish) though English is the most
advanced.  Other groups release new languages for the system.

The system is written in C++ and uses the Edinburgh Speech Tools
Library for low level architecture and has a Scheme (SIOD) based
command interpreter for control.  Documentation is given in the FSF
texinfo format which can generate, a printed manual, info files and

- From this version, the system is now free software and is distributed
under an X11-type licence allowing unrestricted commercial and
non-commercial use alike.

This distribution includes:
   * Full English (British and American English) text to speech
   * Full C++ source for modules, SIOD interpreter, and Scheme library
   * Lexicon based on CMULEX and OALD (OALD is restricted to non-commercial
     use only)
   * Edinburgh Speech Tools, low level C++ library
   * British English Male (for residual LPC resynthesis 8k and 16k versions)
   * 2 American English Male (for residual LPC resynthesis 8k and 16k versions)
   * 4 other voices using MBROLA based diphone synthesis (1 British Male,
     2 American Males and 1 American Female).
   * Castilian Spanish Male (for residual LPC resynthesis 11k version)
   * British English Male (for spike excited LPC resynthesis 10k version)
   * Full documentation (html, postscript and GNU info format)

Festival version 1.4.0 sources, voices, and Linux binaries are
available from
and also from the US mirror at CMU

The Festival home page, offering descriptions of the system, examples
and online demos, can be found at


To run Festival you need:
   * A Unix machine, Festival has compiled and run on Suns (SunOS and 
     Solaris), FreeBSD, Linux, SGIs and DEC Alphas but should be portable
     to any standard Unix machine.
   * A C++ compiler: we have used GCC  version 2.7.2, 2.8.1
     SunCC 4.1 nd 4.2, and egcs.  Other C++ compilers are
     probably possible with perhaps some minor chanages
   * GNU Make any recent version
   * Audio hardware, /dev/audio (8 bit and 16 bit on Suns, Linux 
     and FreeBSD) and NCD's NAS network transparent audio system 
     are supported directly but Festival supports the execution of 
     any Unix command that can play audio files.

There is also support for building the system under Windows NT and
95/98.  We have successfully ran the system complied with Cygnus' GNU
win32 package and Microsoft's Visual C++, instructions are included

We intend to distribute binary distributions for major
architectures particularly Linux and Solaris.  We will *not*
distribute binary versions for Windows NT/95 as we do not have enough
expertise to support them.

New in 1.4.0 version
   * The copyright has changed.  The system is now free.
   * Substantial low level tidy-ups in the Utterance structure
     * Stronger notion of Relation types (lists, trees, mls, and others)
     * XML i/o for relations (good for trees and more complex data exchange)
   * Three new US English voices (simply front ends to MBROLA voices)
   * Features values can now be anything including Features and
     other objects can be cleanly added.
   * SIOD is now "in one directory" rather than one defun
     It can also be used standalone, and adding user objects is easier and
   * New documentation system for Speech Tools, integrating code,
     descriptions and examples (DOC++, Jade, docbook based)
   * Generalization of models (WFSTs, Ngrams, CART, viterbi)
     allowing arbitrary use from Scheme.
   * Initial JSAPI (and JSML) support, stadardising the existing Java
   * "Building Voices in Festival" document describing process of building
     new voices in the system
   * Many other fixes

Alan, Paul, Richard and Rob
20th June 1999

Alan W Black                            email: awb@cstr.ed.ac.uk
Centre for Speech Technology Research   http://www.cstr.ed.ac.uk/~awb
University of Edinburgh                 tel:   (44) 131 650 2787
80 South Bridge, Edinburgh, UK          fax:   (44) 131 650 6351