Re: emacspeak speech server

[tcross@xxxxxxxxxxx]

Dmitry Paduchih writes:
> 
> I have read several postings here stating that speech servers provide
> an abstraction of some generalized speech device for emacspeak. This
> point being amusingly simple and attractive isn't so evident for me
> however. Indeed, emacspeak distribution contains the modules
> 
> dtk-tcl.el
> outloud-css-speech.el
> outloud-voices.el
> 
> and I personally use modules
> 
> mbrola-css-speech.el
> mbrola-voices.el
> 
> which came separately from main emacspeak distribution.
> 
> At least for their names I would assume that emacspeak contains some
> knowledge about implementation related issues: dtk, tcl, outloud, and
> mbrola.

I don't think this disputes the conceptualization model at
all. However, there is a bit of confusion regarding terminology and
exactly what is being referred to by the term emacspeak and speech server.

I will try to clarify how I interpret it. I consider emacspeak itself
to comprise mainly of the lisp source files which start with
the word emacspeak-. The modules you refer to I see as the interface
layer between emacspeak and the speech servers. The fact the emacspeak
package comes with some speech servers, such as those for the dectalk
or outloud does not to my mind break the layered conceptual model. The
fact you use a server which does not come with emacspeak even supports
this model because it indicates that different speech servers can be
used with emacspeak without any need to modify the emacspeak code.
With respect to emacspeak knowing about implementation issues of
speech servers, I disagree ith your argument. Emacspeak itself knows
nothing about mbrola or even tcl for that matter (there have even been
servers written in C which have worked with emacspeak and have not
relied on tcl at all). All that emacspeak cares about is that all
speech servers provide the same interface - emacspeak itself does not
care if you are using outloud, dectalk, mbrolla, doubletalk or
whatever and it knows nothing about the individual differences in
functionality between these different servers such as the ability to
use multiple voices or tones etc.

> I also can't consider layer structure to be the main difference
> between emacspeak and screen readers for MS environment because audio
> environments in MS world do essentially provide very similar
> structure. Take for example Jaws:

I argued that emacspeak was different from screen readers under
windows because it was able to perform speech 'markup' - that is, use
different voices and tones/sounds to indicate different contectual
information such as headings, keywords in programming languages
etc. Software like JAWS or windows-eyes is not able to do
this. These packages are 'dumb' compared to emacspeak because they
have know understanding of the text they are transforming into speech
- to them its just text. Emacspeak is able to be more 'intelligent'
because it is able to identify textual components and provide
additional audible clues about that text.

Tim
>         Dmitry
-- 
Tim Cross
mailto: tcross@xxxxxxxxxxx
phone: +61 2 6772 5973
mobile: 0412 969193
------------------------------------------------------------------------

-----------------------------------------------------------------------------
To unsubscribe from the emacspeak list or change your address on the
emacspeak list send mail to "emacspeak-request@xxxxxxxxxxx" with a
subject of "unsubscribe" or "help"