[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Search]

Re: emacspeak + SAPI4.0 on win32



Yes, I am the same person who started on this back in December. The crush 
of the academic schedule snuffed out any chance for progress until now.

Yes, I have received the Dec-talk manual and much is clearer.

I agree that recognizing and translating the codes is bound to lead to 
headaches. For my immediate benefit I hacked the function dtk-define-voice 
so that instead of storing the dectalk string it is given as an argument, 
it instead ignores the string and installs the name of the symbol as the 
definition. Now I get embedded strings like [Paul} and [Paul-animated] in 
the text to the speech interface. A dirty hack but it's easier for me to 
make progress this way.

I agree with your strawman for the *right* way to proceed. I *may* go down 
that road. I first want to get the thing going by hackery so that I can get 
some experience with this kind of thing. I'm very interested in audio 
interfaces to computers and emacspeak is clearly the best one to date.

I will certainly release my program and source for anyone to use. I was 
thinking of code naming it "tick" since it is riding along on "wonder dog".

My program certainly isn't pretty. I know almost nothing about MSAPI and 
all the machinery on which it is built. I'm just looking at examples and 
the doc and hacking things together. I'm trying to figure out how to 
generate tones now. Beep looks like the right goods on NT but doesn't work 
on Win95. Sheesh... What an OS.....

My main lack now is any idea of how the various voices sound. I guess I'll 
make some fairly arbitrary choices illuminated by the names of the voices 
and later try to match them.

gb


At 08:46 AM 7/12/99 -0700, T. V. Raman wrote:
> >>>>> "Gary" == Gary Bishop <gb@xxxxxxxxxxx> writes:
>
>     Gary> I've got emacspeak10.0 limping along on NT4.0 and
>     Gary> Win98 using Emacs 20.3.1 and the Microsoft Speech
>Wow --this is great to hear --congratulations!!
>
>by   the way, aren't you the same person who wrote to me saying
>you would be doing this sometime towards the start of the year?
>     Gary> SDK 4.0. I've never used or heard emacspeak before
>     Gary> so I have no idea how what I am hearing compares
>     Gary> to the *real* thing.
>
>     Gary> I have 2 questions:
>
>     Gary> 1) If I configure emacspeak10.0 as shipped and try
>     Gary> to run it, emacs complains that
>     Gary> emacspeak-resource-directory is void. I see the
>     Gary> defvar in emacspeak-setup.el but the complaint
>     Gary> happens down in the load of emacspeak-pronounce.el
>     Gary> that happens as a result of loading emacspeak.el
>     Gary> several lines up in emacspeak-setup.el. I moved
>     Gary> the defvar up in emacspeak-setup.el and that
>     Gary> appears to fix it. Is this some subtle bug in
>     Gary> NTemacs or did I do something else wrong?
>
>Sounds like a subtle ntemacs quirk.
>However I do have emacspeak 10.0 loading and running without
>complaints on emacs 20.3.1 (I think --need to check the .1)
>under NT.
>I'll look at this again, and go ahead and move the line up
>if necessary as you found.
>
>     Gary> 2) Is there a document that interprets all the
>     Gary> dectalk commands? I'm trying to puzzle out the
>     Gary> meaning of the various sequences that get inserted
>     Gary> into text strings. I'm guessing these are commands
>     Gary> for dectalk. Perhaps they are tcl commands
>     Gary> instead? I'm not using tcl at all in my
>     Gary> implementation, emacspeak talks directly to my
>     Gary> little app which interfaces to the SAPI. I'm
>     Gary> trying to understand, for example, what was
>     Gary> supposed to happen as a result of this string:
>
>     Gary> pound [:np :dv pr 200 hr 30 sr 50 as 100 qu
>     Gary> 100]ifndef [:np ] S[*] T[*] [:np :dv pr 200 hr 30
>     Gary> sr 50 as 100 qu 100]R_H [:np ]
>
>Jim Van Zandt has already sent you the Dectalk manual --that
>should help.
>The [..] commands are all Dectalk control sequences.
>Some are standard IPA style notation [*] is a morpheme
>boundary for instance.
>
>Here is a strawmen sketch for the "right way" to handle this
>for MS SAPI and other non-dectalk engines:
>
>0) Emacspeak implements a dtk-css-speech module
>--see file dtk-css-speech.el
>--this is an implementation of the necessary structures and
>mappings for implementing aural CSS
>
>1) Many of the  modules in emacspeak that were written
>before dtk-css-speech got created
>use direct dtk codes in the voice definition --e.g. module
>dtk-voices.el
>--these should be updated to instead use dtk-css-speech
>based voices.
>do a grep '[:' *.el to find all the culprits
>2) dtk-css-speech itself provides a css->dectalk mapping
>--that module should be split into two modules
>a) css-speech.el
>b) dtk-css-speech.el
>
>4) Finally, for tts package xxx we can create
>xxx-css-speech.el
>
>The above has been on my todo list for over two years,  but
>I have never found the time, and am probably unlikely to do
>so in the near future.
>
>Depending on how much elisp hackery you want to sign up for,
>I'd love to see this happen--
>if you start it off I'll probably help as we go along.
>
>I would not recommend some kind of ad-hock kluge that
>inspects the codes emacspeak spits out and translate them
>over in the speech server code--
>it quickly turns into a debugging and maintainance
>nightmare.
>
>On  the choice and design of voice personalities
>the MS SAPI kit does have multiple voices,
>but they do not have as many parameters to tweak per voice.
>
>Finally, I assume you're planning to release this thing once
>built as open source --if not please let me know what your
>plans are.
>     Gary> currently it reads all the bracketed text to me,
>     Gary> which clearly isn't what is desired.
>
>     Gary> Thanks for any help.  gb
>
>     Gary> 
> -----------------------------------------------------------------------------
>     Gary> To unsubscribe or change your address send mail to
>     Gary> "emacspeak-request@xxxxxxxxxxx" with a subject
>     Gary> of "unsubscribe" or "help"
>
>--
>Best Regards,
>--raman
>
>       Adobe Systems                 Tel: 1 408 536 3945   (W14-128)
>       Advanced Technology Group     Fax: 1 408 537 4042
>       W14-128 345 Park Avenue     Email: raman@xxxxxxxxxxx
>       San Jose , CA 95110 -2704     Email:  raman@xxxxxxxxxxx
>       http://labrador.corp.adobe.com/~raman/        (Adobe Intranet)
>       http://cs.cornell.edu/home/raman/    (Cornell)
>----------------------------------------------------------------------
>     Disclaimer: The opinions expressed are my own and in no way should be 
> taken
>as representative of my employer, Adobe Systems Inc.
>____________________________________________________________
>
>-----------------------------------------------------------------------------
>        To unsubscribe or change your address send mail to
>"emacspeak-request@xxxxxxxxxxx" with a subject of "unsubscribe" or "help"

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


Emacspeak Files | Subscribe | Unsubscribe