This section taken from the speech-dispatcher manual may help.
ModuleDelimiters "delimiters", ModuleMaxChunkLength length =
Normally, the output module doesn't try to synthesize all incoming text at
once, but instead it cuts it into smaller chunks (sentences, parts of
sentences)
and then synthesizes them one by one. This second approach, used by some
output modules, is much faster, however it limits the ability of the output
module
to provide good intonation.
NOTE: The Festival module does not use ModuleDelimiters and
ModuleMaxChunkLength.
For this reason, you can configure at which characters (delimiters) the text
should be cut into smaller blocks or after how many characters (length) it
should be cut, if there is no delimiter found.
Making the two rules more strict, you will get better speed but give away
some quality of intonation. So for example for slower computers, we
recommend
to include comma (,) in delimiters so that sentences are cut into phrases,
while for faster computers, it's preferable not to include comma and
synthesize
the whole compound sentence.
The same applies to MaxChunkLength, it's better to set higher values for
faster computers.
For example, curently the default for Flite is
FestivalMaxChunkLength 500
FestivalDelimiters ".?!;"
The output module may also decide to cut sentences on delimiters only if
they are followed by a space. This way for example "file123.tmp" would not
be cut
in two parts, but "The horse raced around the fence, that was lately painted
green, fell." would be. (This is an interesting sentence, by the way.)
----- Original Message -----
From: "William F. Acker WB2FLW +1-303-722-7209" <wacker@xxxxxxxxxxxxx>
To: "SPEAKUP Distribution List" <speakup at speech.braille.uwo.ca>
Sent: Tuesday, February 19, 2008 1:19 AM
Subject: Trying out Speech Dispatcher.
> Hi all,
>
> I finally got around to taking a look at Speech Dispatcher. I ran
> into two problems so far. Using either ibmtts or espeak, SD would often
> split words while reading. Configuration would sometimes come out as con
> figuration, and so on. Worse yet, it appears that SD doesn't support
> pitch changes for uppercase letters. When I was finished playing with the
> new configuration, I switched back to Marc Mulcahy's spk-connect-viavoice.
> To my surprise and lack of amusement, I no longer had a pitch change on
> caps. I tried everything I could think of, even resetting
> /proc/speakup/synnth_name from sftsyn to none and back again. Nothing
> helped short of a reboot.
>
> Any ideas?
>
>
> --
> Bill in Denver
>
> _______________________________________________
> Speakup mailing list
> Speakup at braille.uwo.ca
> http://speech.braille.uwo.ca/mailman/listinfo/speakup
