nuance cerence neural tts for NVDA?
|
Josh Kennedy
Hi, Does anyone know if atguys or some other company will be selling the new Cerence vocalizer deep AI voices for NVDA? Nextup sells them for their text aloud product and they sound really good. Voices like conversational Zoey sound quite good.
Josh
Sent from Mail for Windows 10
|
|
|
|
Gene
There is nothing wonderful about those voices.
toggle quoted message
Show quoted text
Zoey sounds as though she's gargling a little and has stilted inflection. She sounds to me as though she is on downers. The other voice I sampled has stilted inflection and artifacts as well. They may be better than the Nuance voices that came before, but until these newer synthesizers sound as good as the Google Assistant or Alexa, or other such very high quality speech found in digital assistants, I'll consider them not worth using. Gene -----Original Message-----
From: Josh Kennedy Sent: Friday, September 11, 2020 8:39 AM To: nvda@groups.io Subject: [nvda] nuance cerence neural tts for NVDA? Hi, Does anyone know if atguys or some other company will be selling the new Cerence vocalizer deep AI voices for NVDA? Nextup sells them for their text aloud product and they sound really good. Voices like conversational Zoey sound quite good. Josh Sent from Mail for Windows 10 |
|
|
|
Jesse Farquharson
That might be your opinion, and that's perfectly fine. But I'm sure there are others who would welcome the additions. The thing is, most of the Nuance voices NVDA uses haven't been refreshed in quite some time. In fact, since NVDA 2019.3, we have not gained voices, we have lost the ability to use many of them, because they have not been ported to Python 3. Ie DecTalk, SoftVoice, RH Voice and so on and so on. As crazy as it sounds, there are some, such as myself, who still want to use them. They may be old, but they do have a certain... charm. The loss of those voices is really what kept me from updating for quite some time. And I'm still a bit sore about it. I wish a greater effort had been taken to bring some of the classics along. |
|
|
|
CARLOS-ESTEBAN <carlosestebanpianista@...>
Hi all. Not is easy make a driver for this voices and NVDA. The voices based in cloud services depending of the internet connection, and a screen reader send many text. Is necessary search some solutions first. Regards.
Carlos Esteban Martínez Macías. Músico (pianista) y también ayuda a usuarios ciegos y con discapacidad visual con el uso de lectores de pantalla y tecnología. Experto certificado en el lector de pantalla NVDA.
Musician (pianist) and help to the blind people, with use of screen readers and technology. Certified expert in the screen reader NVDA.
NVDA certified expert certificate
De: Gene
There is nothing wonderful about those voices. Zoey sounds as though she's gargling a little and has stilted inflection. She sounds to me as though she is on downers. The other voice I sampled has stilted inflection and artifacts as well. They may be better than the Nuance voices that came before, but until these newer synthesizers sound as good as the Google Assistant or Alexa, or other such very high quality speech found in digital assistants, I'll consider them not worth using.
Gene -----Original Message----- From: Josh Kennedy Sent: Friday, September 11, 2020 8:39 AM To: nvda@groups.io Subject: [nvda] nuance cerence neural tts for NVDA?
Hi,
Does anyone know if atguys or some other company will be selling the new Cerence vocalizer deep AI voices for NVDA? Nextup sells them for their text aloud product and they sound really good. Voices like conversational Zoey sound quite good.
Josh
Sent from Mail for Windows 10
|
|
|
|
Ron Canazzi
You wrote:
toggle quoted message
Show quoted text
"There is nothing wonderful about those voices. Zoey sounds as though she's gargling a little and has stilted inflection. She sounds to me as though she is on ..." which only goes to show that beauty or in this case audio clarity is in the eyes/ears of the beholder. I have seen this kind of disagreement about how wonderful or terrible voices/synthesizers sound for years and I have come to the conclusion that every ear hears things differently. With my hearing issues, I think that the Eloquence synthesizer with the male American English voices of Glenn and Reed are superior to most out there. There are those who have voiced vehement disagreement with this analysis: and so it goes. On 9/11/2020 10:23 AM, Gene wrote:
There is nothing wonderful about those voices. --
They Ask Me If I'm Happy; I say Yes. They ask: "How Happy are You?" I Say: "I'm as happy as a stow away chimpanzee on a banana boat!" |
|
|
|
Amir
Agreed 100 percent -- as subjective as it might be. At least compared with what we hear from cloud-based Google/Siri/IBM AI voices, the newly packaged Nuance Vocalizer voices sound vapid and the US English ones have intonation issues.
