Hi Chip,
Sapi, or Microsof voice, has both rate and pitch to be between -10 and
+10 so, you would just have to locate 0 as the middle of the other voices
and go from there, Or just have Sapi as a different parameter.
But, find the middle of all the others and let Microsoft Sapi 0 value be
that middle, or an offset of the others middle...
Now, Sapi only has voice, rate, and volume as there parameters. Pitch is
an Xml command, so, it depends on how you spell out your class.
I gave you my class on Sapi and I only deal with the old, virus
filtered, Sapi 4 as the alternative in an if statement. But, it has all the
possibilities in that. class. No different actually from your MS Office word
class.
So, use my Sapi class and you will get the rest. Other voices don't even
use pitch, but I have not looked into how they use pitch to know...
Bruce
Sent: Thursday, July 14, 2011 8:09 PM
Subject: voice parameters
Hi all,
I've been working with an app which has different voice settings to convey
different information (optionally).
I don't want just to allow the user to edit all the voice parameters
(although I will eventually), but I'd like to come up with reasonably good
guesses as to what would make a different sounding voice for any number of
situations (so the user wouldn't have to do any setup work if they didn't
want to).
I've found varying the tone, which changes what we would call the voice, to
be the best way of achieving this (better than adjusting the pitch; it
yields voices which are more distinguishable from one another and still
understandable).
Unfortunately, I think you have no way of knowing what tone values you
assign will yield a change in the voice. For Eloquence, I think it's tone
values of "I" through "M", and for DEC-Talk Access32 it's "H" through "P"
(I think).
That unfortunately is all I know, and if the user is using some other
synthesizer, then my app will fail. So, I was wondering if people who use
other synthesizers would send me the range of tone settings for their
particular model of synthesizer which yield different voices?
I'd be glad eventually to publish this as a class so others can benefit.
Steve has already started me off with nice code for a class which makes it
easy to switch between various grouped values, and to store them as a voice
"specification".
Just an FYI, I believe you can't switch synthesizers on the fly, the way you
can adjust the tone setting, because you'll get simultaneous speech from
both, or the switch doesn't happen exactly when you need it to (something
like that), so adjusting the tone/pitch of your currently selected synth is
the only way I know to deal with this need in general. I'd be glad to hear
differently though.
thanks for any help.
Chip
