Re: [orca-list] emacspeak, Dectalk etc...

[Michael Whapples <mwhapples aim com>]

On Sun, 2008-07-27 at 12:26 +0300, Veli-Pekka TÃtilà wrote:
> Michael Whapples wrote:
> > It appears that you can use festival from emacspeak. You will need
> > espeakf (notice the "f", not to be mistaken with espeak).
> Ah thanks. I took a look at the Source Forge page and it seems to me the 
> project is dead at the moment. That is, they said it won't be compatible in 
> the future, and is not maintained:
>
> http://sourceforge.net/forum/forum.php?forum_id=793584

OK, I hadn't looked into how possible it would get running. If there is
sufficient call, hopefully someone else may step forward to maintain it,
or to take it and create a separate project based on it, the bueaty of
open source.
> Too bad Linux has half a dozen minimalist speech Apis the user need be aware 
> of, that are partially incompatible. It is a complex beast, but in WIndows 
> you only have SAPI, universally, and I don't hear people complaining.

SAPI in windows is an example of how having one company in charge
dictating standards can work. SAPI does seem to have been widely adopted
by speech enabled applications and does seem to have been a success,
although it does have its problems. An example of this is that SAPI4 I
don't think allowed specifying the audio output to be used by the speech
synthesiser, and nobody could alter this problem with SAPI4 and so
needed to wait for Microsoft to add it in SAPI5. Also SAPI is not the
only speech API (although the main one known of), each synth tends to
have its own API and some applications do use these (eg. NVDA uses the
espeak API to allow use of the copy of espeak distributed with it, jaws
uses the ECI API used by eloquence to lock that copy to jaws, the
software dectalk window-eyes comes with uses the dectalk API, etc). As a
comparison, Linux does not have one company dictating, so you get
multiple tools tackling similar problems but in slightly different ways
(eg. gnome-speech and speech-dispatcher both try to provide a common API
for using speech synthesisers (like SAPI does), but they both have
different designs and different strong and weak points). This gives the
user/developer the choice to select the system which best meets their
needs. Also for those who feel neither really meet their needs and have
the skills to develop software, they can take these projects and modify
them to provide exactly what they want.
> > I use viavoice now
> Is that in Hardy or some other release? That is, I heard that some lib 
> ViaVoice binaries rely on is not available in hardy, so I wonder if it will 
> even work.

I am using it on slackware. I believe that whilst ubuntu does not have
the correct package in Hardy, I have been lead to believe that it is
possible to get working (voxin is meant to provide the package, and for
ttsynth there are notes about where you can obtain the package and do a
local install rather than from a repository).
> > (whilst viavoice has not got the Finish voices you want
> Umm that's odd since both IBm HOme Page Reader and Eloquence, using the same 
> engine, do have basic Finnish support. howabout Dectalk? I like that more 
> than ViaVoice, though not as much as Orpheus. AT any rate, I could even buy 
> it if I knew it would work, but the Orca Wiki says Orca crashes a lot with 
> DecTalk, aaartgh.

I am sorry about that statement that there is no Finish viavoice, somehow I had remembered wrongly, checking 
both voxin and ttsynth's websites Finish is actually listed.

As for dectalk I cannot speak from experience, but there is a driver for software dectalk in gnome-speech and 
an emacspeak driver for software dectalk. I don't know what languages the Linux dectalk supports, and how 
well it works.
>