software speech on Slackware, Debian, and others?

[Chuck Hallenbeck]

David,
I can address some of your questions with reference to Slackware 10.0,
but not all of them.

> 1. I use the "mkinitrd" command to put the "sftsyn" module into the kernel
> at boot time.  The command was:
> "mkinitrd -v /boot/initrd-2.6.x.img 2.6.x --with=speakup_sftsyn
> --with=pcspkr"
>
>    A. Can this be applied to other distros?
>    B. Does "mkinitrd" act different (how) in other distros?
>    C. Should I use another (more portable) approach?
>    D. Do other distros have the PC speaker (pcspkr) compiled as a module?

The above steps are not ones I have had to deal with, and cannot help
with. Sorry.

> 2. I created a /dev/softsynth device using the "mknod" command.  The
> command was:
> "mknod /dev/softsynth c 10 26".
>
>    A. Is this ok for other distros?
>    B. Is the syntax the same?


Yes, and yes.
>
> 3. Festival was already instaled and working in Fedora.
>    A. Any problems getting Festival for other distros?

Yes. I am unable to compile speech_tools, required for compiling
festival, with the GCC 3.3.4 that comes with SW 10.0, and since Flite
works fine, I have not wanted to revert to an earlier GCC.

>
> 4. Flite was not installed in Fedora.  I had to download RPMS, because it
> would not compile properly...
>    A. Other distros?

Flite compiles fine on SW 10.0, once the recommended patch was applied.
Without the patch there is a similar problem with GCC versions as above.
HOWEVER: Compiling Flite takes a hell of a lot of ram and/or swap space.

>
> 5. I downloaded speech-dispatcher, compiled it, and installed it without
> any problems.
>    A. Anyone have problems doing this?

Nope. That went fine. I had to be sure to get version 0.5rc3 though to
make it work. Had a lot of trouble with 0.5rc1 and 0.5rc2.

>
> 6.  speechd_up compiled and installed without problems also...
>    A. ...?

No problem here.

>
> 7. Just ran:
>    A. # speech-dispatcher
>    B. # speechd_up
> I had software speech working at this point...

Here is what I did to get software speech going:

I recompiled the default kernel selecting "sftsyn" as a module rather
than compiled into the kernel. Not sure if this was still necessary; it
used to be necessary. I also selected "ltlk" into the kernel and
selected it as my default synth. So it comes up talking on the
Doubletalk external, which I normally just silence once I see it is
working.

In my /etc/rc.d/rc.local file, I have included these statements:

export PATH=/usr/local/bin:/usr/local/sbin:$PATH
speech-dispatcher
speakwith sftsyn

These statements start speech-dispatcher, which remains running
thereafter. Then the "speakwith" script is run with the argument
"sftsyn" to switch the ltlk out and the sftsyn in. The script may later
be run (as root) to switch to any other supported and installed synth.
The script is available from:

http://www.mhcable.com/~chuckh/speakwith

One more detail, the speakwith script requires Steve's "speakupcomf"
script to be moved from the speakup directory to /usr/local/bin, where
it can restore user selected settings whenever a synth change is made.
Of course speakwith also installs and removes speechd-up when needed.

HTH

Chuck



-- 
The Moon is Waning Crescent (1% of Full)
Home page at http://www.mhcable.com/~chuckh
Speakfreely at hhs48.com:2074