Re: Speech crackling/breaking up in latest NVDA


Speech output is one of the major components of NVDA. However, I would say that interpretation is a slightly higher priority, along with responsiveness in how it is carried out. Not that I don't value speech issues as something to be ignored - it is a major priority. I wrote the below statement, recognizing that Python Software Foundation itself no longer supports Python 2.7, which has huge implications even for NVDA at some point. Older NVDA releases will work fine, but we cannot ignore the speed on which we drive the highway.

-----Original Message-----
From: <> On Behalf Of Tyler Spivey
Sent: Thursday, August 6, 2020 2:38 PM
Subject: Re: [nvda] Speech crackling/breaking up in latest NVDA

It's perfectly fine to run 2019.2.1. It should keep working for a while.
If it works for you, then great. One thing you might have a problem with is finding copies of addons that work with it.
The core functionality of NVDA is to speak, so this seems pretty important to get right.
See the comment I made to #11061 this morning for an easily reproduceable test case.

On 8/6/2020 2:31 PM, Joseph Lee wrote:

It’s up to you, but at this time, I advise not going to anything
earlier than 2019.3.



*From:* <> *On Behalf Of
*Sharni-Lee Ward
*Sent:* Thursday, August 6, 2020 2:30 PM
*Subject:* Re: [nvda] Speech crackling/breaking up in latest NVDA

Would it be better for me to roll back my NVDA then? Or should I stick
this out and hope improvements come along eventually? I can assure you
I didn't have this problem before I updated, but I was concerned the
minor accessibility issues I was having might be due to my having an
outdated version of NVDA on my system.

On 7/08/2020 2:36 am, Joseph Lee wrote:


Part of these have to do with speech refactor. Python 3.7 did
contribute somewhat, but we’re basically talking about multiple
changes at once. At the moment NV Access folks are working hard on
improving performance and fixing outstanding Office usability bugs.



*From:* <>
<> <> *On Behalf Of *Sean
*Sent:* Thursday, August 6, 2020 9:31 AM
*To:* <>
*Subject:* Re: [nvda] Speech crackling/breaking up in latest NVDA

Unfortunately, the old stability disappeared with the move to Python3.

I have an Issue about this. But the NVDA team said this would be a
difficult / complex process.
Nothing can be done right now.


On 06/08/2020 07:49, Sharni-Lee Ward wrote:

So I just finally decided to brave the Python 3 hurdle, and
updated from 2019.2.1 to 2020.2. I've updated the addons I could
and wanted to keep, and they're both working fine.

I am, however, having an odd little issue with the synth's
speech. I'm using ESpeak NG, and the latest NVDA. I'm on Windows

My issue is that occasionally, most often at the ends of
sentences and on the word image in the context of that being an
object, the speech will break up slightly, for a brief but
noticeable fraction of a second. It happens irregularly, and I
don't know what to do to fix it. I asked on a Discord server I'm
a member of, and someone there told me to switch audio output
from Microsoft Sound Mapper to my speakers/headphones. I did
this, but it hasn't helped.

I've also noticed a slight but noticeable lag when deleting
characters, and in single-letter navigation. In the latter, It
takes most of a second before NVDA announces the item the cursor
has moved to after the letter was pressed.

Any help regarding these issues would be sincerely appreciated.



* Email:
* GitHub: SeanTolstoyevski

👨‍🦯I’m programmer. I coding often Python, sometimes Go and rarely

Join to automatically receive all group messages.