INSTALL

The requirements:
=================

You will need these components to compile Speech Dispatcher:  
 - glib 2.0  (http://www.gtk.org)
 - libdotconf 1.3 (http://github.com/williamh/dotconf)
 - libsndfile 1.0.2 (http://www.mega-nerd.com/libsndfile/)
 - pthreads
 - gcc and GNU make (http://www.gnu.org)
 - festival-freebsoft-utils 0.3+ (optional)
	(http://www.freebsoft.org/festival-freebsoft-utils)
 - Python 3, PyXDG (optional)
        (http://www.python.org)

Only if you are building from Git, you also need these:
 - automake 1.11+, autoconf, autopoint, pkg-config, libtool, makeinfo, texinfo
   and gettext

We recommend to also install these packages:
 - PulseAudio (http://www.pulseaudio.org)
 - Festival (http://www.cstr.ed.ac.uk/projects/festival/)
 - Espeak (http://espeak.sourceforge.net/)
 - Gettext 0.19.8+ (https://www.gnu.org/software/gettext/) for internationalization support.

These packages are known to work with Speech Dispatcher:
 Software synthesizers:
 - Festival (http://www.cstr.ed.ac.uk/projects/festival/) -- an extensible
     speech synthesis engine with support for various languages (English,
     Spanish, Italian, Finnish, Czech, Russian, Hindi, some African languages
     and others)
 - Espeak (http://espeak.sourceforge.net/) -- a light and fast multi-lingual
     synthesizer with very good support for various Speech Dispatcher
     features.
 - Flite, Cicero, IBM TTS, Dectalk Software, Ivona, SVOX Pico

 User applications:
 - Orca (http://www.gnome.org/projects/orca/) -- AT-SPI screen reader (Gnome)
 - speechd-el (http://www.freebsoft.org/) -- Emacs speech interface
 - Linux Screen Reader (LSR (http://live.gnome.org/LSR)
 - BrlTTY (http://mielke.cc/brltty/)
 - ...

Building Speech Dispatcher:
===========================

(for instructions about how to build from Git or Git snapshot see below)

If you want to use Speech Dispatcher with Festival (recommended), you
also need to install festival-freebsoft-utils version 0.3 or higher!

If you want to use Speech Dispatcher with Ivona, you need to install
libdumbtts

To compile the whole project, including Speech Dispatcher server, all modules
and clients, simply extract the distribution archive and change to 
speech-dispatcher-<version> directory and run "make all" command as follows:

 $ tar -xzvf speech-dispatcher-[version].tar.gz
 $ cd speech-dispatcher-[version]


Notes about Pulse, LIBAO, ALSA, OSS and NAS support
============================================

Speech Dispatcher's default audio output system is the first discovered
during configuration started with the Pulse Audio and followed by libao, ALSA,
OSS and NAS. It is possible to explicitly setup default audio output system
while building Speech Dispatcher, see configuration options.
This value can be redefined during run time in the speechd.conf.
If multiple audio output systems separated by comma are defined as default
audio output system e.g. "pulse,alsa" an automatic fallback will be done.

Notes about Espeak support
==========================

You need to install Espeak libraries to be able to use the native
output module. Some distributions like debian provide
those libraries in a different package (libespeak1 and libespeak-dev).

Notes about Flite support
=========================

You don't need to do anything special to compile Speech Dispatcher
with Flite support, Flite will be detected automatically when running
./configure. However, make sure that you have correctly installed
flite libraries and examine the output of ./configure carefully.

You should have configured flite with
    ./configure --enable-shared
to be able to use the flite plug-in. If you aren't sure if you have done this,
check for example for libflite_cmu_us_kal.so in your library directory.
(It's also possible to use Speech Dispatcher with Flite static libraries,
but it is generally not a good way unless you know what you are doing.)

Notes about IBM TTS support
===========================

In order to build the propriatary (non-free software) ibmtts output
module, you will need one of two things:

  1.  An install of the proprietary commercial IBM TTS synthesizer (for Linux),
or

  2.  The sdk for IBM TTS, which contains the headers and a "stub library" for
      linking against. This SDK is Free Software and freely available, however
      it doesn't contain the synthesizer. It will enable you to build the
      module, but this module will give you no sound until you install the
      synthesizer itself.

The sdk can be obtained here:

      http://ibmtts-sdk.sourceforge.net/

      It isn't obvious on that website, but you download the sdk using cvs:

      cvs -d:pserver:anonymous@ibmtts-sdk.cvs.sourceforge.net:/cvsroot/ibmtts-sdk login 

      cvs -d:pserver:anonymous@ibmtts-sdk.cvs.sourceforge.net:/cvsroot/ibmtts-sdk co sdk

      The sdk is licensed using a BSD-style license.

You will also need libsndfile in order to enable sound icon emulation with the
ibmtts output module.  (As of Apr 2006, the Debian packages you need are
libsndfile1 and libsndfile1-dev.)