Best, Amir |
|
|
|
Leslie <soundsofmusic@...>
toggle quoted message
Show quoted text
From: Gene
Sent: Friday, September 11, 2020 7:23 AM To: nvda@nvda.groups.io Subject: Re: [nvda] nuance cerence neural tts for NVDA?
There is nothing wonderful about those voices. Zoey sounds as though she's gargling a little and has stilted inflection. She sounds to me as though she is on downers. The other voice I sampled has stilted inflection and artifacts as well. They may be better than the Nuance voices that came before, but until these newer synthesizers sound as good as the Google Assistant or Alexa, or other such very high quality speech found in digital assistants, I'll consider them not worth using.
Gene -----Original Message----- From: Josh Kennedy Where do you find them and how much are they? I loved the nuance voices that I heard before. I know you don’t like them, Jean, but I’d like to hear samples. I’m one of those people who like to change voices every once in a while. I have many voices on my phone. They add variety as I work.Sent: Friday, September 11, 2020 8:39 AM To: nvda@groups.io Subject: [nvda] nuance cerence neural tts for NVDA?
Hi,
Does anyone know if atguys or some other company will be selling the new Cerence vocalizer deep AI voices for NVDA? Nextup sells them for their text aloud product and they sound really good. Voices like conversational Zoey sound quite good.
Josh
Sent from Mail for Windows 10
|
|
|
|
Devin Prater
I kinda think these specialty TTS companies are losing steam. I don't think they can keep up with Apple, Google, or even Amazon in this. Devin Prater sent from Gmail. On Fri, Sep 11, 2020 at 9:57 AM Leslie <soundsofmusic@...> wrote:
|
|
|
|
Afik Souffir <afik.souffir@...>
Hi,
How can you download this voice?
From: nvda@nvda.groups.io <nvda@nvda.groups.io> On Behalf Of
Devin Prater
Sent: Friday, September 11, 2020 6:29 PM To: nvda@nvda.groups.io Subject: Re: [nvda] nuance cerence neural tts for NVDA?
I kinda think these specialty TTS companies are losing steam. I don't think they can keep up with Apple, Google, or even Amazon in this. Devin Prater sent from Gmail.
On Fri, Sep 11, 2020 at 9:57 AM Leslie <soundsofmusic@...> wrote:
|
|
|
|
Amir
Nicely put. Granted, Apple relies on them, Nuance in this case, for VoiceOver, but its own Siri Female voice is head and shoulders above what Vocalizer provides. Of course, I mean the Siri voice which VoiceOver can't use as there's a difference between the Female Siri voice for VoiceOver and the Female Siri voice which utilizes deep learning algorithms. Moreover, the best performers are also gobbled up by larger companies (like what happened to Ivona with Amazon). The rather worrying point, at least for screen reader users is that though Google/Microsoft/IBM have achieved a lot in terms of bringing oomph and naturalness to their TTS engines via deep learning, they can't be used offline due to various limitations and can't be used online due to the delay they generate for screen readers.
Best, Amir |
|
|
|
Jesse Farquharson
I personally don't think I'd mind the delay, honestly. If it meant access to new voices I've not heard before which could make reading documents and such more pleasant, I think I'd gladly take it. |
|
|
|
Amir
But it might be a serious concern -- guess it is indeed. People are complaining about 70-millisecond delays with SAPI 5 voices on NVDA's Github issues page, and these deep-learning voices can generate substantially longer pauses given their online operation. If you throw some paragraphs at them to read longer passages, the delay might be passable, but for navigation, the essence of screen reader use, that will become a headache. What's more, they are not free. You can use them up to a certain point for free, but all of them charge users per a specified number of characters synthesized after that. And the screen reader usage model requires lots of synthesis -- even for typing stuff.
