Re: Reading of numbers


Robert Mendoza
 

Hi, Quentin!

In relation to my earlier post, this happened when reviewing the said numbers not by doing typing characters. This was tested to different speech synthesis which are the eSpeak ng and Microsoft SAPI5.

Robert Mendoza
On 4/12/2017 4:23 PM, Robert Mendoza wrote:

Hi, Quentin!

Sorry for the late response. I would like to add to my note that after doing a test to the speech synthesis for the Microsoft version 5, I got the same result of reading or having output. If there is a chance do you have means to mitigate this to fix this in the speech dictionaries so that nvda could read it precisely and correct this occurrences.  Again, the speech mentioned for the version 5 is the one that is built to the version 2017.1.

Robert Mendoza
On 4/12/2017 1:52 PM, Robert Mendoza wrote:

Hi, Quentin!


The test normally occurs when using in the eSpeak ng synthesizer which I usually used with with my test, but never had a chance to play it to Microsoft speech version 5. Ever since I never had a chance either the speech settings is change as well.

Robert Mendoza
On 4/12/2017 1:29 PM, Quentin Christensen wrote:
Hi Robert,

Can you please clarify what behaviour you are getting vs what you expect exactly?

When I write 1.49.1 in Word or NotePad, exactly what is read depends on the synthesizer I am using:

Using eSpeak NG, NVDA reads it as "one point four nine point one"
Using Microsoft Speech API Version 5, NVDA reads "one dot forty-nine dot one"

The exact text read will depend on the synthesizer, unless you have created an entry in the speech dictionary which covers the text encountered.

I did try just 1.49 and both synthesizers read it as "one point four nine".  At a guess, both synthesizers see this is a decimal number and read the numbers after the decimal point individually (since you wouldn't usually say "one point forty-nine".  There isn't as clear a rule on numbers with multiple dots / decimal places.  For instance, I've heard "255" in the middle of an IP address pronounced as both "two five five" and "two hundred and fifty-five".

Regards

Quentin.

On Tue, Apr 11, 2017 at 11:15 PM, Robert Mendoza <lowvisiontek@...> wrote:
While testing earlier, just noticed that nvda reports the set of number 1.49.1
Instead of saying this proper by, 1 point 49 point 1

Normally, this is happening when the first one number occurs. Or, I may say the one's place value of digit after the period, as stated above mentioned sample.  Is there a way to correct it to the speech settings to make it more precise of readings numbers. I appreciate your advice. or, perhaps someone could replicate this case. Otherwise, this is a sort of a fix that I would needed to one users or primarily errors in nvda. Again, I'm using the latest version of release of 2017.1 to my windows 10 Pro machine.


--
Robert Mendoza







--
Quentin Christensen
Training and Support Manager

Basic Training for NVDA & Microsoft Word with NVDA E-Books now available: http://www.nvaccess.org/shop/

www.nvaccess.org 
Facebook: http://www.facebook.com/NVAccess 
Twitter: @NVAccess 




Join nvda@nvda.groups.io to automatically receive all group messages.