Re: I take it there have been no improvements to the flawed espeak ng in 2017.4?


Sharni-Lee Ward
 

Oh, yeah, I forgot the American Espeak is godawful. I always use the default English voice, as I am Australian and it is the closest to my own accent I can get, while also not being painful to listen to. I tried the American accent a long time ago on Espeak and couldn't stand it. I don't know who's American or not on here as I only read message chains that are relevant to me.


On 27/03/2018 8:09 AM, Devin Prater wrote:
I agree completely, no matter how much users mess with voice files, eSpeak will always be a British synth that will be discomfort for new screen reader and NVDA users. Now, we have the One-core voices, but in Windows 1803 and the latest version of NVDA, they sound like chipmunks. So yeah, I'll just have to hope NVDA gets an update to fix that, or Microsoft fixes it. .




On Mon, Mar 26, 2018 at 4:01 PM -0500, "Gene" <gsasner@...> wrote:

The add on you were sent is the illegal add on which uses the same inferior implementation as JAWS does.  I don't use it but I've heard it.
 
Gene
----- Original Message -----
Sent: Monday, March 26, 2018 3:18 PM
Subject: Re: [nvda] I take it there have been no improvements to the flawed espeak ng in 2017.4?

I'll admit I've not heard Eloquence outside of Jaws and the addon I was sent some years back. I didn't know the version in Jaws was different to Eloquence on its own. The one in the addon I have doesn't pause very long between sentences. Also, I just do not like to speed up my synth too much. I still think of how some of you can listen to something at the highest rate possible and understand it. I've tried increasing it a little, but it's just not enjoyable for me.


That said, I'm now very used to Espeak, and it may not be an improvement over Eloquence in the opinion of some, but it was definitely an improvement over Microsoft Anna, which was what I had been stuck with via Narrator until I discovered NVDA. Espeak is the voice of NVDA to me now, and other synths fall short in many ways.


On 26/03/2018 7:20 PM, Gene wrote:
You wrote:
I don't like Eloquence because it isn't as crisp as Espeak. I have compared
it to A.M radio in the past, and Espeak to FM/CD Quality sound.
 
That is a very inaccurate comparison.  E-Speak has more high end in the s sounds and the voice may have a little more high end but hardly comparable to FM radio or CD quality sound.  and why is the difference of perhaps two thousand or three thousand KHZ, by my estimate, from, let's say, around 7000KHZ in Eloquence to about nine hundred or one thousand KHZ in E-Speak of any real importance?  I would imagine you talk on the phone without worrying about such things. 
 
Eloquence has much better understandability of speech at higher speeds, and pronounces a lot more words correctly. 
 
Also, I don't know if you are talking about Eloquence as modified by the producers of JAWS or Eloquence, as it is unmodified.  The JAWS designers customized it.  If you are saying it isn't crisp because you are thinking of JAWS Eloquence, that's not an accurate representation of what unmodified Eloquence sounds like. Also, the JAWS modifiers took out pauses at periods.  That seriously degrades reading, that the modified version doesn't pause at periods. 
 
You may be pleasantly surprised if you try the demo of the HTML version.  If you haven't heard unmodified Eloquence, you don't know what it really sounds like. 
 
Gene


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