Responsiveness and ViaVoice Outloud

["T. V. Raman" <ramantv@xxxxxxxxxxx>]

You're the second to raise this --
Bart Bunting  posted a similar message to the viavoice dev
list.

I've got emacspeak working with viavoice outloud at
present-- and stop is definitely not a problem --it stops as
quickly as you would expect.

I'd make the same suggestions to you as I did to Bart--

a) make sure you've got the engine in a mode where it
synthesizes as text is added via the ECIAddText call;

b) Make sure in your tests that you are queueing complete
clauses in the AddTexT call 

At this point the only place I see some performance
degradation is when speaking individual characters 
--with immediate flushing of speech turned on in emacspeak
as you type,
the engine keeps pace to the extent that you can hear about
every alternate letter with character echo turned on.

I did a quick test on Notepad udner JFW on win98 and see
similar behavior.

--Raman

>>>>> "Bryan" == Bryan Smart <bsmart@xxxxxxxxxxx> writes:

    Bryan> Hello everyone.  I've been working on a speech
    Bryan> server for Emacspeak to interface it with the
    Bryan> version of ViaVoice Outloud for I386 Linux.  I
    Bryan> have had a fair amount of progress, but am
    Bryan> disappointed with the responsiveness of the
    Bryan> ViaVoice Outloud program.  This lack of
    Bryan> performance is especially lacking in the areas of
    Bryan> the time it takes ViaVoice Outloud to respond to
    Bryan> a stop speech request, as well as the length of
    Bryan> the initial delay required to synthesize single
    Bryan> letters (making arrowing in files a difficult
    Bryan> process).  It is an improvement over Festival and
    Bryan> Mbrola, but the Linux version of ViaVoice Outloud
    Bryan> does not match up to performance of similar
    Bryan> versions of the software for Windows.

    Bryan> As I recall, someone on this list commented that
    Bryan> Eloquence, the incarnation of ViaVoice Outloud
    Bryan> distributed by Henter-Joyce, had been extensively
    Bryan> modified in order to produce superior
    Bryan> responsiveness.  This may or may not be so, but
    Bryan> the version that was included with my copy of JFW
    Bryan> certainly outperforms the Linux port of ViaVoice
    Bryan> Outloud.  Can anyone speculate as to the cause of
    Bryan> the differences in performance?

    Bryan> I doubted the truth of the assertion that H-J had
    Bryan> "modified" ViaVoice Outloud, as I assumed that
    Bryan> IBM would license a product to Henter-Joyce, but
    Bryan> would not give them the source and allow them to
    Bryan> distribute a hacked version of the product.

    Bryan> Also, there is the matter that the Linux version
    Bryan> of ViaVoice Outloud is a beta version, but,
    Bryan> considering that most of the basic TTS engine
    Bryan> would have survived the port in a mostly
    Bryan> unchanged state, I was hesitant to blame the lack
    Bryan> of speed on the program's beta status.

    Bryan> Basically, I'm ready to listen to any technical
    Bryan> tips, rambling, rumors, or any other information
    Bryan> that people are willing to give about ViaVoice
    Bryan> Outloud, any "hacks" by the screen reader
    Bryan> vendors, and the Linux port.

    Bryan> Best, Bryan

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

-- 
Best Regards,
--raman

      
Email:  raman@xxxxxxxxxxx
WWW:    http://cs.cornell.edu/home/raman/             
PGP:    http://cs.cornell.edu/home/raman/raman.asc 

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