Best, Amir |
|
|
|
Jesse Farquharson
I understand the delays aspect of things, for sure. Which is why I wouldn't use it for navigation. If NVAccess were to post a disclaimer somewhere, I'm sure that might help things. Or, different voices could be specified for different use cases, sort of akin to how... dare I say it, JAWS does it. In terms of the cost, for sure, that could be a definite barrier. However, I'm sure something could be worked out in that regard. |
|
|
|
Amir
OK. Let's see what happens in this regard in the near future. I also like IBM TTS (Watson) and the default Google Neural voice.
Best, Amir |
|
|
|
Gene
Yes, that is my opinion. I haven't followed older voices but I understand why those who like them and use them don't want to lose them.
toggle quoted message
Show quoted text
As I think about it, one reason I expressed myself as I did is because the voices are claimed to be real improvements and I don't think they are significantly better. Those who regularly use Nuance voices may differ because they are so much more familiar with the voices now widely used as opposed to these. I'll be curious how regular users react to the samples on the page. Gene -----Original Message-----
From: Jesse Farquharson Sent: Friday, September 11, 2020 9:31 AM To: nvda@nvda.groups.io Subject: Re: [nvda] nuance cerence neural tts for NVDA? That might be your opinion, and that's perfectly fine. But I'm sure there are others who would welcome the additions. The thing is, most of the Nuance voices NVDA uses haven't been refreshed in quite some time. In fact, since NVDA 2019.3, we have not gained voices, we have lost the ability to use many of them, because they have not been ported to Python 3. Ie DecTalk, SoftVoice, RH Voice and so on and so on. As crazy as it sounds, there are some, such as myself, who still want to use them. They may be old, but they do have a certain... charm. The loss of those voices is really what kept me from updating for quite some time. And I'm still a bit sore about it. I wish a greater effort had been taken to bring some of the classics along. |
|
|
|
Gene
There are many without hearing loss who have the same opinion. I'll be curious how people react to the voice samples.
toggle quoted message
Show quoted text
Gene -----Original Message-----
From: Ron Canazzi Sent: Friday, September 11, 2020 9:35 AM To: nvda@nvda.groups.io Subject: Re: [nvda] nuance cerence neural tts for NVDA? You wrote: "There is nothing wonderful about those voices. Zoey sounds as though she's gargling a little and has stilted inflection. She sounds to me as though she is on ..." which only goes to show that beauty or in this case audio clarity is in the eyes/ears of the beholder. I have seen this kind of disagreement about how wonderful or terrible voices/synthesizers sound for years and I have come to the conclusion that every ear hears things differently. With my hearing issues, I think that the Eloquence synthesizer with the male American English voices of Glenn and Reed are superior to most out there. There are those who have voiced vehement disagreement with this analysis: and so it goes. On 9/11/2020 10:23 AM, Gene wrote: There is nothing wonderful about those voices.-- They Ask Me If I'm Happy; I say Yes. They ask: "How Happy are You?" I Say: "I'm as happy as a stow away chimpanzee on a banana boat!" |
|
|
|
Gene
The web page where you can read about the voices and find samples is:
toggle quoted message
Show quoted text
https://nextup.com/cerence/ Gene -----Original Message-----
From: Leslie Sent: Friday, September 11, 2020 9:57 AM To: nvda@nvda.groups.io Subject: Re: [nvda] nuance cerence neural tts for NVDA? Sent from Mail for Windows 10 From: Gene Sent: Friday, September 11, 2020 7:23 AM To: nvda@nvda.groups.io Subject: Re: [nvda] nuance cerence neural tts for NVDA? There is nothing wonderful about those voices. Zoey sounds as though she's gargling a little and has stilted inflection. She sounds to me as though she is on downers. The other voice I sampled has stilted inflection and artifacts as well. They may be better than the Nuance voices that came before, but until these newer synthesizers sound as good as the Google Assistant or Alexa, or other such very high quality speech found in digital assistants, I'll consider them not worth using. Gene -----Original Message----- From: Josh Kennedy Where do you find them and how much are they? I loved the nuance voices that I heard before. I know you don’t like them, Jean, but I’d like to hear samples. I’m one of those people who like to change voices every once in a while. I have many voices on my phone. They add variety as I work.Sent: Friday, September 11, 2020 8:39 AM To: nvda@groups.io Subject: [nvda] nuance cerence neural tts for NVDA? Hi, Does anyone know if atguys or some other company will be selling the new Cerence vocalizer deep AI voices for NVDA? Nextup sells them for their text aloud product and they sound really good. Voices like conversational Zoey sound quite good. Josh Sent from Mail for Windows 10 |
|
|
|
Mary Otten
Clearly, speech preference is in the ear of the beholder. But I would not pay for any of these voices. I love the idea, and I'd pay for something that made the book-reading experience or even long article reading great. But I don't see where these do any better than other voices I already have access to. And why the heck do they insist on making so few male voices anyway? All 3 U.S. English ones are female. Nuts!
