WebAIM - Web Accessibility In Mind

E-mail List Archives

Re: best text to speech ap to show my non-geek sighted friends?


From: shawn klein
Date: Mar 2, 2008 4:50PM

OK Phil. I'll be looking forward to trying it when the
instructions are accurate. Yes 400 ms is mighty slow
when you're used to a response time from a screen
reader that seems instantaneous, but I know full well
why it's this way when the screen reader is on a
server half a world away. I'd expect it to be a
virtual tortoise through a dial-up connection, given
the fidelity of the voice, well is the voice streamed,
or is a temporary copy loaded on the user's PC? I'm
wondering how it'll work without an input mode? Do you
just turn off the hotkeys when you want to input
something? Yeah the ability to change the voice, and
the speed of the voice, would be awesome. Later I'll
sign up and try that. I wonder what voice that is? At
first I thought it sounded like Scansoft Emily, but
then I compared them and it doesn't quite. VW Kate,
which is an American voice, and Scansoft Australian
Karen are 2 of the best and clearest voices I've yet
heard personally. It's mostly a process of getting
used to a voice's idiosyncrasies. Eloquence is the
voice I normally use, and it pronounces some things
incorrectly but I've gotten used to the way it does. I
could change every single mispronounced word in the
dictionaries, but I'm a bit lazy in that regard. When
I go to a friend's place where they have a Decktalk
express some pronunciations throw me for a loop, even
though they have no trouble understanding it. In a
pinch I can even understand Spanish or Portuguese read
by US English eloquence, or English read by the
Spanish or Portuguese eloquence voice. It's just what
you're used to. Well actually I'm pretty new to
Firefox so I had to look to see what Firebug was. No I
don't have that one. The only plugins I've actually
manually installed are Adobe flash, and adobe reader.
Well, I don't know which text is cached and which
isn't, but it might be useful to A. have the system
beep when the user has pressed a key but has happened
on blank space with nothing to read, and b. a sort of
rising series of beeps while uncached text is loading,
like the beeps you hear with NVDA while a page is
loading. An audio progress bar so to speak. One more
dumb question, why don't you use arrow key navegation
as well?

>>message truncated ===> Subject: Re: [WebAIM] best
text to speech ap to show
> my non-geek sighted
> friends?
> Date: Sat, 1 Mar 2008 14:18:34 +0000
> To: "WebAIM Discussion List"
> From: "Phil Teare"
> Thank so much for all the feedback.
> Some responses:
> On 01/03/2008, shawn klein < <EMAIL REMOVED> > wrote:
> >
> > Thanks Phil. The blog is interesting, but a bit
> slow
> > to respond. This may just be because my computer
> is
> > 4-1/2 years old with only 768 MB of ram. My
> download
> > speed is 1.5 mb/s, but my upload speed is only 256
> > kb/s, this could also effect things.
> This shouldn't usually be a problem. Text that is
> not yet cached (the blogs
> cache was deleted two days ago, so most of it) will
> take a little while to
> respond, but is always usable - within reason). Text
> that is cached should
> read very quickly (sub 400ms). We're currently
> working on reducing this
> further, and for uncached text). The biggest
> performance hinderance is if
> your using Firebug or similar on FF. Its still
> usable, but much slower if
> you have firebug installed and active. Can I ask if
> you had this or similar
> running?
> I had a bit of
> > trouble understanding some of what the voice was
> > saying in regards to the hotkeys, but I couldn't
> make
> > it spell what I missed.
> I've added to the roadmap. It will make it into 2.5
> all being well (should
> be released at the end of the summer).
> I played it 3 times before I
> > got it, all but one. It seems to be saying that X
> > takes you to the next dive. Although it seems to
> take
> > you to the next line of text.
> Sorry, my fault. Chatterblogs is beta, and the
> instructions are out of
> synch. I'll see to it.
> I had to think for a
> > second before I realized that full stop must be
> > period.
> Interesting point. We're growing in the US (I'm in
> the UK), so I'll make
> sure instructions get more properly localized.
> Although, if you're on a uk
> site with a UK voice, the laguage with be UK
> centric.
> I tested NB, which put me in input mode, but
> > then f8 didn't work.
> Again, instructions out of synch. Sorry. 2.0b is
> looking into simplifying,
> and not using an input mode at all. I'll let you
> know when I've polished
> this, and updated the instructions accordingly, if
> you would like to give
> more feedback, I would be hugely grateful.
> ... may be because I was using
> > Firefox.
> I go to great lengths to make sure FF and IE are as
> feature, function, and
> frills idendical as is possible. Safari will
> sometimes be out of sink, but
> degrade graceful, to have as much function as is
> possible (e.g. it doesn't
> currently have realtime word by word highlighting,
> but FF and IE do).
> Natural voices are cool, but I think they
> > have a ways to go.
> I realy like our voices, but voice easthetics will
> always be very
> subjective. Some of the US voices are a little more
> inteligable, while
> retaining the fuidity and elegance. The Talklet
> system currently allows web
> account users to change the voices centrally, we're
> also looking at allowing
> any user to change voices. Licensing issues are the
> only hold up there
> really, but we're looking at this.
> We also have our own bespoke voices. So larger
> customers can even get their
> own done (which adds to the variety for everyone
> else).
> My
> > biggest issue was some lack of feedback. When you
> are
> > working with something that's slow to react, lack
> of
> > feedback is troublesome. I think that's what it
> was
> > anyway, sometimes I would press x to go to the
> next
> > line and would hear nothing.
> x is deprecated in 2.0. Sorry, again, its the
> instructions on the beta that
> should be changed. I'll get on it. However, we want
> to key the keys a
> minimal as possible, so as to not overwhelm new
> users.
> Hope you found it useful.
> Best
> Phil
> --
> Phil Teare,
> CTO & Chief Architect,
> http://www.talklets.com from Textic Ltd.
> (44) [0] 208 4452871
> >