[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Announcing Festival-1.2.95 BETA (fwd)



---------- Forwarded message ----------
Date: Fri, 31 Jul 1998 11:52:09 +0100
From: Alan W Black <awb cstr ed ac uk>
To: festival-beta cstr ed ac uk
Subject: Announcing Festival-1.2.95 BETA


        The Festival Speech Synthesis System version 1.2.95 BETA
         and Edinburgh Speech Tools Library version 1.0.95 BETA
                             31st July 1998

We are pleased to announce the BETA release of version 1.2.95 of Festival
and version 1.0.95 of the Edinburgh Speech Tools Library.

This announces a BETA release and is sent to you because you
are working closely with us or have expressed interest in early
releases.  If you have time we would be grateful if
you could download this version of the system and test it
in your environment.  We will try to incorporate any feed
back in the full release which we hope release in two weeks.

This release is available from 
    ftp://ftp.cstr.ed.ac.uk/pub/prot/BETA-festival-1.2.95/

Please don't redistribute this version as it is still incomplete, but
feel free to pass this url on to people you think will be interested
in this new version 

I will add a file LATEST in that directory with any fixes, notes
etc.  I will also release .96 if necessary, but will mail the group
if that happens.

You bascially need all the files here to run the tests, though only
one of the 16k or 8k versions of rab and ked diphones.

We are particularly interested in compilation/configuration problems,
as well as integration with your existing modules.  If you can (or
cannot) successfully compile the system and make it pass its tests
(gnumake test in the speech_tools/, and festival/ directories) we
would be like to know.

Please address any questions directly to awb cstr ed ac uk, though
questions about the Windows port should go directly to
rjc cstr ed ac uk

Thanks for your help

Alan, Paul and Richard
31st July 1998

------------------------------------------------------------------------

New in this version
   * New (improved) diphone synthesizer
   * Better letter to sound rules (trained) for lexicons
   * New utterance architecture
   * Probabilistic parser
   * Improvements in training tools
   * Improvements in signal processing routines 
       (can now build diphone databases from our tools alone)
   * Tilt intonation model support
   * Time and space efficiency
   * Sable/XML markup language support (optionally built in)
   * Basic Java bindings (loading Speech Tools into Java and
       java client access to festival server)
   * Many other fixes

What has still to be done before the full release
   * Documentation re-read and correction, still much of
       it is old or placeholders
   * Tilt training and prediction need to be re-integrated
   * Windows NT versions needs some more testing
   * Java examples and initial JSAPI-like interface
   * Sable updates to 0.2 (and aligned with ICSLP paper)
   * Documentation of Cluster Unit selection example (and completion)
   * Removal of some development code

We have successfully run the system under the following systems

Sun Sparc Solaris 2.5.1/2.6:
   GCC 2.7.2, GCC 2.6.3, GCC 2.8.1, SunCC 3.01, SunCC 4.0
Sun Sparc SunOS 4.1.3:
   GCC 2.7.2
Intel Solaris 2.5.1:
   GCC 2.7.2
FreeBSD for Intel 2.1.7 and 2.2.1:
   GCC 2.7.2.1, GCC 2.6.3
Linux (2.0.30) for Intel (RedHat 4.[01]/5.[01]):
   GCC 2.7.2, GCC 2.7.2/egcs-1.0.2
Windows NT 4.0:
   GCC 2.8/egcs ??? (from Cygnus GNU win32 b19), Visual C++ 5.0.

We are particularly interested in tests under new OSs such as
IRIX, OS/2, HPUX, under which previous versions of the system did
run and we feel these should be close.

As for new C++ compilers, good luck, it has been harder this time
to get a version that works under all compilers as the compilers
available seem to be more different than they were last year.

Installation instructions for people who don't read INSTALL and
are familiar with installing software.

   gnutar zxvf speech_tools-1.0.95.tar.gz
   cd speech_tools/configp
   cat config-dist >config  (edit it? for READLINE)
   cd ..
   gnumake info    ( will display what we think the configuration is)
                     if its not ok read the INSTALL information about
                     changing config/config or read config/ReadMe)
   gnumake 
   gnumake test
   gnutar zxvf festival-1.2.95.tar.gz
   gnutar zxvf festvox_rablpc16k.tar.gz
   gnutar zxvf festvox_kedlpc16k.tar.gz   (or the 8k versions)
   gnutar zxvf festvox_don.tar.gz
   gnutar zxvf festlex_CMU.tar.gz
   gnutar zxvf festlex_OALD.tar.gz
   gnutar zxvf festlex_POSLEX.tar.gz
   cd festival/config
   cat config-dist >config (almost everything comes from EST so 
                            youn probabaly don't need to edit this)
   cd ..
   gnumake info
   gnumake
   gnumake test

Of course this a assumes a normal machine (i.e. one where this works :-)
See speech_tools/INSTALL and festival/INSTALL for real details though
admittedly they are still a littel rough.





   
   



[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]