For the past several months, Code Factory Eloquence under NVDA spells out all the text on control elements like buttons or menus (including the nVDA main menu). I can't remember if it started after a Windows update or NVDA update, but it doesn't exhibit the behavior in Narrator. Anybody else run into this?
Comments
Eloquence for NVDA.
If you type that into google, you should find ways to get eloquence that won't have these issues.
IBM TTS on GitHub
You can google the repository for this, but GitHub has a publicly available freeware called IBM TTS. It is free without libraries, as those are licensed if I recall, but the main add-on is available and is esentially an Eloquence clone.
Why do I mention this? Because i use this with my Network engineer Cert course i am in, and do not have the spelling controls issue that you are experiencing.
HTH. 🙂
I didn't have this bug
with codefactory and vocalizer until nvda remote to sb machine and it spelling most controls at loud. Unfortunately this company is seemingly broke and/or bancrupt. With the grey area with eloquence and vocalizer legally and nvaccess unable to properly pay a lawyer and fight for their rights I very might use that on blind help as much as I hate the idea. Nobody is talking about the real danger of a lack of inovation in tts world aside ai and expressiveness which for day to day screen reading use we don't need / care about. .. hmm just my opinion
Eloquence for the win.
It kind of blows my mind that a speech synth I've been using since the 2000s still holds up today but it does.
I'd not be surprised at all if I keep using eloquence for the wrest of my life, something hugely and I mean hugely game changing would have to come along for me to change my mind.
Clarification
Just to be clear, I was not referring to what BlindHelp has to offer with my previous post. I was actually referring to the GitHub page where you can obtain a license free, copy of IBM TTS (again without libraries). GitHub is owned by the Microsoft Corporation, so I am skeptical that they would allow pirated software. 🤷
On a side note, Tiflotecnia now runs Vocalizer voices for NVDA. More info can be found here.
HTH. 😃
Thanks
@Lottie, you were very helpful, actually.
I *hate* trying to use the NVDA Google Group site, because I find it (and all Google products) hard to use: links often don't work until exiting browse mode, the page has inadequate structure, etc. It took me an hour searching the NVDA forum to find the relevant info, largely because the first page of results comprised people being dickish to people asking the question. They spent enormous amounts of energy chastising people for their ostensible lack of effort or, I guess, lack of skills, when they could have just answered the question. That's my #1 reason for being so grateful AppleVis added this site area, both because the site is more usable and the community so much more humane.
For anyone else still struggling with this, the answer is to uncheck "use spelling functionality" in the voice dialog (NVDA+Control+V, I think). I had turned off other spelling options without success, but guess I hadn't spotted this one. Also, there may be a new version of SAPI eloquence that fixes the issue. I haven't downloaded it.
I'm also going to look into the IBM TTS. FWIW, the algorithms that have become known as Eloquence date back to the DECTalk hardware synth, which I used in the mid-90s. I use Eloquence on Mac, too. It's the culmination of what are called formant speech synthesizers. Concatenative synthesizers like Alex and Windows OneCore aren't going to get any better. But I can see myself going with neural voices as the tech progresses. Neither my brain nor my blood pressure can handle super-fast speech anymore.
Irony
It is interesting, and frustrating, how some speech engines will do (what I would consider a basic function) different from one another. Example, Vocalizer voices choke in any kind of 'Terminal' interface, whether it's something basic like Command Prompt, or something as robust as Cisco's IOS. Not to be confused with Apple's iOS (note the spelling).
I have to use IBM TTS to do anything Command Line, or otherwise 'code' related, or else I will likely break something.
Badly. 😔