[Sugar-devel] etextRead activity and espeak

[James Simmons]

Tony,

Read Etexts does *not* have speech disabled.  If you install 
speech-dispatcher and configure it to use espeak (NOT espeak-generic) 
then it will work.  What happens in my code is I try to import 
speech-dispatcher python and if I don't find it then I hide the Speech 
tab on the toolbar.  Hemant Goyal was working on creating RPMs for 
speech-dispatcher for Fedora a few months ago, and it may be possible to 
do a yum install to get it.  It will drag in a bunch of other stuff you 
don't need, though, based on my experiences with the RPMs he gave me to 
try out.

The reason the Activity needs speech-dispatcher rather than just espeak 
alone is that the Python code needs to be called back on word breaks.  
That's the only way to do karaoke highlighting, and espeak by itself 
won't do it.  I would also warn you that the highlighting is not 
perfect.  Sometimes it skips a word.  The way I wrote the code keeps it 
from falling behind and highlighting the wrong word, but it will skip 
one every so often, at least a couple per page.  The folks on the 
speech-dispatcher mailing list seemed to think that festival might work 
better than espeak as far as these callbacks were concerned.

The other thing to warn you about is that multithreaded Python apps 
prevent the power management code from kicking in, or at least this 
happened with the version of Python that the XO was using when I wrote 
the Activity.  This may be resolved now; I can't say.  There was some 
talk here that a newer version of Python would resolve the issue.

I stopped working on these Activities in late September of last year 
when my basement experienced some flooding.  It was January before the 
mold remediation was done and I was able to go back down there to 
program.  Also, I wiped out both Linux installs I was using with espeak 
and replaced them with Fedora 10, so I don't currently have 
speech-dispatcher installed and configured, and it will take me awhile 
to remember what I did.  I was really counting on Hemant to get SD ready 
for the XO, but he isn't responding to my emails these days.

So, in summary, this may turn out to be more of a do-it-yourself project 
than you had hoped for.  I would dearly love to have Read Etexts used by 
kids in Nepal, so be assured that I will do everything I can to help.  
If you want to experiment with what I created the best way would be to 
get an old computer that runs sugar-jhbuild or use Fedora 10 with the 
Sugar packages and install speech-dispatcher.  Get that working before 
you try using an actual XO.

I've been thinking about Read Etexts ever since Kindle 2 was announced 
and I heard it has text to speech built in.  I have always thought that 
really good support for ebooks would make the XO a Kindle-killer.  It 
would cost less, be readable in sunlight, give you access to thousands 
of free books, and when you get one a kid in a foreign land gets one 
too.  What's not to like?

Sorry I can't be more encouraging,

James Simmons


Tony Anderson wrote:
> Hi,
>
> The educational team at OLENepal would like a 'karaoke-style' text 
> reader above all other activities on the XO. This is primarily for 
> English. Looking at the activity code, it looks like the speech code 
> is there but disabled.
>
> Would it be possible to release a version that supports espeak without 
> the Speech-Dispatcher pending its implementation in Sugar? This would 
> give us a chance to experiment with creating 'read-aloud' stories.
>
> Tony
>
>
>
> > Thanks for your response.  I never questioned that there was still
> > interest in TTS on the XO.  What I was wondering is if there was any
> > progress made by Hemant Goyal or anyone else in getting the
> > Speech-Dispatcher software included with the Sugar distribution, if the
> > newer version of Python that resolved the power management issue was
> > included, etc.  I've sent a couple of emails to Hemant and haven't 
> heard
> > back from him.  I was wondering if he was still working on these 
> things,
> > or if someone else had taken over his work, etc.  He was making RPMs 
> for
> > Fedora for installing speech-dispatcher.
> >
> > James Simmons
>