Notes about SVOX Pico support
=============================

SVOX Pico software can be obtained from
http://git.debian.org/?p=collab-maint/svox.git

SVOX Pico documentation can be found under
http://android.git.kernel.org/?p=platform/external/svox.git;a=tree;f=pico_resources/docs
It includes three manuals:
- SVOX_Pico_Lingware.pdf
- SVOX_Pico_Manual.pdf
- SVOX_Pico_architecture_and_design.pdf

To build SVOX Pico, fetch the debian-sid branch with

git checkout debian-sid

then apply Debian patches with

QUILT_PATCHES=debian/patches quilt push -a

then go to the pico subdirectory and type
chmod +x autogen.sh
./autogen.sh
./configure
make
make install

Notes about cxxpiper support
============================

The cxxpiper module uses the Piper synthesis
https://github.com/rhasspy/piper/

Upstream piper however does not provide an API that can really be used nicely,
so the build is quite convoluted. The following should be working:

wget https://github.com/rhasspy/piper-phonemize/releases/download/2023.11.14-4/piper-phonemize_linux_x86_64.tar.gz
sudo tar -C /opt -xvf piper-phonemize_linux_x86_64.tar.gz
wget https://github.com/rhasspy/piper/archive/refs/tags/2023.11.14-2.tar.gz
sudo tar -C /opt -xvf 2023.11.14-2.tar.gz
CXXFLAGS=-I/opt/piper_phonemize/include LDFLAGS=-L/opt/piper_phonemize/lib ./configure --with-piper=/opt/piper-2023.11.14-2
make
make install

Continue building Speech Dispatcher
===================================

On some systems, before running make, you have to set the environment variable
SED to your binary of the sed application. This is usually done as:

$ export SED=sed

$ ./configure
$ make all

If everything worked ok, you can install all the parts to the default
path by running (as root):

 # make install
 # ldconfig

If it doesn't work, please see "Building from Git" if this is your case. If
not, you can ask us on speechd-discuss@nongnu.org .

Building from Git
=================

If you are building directly from Git or from some Git tarball snapshot, you
must first create the ./configure file. Please make sure you have all the
necessary tools listed in the requirements at the beginning of this README
in their appropriate versions. Automake and autoconf are especially version
sensitive.

Then please do:

 $ cd speech-dispatcher # where speech-dispatcher is your Git clone
 $ ./build.sh

Now the configure file should be created and you can proceed like with
an ordinary instalation.

Installing Festival Speech Dispatcher interface:
================================================

You need to install the new version of festival-freebsoft-utils (0.3
or higher).

Please make sure that Festival server_access_list configuration
variable and your /etc/hosts.conf are set properly. server_access_list
must contain the symbolic name of your machine and this name must be
defined in /etc/hosts.conf and point to your IP address. You can test
if this is set correctly by trying to connect to the port Festival
server is running on via an ordinary telnet. If you are not rejected,
it probably works.

Example of the server_access_list configuration line
in festival.scm:
(set! server_access_list '("127.0.0.1" "localhost" "chopin"))

Testing
=======

You can try to execute Speech Dispatcher by
 $ speech-dispatcher

If Speech Dispatcher is running, you can test if it's working by:
 (on some other console)
 $ cd speech-dispatcher-[version]/src/tests
 $ make check
 $ ./run_test basic.test
(You can also try other tests, if you like.)

Also try
 $ speech-dispatcher --help
to see how you can run it as a standalone program and possibly see
what's wrong.

Troubleshooting
===============

If you think something is missing, try additionally:

 $ mkdir PREFIX/etc/speech-dispatcher
 $ cp -r config PREFIX/etc/speech-dispatcher
 $ mkdir PREFIX/include
 $ cp clients/libspeechd.h PREFIX/include/

where PREFIX is your installation path prefix (defaulting to /usr/local).

Building in more detail:
========================

The hierarchy of Makefiles is used to build each of the parts. The top
level Makefile allows you to compile some extra parts of the project.
These are:

    - doc ........ build documentation in all supported formats
                   (see doc/README for more information)
    - src/server ..... build speachd server - this is the same as
                   cd src/server; make all
    - src/modules .... build all speachd modules - this is the same as
                   cd src/modules; make all
    - src/clients .... build all clients - this is the same as
                   cd src/clients; make all
    - clean ...... remove all files except the sources (does not
                   clean documentation)
    - all ........ build server, modules and clients

If you are building directly from git or from some git tarball snapshot, you
must first create the ./configure file. Please make sure you have all the
necessary tools listed in the requirements at the beginning of this README
in their appropriate versions. Automake and autoconf are especially version
sensitive.

Then please do:

 $ cd speech-dispatcher-[version]
 $ ./build.sh

Now the configure file should be created and you can proceed like with
an ordinary instalation.



Copyright (C) 2001-2012 Brailcom, o.p.s
Copyright (C) 2010 Rui Batista <ruiandrebatista@gmail.com>
Copyright (C) 2010 William Hubbs <w.d.hubbs@gmail.com>
Copyright (C) 2010-2016 Andrei Kholodnyi <andrei.kholodnyi@gmail.com>
Copyright (C) 2015 Luke Yelavich <themuso@themuso.com>
Copyright (C) 2018-2020, 2025 Samuel Thibault <samuel.thibault@ens-lyon.org>

This program is free software; you can redistribute it and/or modify it under
the terms of the GNU General Public License as published by the Free Software
Foundation; either version 2 of the License, or (at your option) any later
version.

This program is distributed in the hope that it will be useful, but WITHOUT ANY
WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A
PARTICULAR PURPOSE.  See the GNU General Public License for more details (file
COPYING in the root directory).

You should have received a copy of the GNU General Public License
along with this program.  If not, see <https://www.gnu.org/licenses/>.