
This is Epos, a freely distributable language independent Text-to-Speech
system. See doc/COPYING for distribution conditions.

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

Compiling and installation on a UNIX machine:

 1. ./configure
 2. make
 3. make install   (with sufficient rights to install cfg files under /usr/local/share/epos)
 4. install speech segment inventories (see cfg/inv/README for an explanation;
	untar them starting from /usr/local/share/epos/inv)

Running on a UNIX machine:
 eposd       starts a speech daemon  
 say-epos    controls this daemon, e.g. "say-epos 'Some text'" means "tell eposd to
	     say 'some text'".

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

Compiling with Visual C++ and installation on a Windows NT (Windows 2000 etc.) machine:

 1. cd arch\win
 2. copy cpconv2.exe into arch\win if you require the conversion of
	configuration files into code page 1250 (this is optional)
 3. .\configure.bat
 4. cd ..\..\src
 5. run VC and open the workspace file epos.dsw
 6. build all the projects - at least instserv, eposd and say-epos
 7. install speech segment inventories (see cfg/inv/README) if required
	(copy them to the cfg\\inv directory and untar them there)
 8. run instserv.exe (it should compile into the same directory as eposd.exe,
    else it won't find eposd.exe) with sufficient rights to install Epos
    as a Text-to-Speech service and to write a pointer to its configuration
    to the registry
 
Then try to run "say-epos.exe some text" with sufficient privileges to hear Epos speak.

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

The default configuration speaks Czech using the local sound card and
the public TTSCP server at epos.ure.cas.cz. If you don't have a sound
card, you might wish to try "say-epos -w 'Some text'" to create said.wav
which is an MS Windows-style .wav file. If you are not connected to
the Internet, or epos.ure.cas.cz is unreachable or misconfigured 
Epos will revert to a low quality voice distributed with Epos for this
purpose.  (In some versions of Epos the default voice is supposed
to be installed locally.  Be creative.  Use "say-epos -l" to find out
the order of the voices.  Use "say-epos --voice xxxxxxx 'Some text'" to
use a specific voice.)

System requirements:

 Any decent C++ compiler, for example gcc 2.8 or later.
	Good choices also include MS Visual C++, Watcom C++; otherwise
	"decent" means "recent".
 (If using the GNU assembler, gas version 2.9 or newer is recommended.)

More info:

 1. read doc/*.sgml (or, in the compiled form: epos.dvi or epos.html or epos.txt)
 2. there are other relevant files under "doc"
 3. browse http://epos.ure.cas.cz/ (mostly identical to (1))
 4. have a look at the files under cfg (or /usr/local/share/epos)
 5. read the source under src
(The last two steps are optional.)

Getting help:

 0. The order of the following steps is important.
 1. Download a recent version.
 2. Get more info as outlined above.
 3. Mail to epos@speech.cz
    Mention the version you're running or trying to run
    (to be found in common.h or by "eposd -v") and describe your
    problems. If your mail bounces, you might try geo@cuni.cz instead.

Bug reports:

	If you think you found a technical bug, you can send it
	directly to epos-dev@speech.cz.  This is a developer
	list.  The list language is English. Consider subscribing
	to the list first so that you miss no answers if somebody
	forgets to CC you.  Again, be sure to mention the version
	you are trying to run.