toggle quoted message
Show quoted text
Mary On 9/11/2020 10:14 AM, Gene wrote:
The web page where you can read about the voices and find samples is: |
|
|
|
Louise Pfau
Hi. I want to clarify that synthesizers like "Vocalizer" and "Eloquence" are paid products when used in conjunction with NVDA. I've got the "Vocalizer Expressive" voices that came as a series of free downloads for JAWS 16, and I really like them. I don't think they're listed as available synthesizers for NVDA, even though they're installed on my computer.
Thanks, Louise |
|
|
|
valiant8086
Hi.
toggle quoted message
Show quoted text
I have to admit that I'm not impressed by these voices. They're too jerky, even more so than the previous expressive versions. Even the conversational one is pretty bad about it, though i I like how it allows the artifacts of her voice to show through, you can actually hear her vocal artifacts like at one point her voice croaked a little, but it was clearly the real deal rather than concatenation problems. The built in google voices on Android are starting to get it right, the male one is expressive and reasonably smooth while some of the vocal artifacts are being represented. In my opinion, most of the concatenated voices are not reproducing very well. They're nearly all jerky, and their intonation is just goofy. And that goes for everybody. About the only ones who don't screw up the inflections is the ones that don't actually do much of it to begin with, Cepstral comes to mind. For Nuance, a couple of the later versions of Tom are pretty good, as is Samantha, but the rest meh. Ivona is amazing at not being jerky, but they somehow are difficult to understand, like they're just muttering along or something although Sally isn't quite so bad as Joey and the others. I vaguely think the Alexa voice is based off of one of the Ivona voices but I don't know that for a fact. She is an incredible voice, but being a female, I understand very little that she says, significantly reducing the actual usefulness of the echo devices for me. I do much better using the male voice for the google home devices. I prefer the one called pink although it has been a while and that voice may not be called that anymore. Out of all voices I still find Eloquence and ESpeak, in that order, to be the easiest to understand for my hearing. I even have difficulty understanding DeCtalk. The only DeCtalk voices I really understand well are the earliest ones like 1.8, and the speechmaster2000 that sounds vaguely similar to very old DeCtalk. Anything newer has a really odd artifact where it starts tumbling when it gets to read anything that is longer than 5 words or so. The syllables get shortened and it stops anunsiating clearly, which, by the way, is another thing that I dislike about the Nuance voices they all tend to do that too, one notable exception is Tom. Apple's Alex voice does this badly although it can be somewhat reduced by running at a very slow speed. It's like those voices get distracted with what they are reading and they forget to read out clearly, and so they just start muttering to themselves. Ivona's voices did not do that. One of my favorite voices that I forgot to mention is the LH True Voice peter. Very plane, sharp, smooth and consistent the only problem was all the spelling and the excessive inflection when encountering exclamation marks. That is one somewhat human sounding voice that I can actually understand well enough to read emails with. The main reason that I use Android is because I can put Eloquence on it. Using the iPhone with the Alex voice is doable, but tends to be a lot more stressful just for struggling to understand him. If I'm in a car there is an even wider gap between them, because I am so good at understanding Eloquence that I only need bits and pieces to make out what is being said where other voices I really need to hear everything very clearly and even that's no guarantee. Of course, ESpeak is also good enough to use I just hate the sound of it. RH voice is getting there, it vaguely reminds me of the above mentioned LH True voice, but it's too much bass and too little mids and highs. They're a bit quiet. I don't find myself easily able to read emails with them yet. Cheers: Aaron Spears, AKA Valiant8086 General Partner at Valiant Galaxy Associates "we make (VERY GOOD AUDIOGAMES) for the blind comunity" http://valiantGalaxy.com On 9/11/2020 1:14 PM, Gene wrote:
The web page where you can read about the voices and find samples is: |
|
|