speaking personally?
I am confused that this feature even needs requesting.
Every single DOS screen reader I own has a setting where the user decides
how numbers are spoken. single digits or words.
Further speaking personally I feel separating the idea, text to speech,
instead of drivers written to enhance synthesizer quality abilities leads
to this problem too.
Pronunciation's in text to speech only tools is a fine example.
Present and present are announced based on context, dessert and desert
those sorts of things, largely because my screen reader programs draw about
the capabilities of the synthesizer source I am using.
What frustrates me about muddying how such technology works is the way
outsiders respond when a situation arises.
Speakup is less than a quality screen reader if it does not easily provide
a way for an individual to set rate, pitch, inflection, number
pronunciation's, symbols anagrams and so forth.
I recently had someone from my senator's office tell met hat an embedded
form was supposed to work with screen readers..as if all of them are the
same work the same, and interact the same regardless of platform. Never
mind that the embedded nature meant the form would not appear..which I
demonstrated.
If Linux in general is going to be taken seriously, speaking personally,
and Speakup in particular is to be considered a solid tool?
these features need to be present.
text to speech is just what it says a spoken rendition of a plain text
file, which is why the pronunciation can be so dreadful. It takes a
quality screen reader program to provide all the extras, mixing the
terms only leads to poor results.
Best,
Karen
On Sat, 1 Jun 2024, Chime Hart wrote:
Hi Martin: While I basicly agree with your analysis, I would say Vocal-Eyes
in DOS had it completely right. Your comments about NOAA Weather Radio, it
actually sounded better when they were useing DecTalk, however sales of
Radios had dropped. This particular AT&T Natural Voice has odd seggment jumps
or even hicups.
Chime
