Comment 1 by jteh on 2013-07-07 22:38
We can't do anything about this. Those voices are slow, as natural voices often are.
Changes:
Added labels: cantfix
State: closed

Comment 2 by Animesh.mr on 2013-07-08 05:45
ok but it is working well with speech hub i am using it

Comment 3 by jteh on 2013-07-08 05:55
Please elaborate. Are you saying you are using Microsoft Speech Platform via Speech Hub? In what way is it faster with Speech Hub? Are you talking about the time between pressing a key press and hearing a response, the time taken to silence speech or something else?
Changes:
Removed labels: cantfix
State: reopened

Comment 4 by jteh on 2013-07-08 05:58
Also, try to describe the time difference; e.g. it takes half a second to respond with NVDA's synth driver but only a quarter of a second with Speech Hub, etc.

Comment 5 by Animesh.mr on 2013-07-08 11:11
it is fine with speech hub and wile typing it is as fast as espeak but some times take lintel time to read same as neo speach jooli would take

Comment 6 by briang1 on 2013-07-08 16:51
I have not noticed speech hub being any faster than native, though the actual intonation is sometime different on the same text. I don't know how Speech hub presents the data to it.
It is quite sluggish but not bad if you want to listen to a document. I still find it annoyingly slow for navigation though, even on a fast machine, as I suspect it was originally developed for its clarity for use on telephone aanswering systems etc, rather than for the likes of us.
I seriously doubt much could be done in nvda. Maybe you need to ask the speech Hub folk about what they are doing if you feel convinced it is in fact faster.

CC @jcsteh

@jcsteh do you believe this is still something that NVDA can not fix? Or has a fix for sapi voices already been added?

The specific nature of the performance problem here (as compared to other software using the same synthesiser) was never clarified. #8174 does implement a fix which allows certain Microsoft voices to be interrupted faster. That may or may not be relevant here; I'm not sure. Either way, in order for this to be actionable, this needs to be re-tested and the exact nature of the issues clarified.

@Animesh.mr could you please describe a bit more what exactly the issue is? Is the speech reacting slow when you navigate with arrow keys? Or is the interuption between sentences and words too Long compared with other Synthesizers?
According to my last tests on Windows 10 with NVDA 2019.1 Beta 1, Microsoft speech platform works quite good in Terms of Performance.

I did a simple test:
In nvda I am using the voice of the one core synthesizer, I set the voice speed to 100%.
Now I logged into the narrator, went through speech settings and changed the voice speed to 100%.
The speed reached by the narrator is enormous, much greater than the speed achieved by the nvda, ambush the readers at 100%.
I expect nvda when set to 100% using the one core synthesizer will achieve much higher speeds, similar to that of the narrator at 100%.
It is a question of speed of the voice, the problem of pauses when finishing sentences is much better after the corrections implemented in the nvda.

I tested with the old speech engine on Windows 10 and the performance is quite good. @fernando-jose-silva the speed of the voice is covered in a different issue.

I am afraid the initial author is not available anymore and we are not exactly sure what the issue actually is. So if someone can reproduce an issue similar to this, please open a new issue. I close this one for now.