Speechd-el: Word Echo, Smarter Navigation, Visual Performance?

[Milan Zamazal]

>>>>> "VT" == Veli-Pekka T?til? <vtatila at gmail.com> writes:

    VT> 1, lack of word echo: There's no word echo in speechd-el for
    VT> immediate feedback when typing, That's just majorly bad for me,
    VT> since I've edited with word echo about 10 years in quite a
    VT> number of apps. Even with the fastest Espeak setting, char echo
    VT> is too slow for me and it is hard to mentally parse the
    VT> individual letters to words. I can pick out most ENglish and
    VT> Finnish typos by pronunciation when word echoed, but character
    VT> echo doesn't make that process quite as straight forward and
    VT> transparent.

Well, this feature is easy to add.  So I've implemented it and it's in
speechd-el CVS now.  It can be enabled by setting new speechd-speak-echo
variable.

    VT> SO ideally I'd like to tell Emacs, move to the next sentence,
    VT> and have speechd-el read me the whole of the next sentence, as
    VT> that's what my movement implied. It does not do that, reading
    VT> the destination line after the navigation in stead.

speechd-el already works for me this way.  For example, when I press
`M-e' Emacs moves to the next sequence and reads it.  The same applies
to paragraphs.  Did you change any speechd-el basic reading settings?

speechd-el doesn't provide any facility for chapter reading, because
there is no general concept of "chapters" in Emacs.  But if any
extension command defines chapter movement commands, it should be easy
to add a custom speechd-el definition making the commands behave in a
similar way as sentence or paragraph movements.  Tell me if you want to
know details.

    VT> Alternatively, on the reader side, you could just read whatever
    VT> is between the current and previous cursor position when the
    VT> cursor moved, which would amount to much the same thing without
    VT> the reader having to understand app logic other than getting the
    VT> whole doc contents. There are some problematic boundary cases
    VT> such as first and last unit, but still I think the genral idea
    VT> might be useful and so far unimplemented.

This may not be that easy.  Cursor can move for many reasons, even
within interactive commands, and I'm not sure the user would be always
happy with such a behavior.  But it may be possible to try it as an
option.

    VT> One nice addition would be a command, stop the speech and drop
    VT> the cursor to the pos the reader was reading at the moment. It
    VT> would be pretty neat in hands-free reading a whole chapter,
    VT> wishing to pause in the middle for slight editing. Later on, you
    VT> could then continue from where you left off by using the next
    VT> section command again.

Yes, this would be nice.  It's possible to implement this feature as
Speech Dispatcher already contains support for index marks.  But it's
not trivial and I'm afraid I won't be able to implement this feature
anytime soon.  OTOH it shouldn't be very difficult too, so if anybody
would like to help with it, tell me.

    VT> 3. Performance on older machines: 

Do you speak about your Pentium laptop with speechd-el Ubuntu package?
I guess speechd-el performance should be acceptable on it.  But I don't
know, I don't have experience with this nor do I have any old hardware
around to test it.  Perhaps Orca makes the machine busy enough to reduce
the performance below the acceptable level?

Regards,

Milan Zamazal