Re: [orca-list] gnome-speech seems to block orca

[Michael Whapples <mwhapples aim com>]

I will try with the espeak driver, but I won't be trying on the machine
I am using to write this email as I use this machine for work and don't
want it to go wrong (although I am confident I would be able to do
everything fine).

Regarding the cursoring through emails, as an example, if I cursor on to
a message and here who it is from (just the name not the email address,
so for you it would be "Willie Walker") and then at that point I know it
isn't the message I am looking for, so I press the up arrow again to go
to the message before (but as my actions aren't perfectly instant, it
has just started speaking the email address, may be "<William" before
the key is actually pressed) then the speech will continue to read the
email address, and then will speak the one I moved to, but the Braille
doesn't move until the one I move to has had the email address spoken. I
don't know whether you would call that cursoring fast, but it certainly
is a way that one may use the system.

I have noticed delays in response else where when cursoring, one I know
of (because I use it so much) is eclipse in the code editor.

I have never noticed this sort of problem when using speech-dispatcher.
The nearest it has been with speech-dispatcher is when using firefox
that if cursoring fast (I mean to the point where speech may not be
output for a line) orca has been known in the past to drag the cursor
back to where it was before that load of fast cursor movement.

Michael Whapples
On Thu, 2008-06-12 at 10:30 -0400, Willie Walker wrote:
> In looking at the gnome-speech driver for ViaVoice (which I think what 
> is used for the IBMTTS engine), I see calls to ViaVoice's eciAddText 
> method directly in the 'say' method.  If the call to eciAddText can 
> block, then this could be what is causing the problem.
>
> We saw something similar with the gnome-speech driver for eSpeak, and I 
> worked up a patch yesterday to try to avoid the problem: 
> http://bugzilla.gnome.org/show_bug.cgi?id=535493 (the patch only 
> applies to the gnome-speech driver for eSpeak).
>
> Do you have other speech engines you can try?  If eSpeak is one of 
> them, you'll need to apply the patch from the above bug first and 
> rebuild/reinstall gnome-speech.  If you're not familiar with doing 
> that, however, then don't do it.  I often forget how dangerous it is to 
> place a knife in someone's hands and tell them to cut a piece of rope 
> -- it's amazing, but somehow they sometimes end up slashing their 
> wrists and hanging themselves in the process.  :-)
>
> In addition, you mention that you experience this when arrowing through 
> the list of e-mails in evolution.  Are you arrowing really really fast? 
>   Do you experience this in other applications?  Is it associated with 
> the speed at which you are typing and/or if you press and hold a key to 
> cause autorepeat to kick in?
>
> Will
>
> On Jun 12, 2008, at 8:24 AM, Michael Whapples wrote:
>
> > Hello,
> > I will try and word this carefully so it doesn't get misunderstood.
> >
> > I am using gnome-speech with IBMtts, and I have modified the bonobo
> > server file
> > (/usr/lib/bonobo/servers/GNOME_Speech_SynthesisDriver_Viavoice.server)
> > to use the Viavoice driver with ALSA's oss emulation.
> >
> > When using orca, at times speech starts and it seems impossible to
> > interupt the speech. It appears as if the speech output is blocking
> > further execution in orca, until the speech finishes, or possibly when
> > it is reaching a join of speech strings sent to gnome-speech. An 
> > example
> > I can give is when I am cursoring through the list of emails in a 
> > folder
> > in evolution, orca starts speech of the item I move to, braille does 
> > not
> > change at this point, and no key presses seem to respond, until a good
> > way through the row of the list item being spoken, at which point
> > braille catches up and keys start responding. In fact when I say key
> > presses don't respond, I mean orca reports nothing as happening, it
> > continues what it is speaking as if nothing happened, but at the end 
> > the
> > key presses then all seem to happen.
> >
> > I have never noticed this with orca and speech-dispatcher, but at the
> > moment gnome-speech seem better (a part from this problem) as there are
> > a couple of things with speech-dispatcher and the IBMtts module that
> > make it not so good.
> >
> > Is this problem an orca bug or a gnome-speech bug? Does a bug report
> > already exist for this, or shall I create one.
> >
> > Michael Whapples
> >
> > _______________________________________________
> > Orca-list mailing list
> > Orca-list gnome org
> > http://mail.gnome.org/mailman/listinfo/orca-list
> > Visit http://live.gnome.org/Orca for more information on Orca