Re: [orca-list] speech dispatcher - still confused

[Hynek Hanke <hanke brailcom org>]

Dear Derek,

one of the problems you face is that Speech Dispatcher
is not well integrated into the distributions, so you need
to do some installation and settings of things you should
not normally be asked to do manually. There is definitely
work waiting to be done to make these things easier.


> I've been having trouble with port numbers changing between 6560 and 
> 6561. What files is that variable set in?
Speech Dispatcher is a server and Orca or other applications are clients.
The port is a communication channel, so both the server and the client
need to know on which port they can meet. In Speech Dispatcher you
can set the port in the speechd.conf file. The default for system wide
Speech Dispatcher is 6560 and user Speech Dispatcher is 6561. Depending
on what Speech Dispatcher you want to run/use in the client you need
to tell the client which port it should communicate with it. You can either
do it in the settings dialogs in the client (I don't know if it is possible
with Orca) or set the environment variable SPEECHD_PORT, which should
be respected by all clients. One way to do this is to add the line

    export SPEECHD_PORT=6561

into your ~/.xsession file. This file needs to terminate with the line

    gnome-session

or similar.

You need to restart your gnome-session then and can check what is
the current state of the variable by executing in the terminal

env | grep SPEECHD_PORT
> Also, I know there are several different sound systems I can select 
> from. Do I need to use alsa, oss, or pulse?

You can use either of them, but ALSA is currently the most convenient 
option.

> Also it seems to be built around espeak

No, espeak is the default option, but not in any way the only one.

> as when I managed to get it to connect I heard my test message from 
> espeak, nothing from ibmtts
Either you have set espeak as the default or your ibmtts is not working.
Check with

spd-say -o ibmtts "Is this IBM TTS or espeak?"

If you hear espeak not IBM TTS, your IBM TTS is not working with Dispatcher.
> , and an error message from a "dummy module" when I didn't uncomment 
> the addmodule lines in the speechd.conf. As for orca, I've looked at 
> debug.out and see where speech-dispatcher interface is not installed. 
> apt-get install speech-dispatcher and apt-get install python-speechd 
> were the two things I saw that I needed. Is there perhaps more?

I don't think so. Very likely, Orca just didn't see Speech Dispatcher
because you didn't set the SPEECHD_PORT environment variable.

>  I'm curious if ibmtts is supported by speech dispatcher or if just 
> the framework is in place and the support has yet to be written in.
IBM TTS is supported and should work. There are some problems
with it but they are difficult for us to fix given that IBM TTS is a
proprietary not free software, it needs special old libraries, doesn't
have a 64 bit binary etc.

> I see things like "A command like spd-say hello in a terminal should 
> cause espeak to say the word hello."
I think this is a line from the manual for instalation procedure where
it is meant that at the early stage, espeak is the default synthesizer,
not that espeak is the only synthesizer to use with speechd once you
have it up and running.
> There, espeak is mentioned as if it should be the synthesizer that is 
> used.

No.

> Will I need to do further editing to make it talk while other audio is 
> being played?

No, this should be the default behavior with ALSA (not OSS).

With regards & hope it helps,
Hynek Hanke