The voice behind text to speach
|
Chris Shook
Hi,
Does anyone know who the voices are behind the new Vocalizer voices? That is, Oliver UK, Susan, Nathan, etc. |
|
|
|
Brian's Mail list account
Why do you think they might have a fan club?
toggle quoted message
Show quoted text
That would be a neat twist. Brian bglists@... Sent via blueyonder. Please address personal E-mail to:- briang1@..., putting 'Brian Gaff' in the display name field. ----- Original Message -----
From: "Chris Shook" <chris0309@...> To: <chat@nvda.groups.io> Sent: Friday, April 19, 2019 9:50 PM Subject: [chat] The voice behind text to speach Hi, |
|
|
|
Am I safe in presuming, based on what's been said here, that Vocalizer voices are not synthesized, but "assembled" from recorded sound samples of a single human voice?
-- Brian - Windows 10 Home, 64-Bit, Version 1809, Build 17763 A learning experience is one of those things that say, "You know that thing you just did? Don't do that." ~ Douglas Adams (1952-2001) |
|
|
|
JM Casey
Yes, that’s apparently so. I was using the Daniel (british english) vocalizer voice for a while and I found out the name of the actor whose voice was sampled – John something? Damn, I can’t recall, now. But yes, it’s not a pure synthesised voice at all; they have very human character but imo can’t really be used properly at fast speech rates.
From: chat@nvda.groups.io <chat@nvda.groups.io> On Behalf Of Brian Vogel
Am I safe in presuming, based on what's been said here, that Vocalizer voices are not synthesized, but "assembled" from recorded sound samples of a single human voice? Brian - Windows 10 Home, 64-Bit, Version 1809, Build 17763 A learning experience is one of those things that say, "You know that thing you just did? Don't do that." ~ Douglas Adams (1952-2001) |
|
|
|
On Sat, Apr 20, 2019 at 01:10 PM, JM Casey wrote:
they have very human character but imo can’t really be used properly at fast speech rates.No surprise there. Even with audio methods that allow you to speed things up without distortion, there are definite limits when one is piecing together really short samples to get continuous speech. Truly synthesized speech is generally much more "pliable" when it comes to pushing it to hyper speed (which I still marvel at the fact that many of you who use screen readers and have for years can actually understand it. I can't process at that speech rate for anything, and with slight hearing loss in the high frequencies it's even more difficult than it once might have been). -- Brian - Windows 10 Home, 64-Bit, Version 1809, Build 17763 A learning experience is one of those things that say, "You know that thing you just did? Don't do that." ~ Douglas Adams (1952-2001) |
|
|
|
Gene
As I understand it, the newer kind of synthesized
voices are assembled from a voice having been recorded saying all sorts of
things. And that, it seems to me, is why they sound so horribly artificial
in terms of using proper intonation. It is using and combining words in
its library as well as sounds letters, etc. Eloquence and Dectalk, being
the older kind of synthesizer, I'm not sure exactly how they work, can do a much
better job of having proper or reasonable intonation. I don't know if
there are better ones, But I hate listening to the new kind of synthesized
voices I've heard, with their improper intonation, sounds in words or in
connecting words that sound as though they are being spliced together, and
improper pause lengths.
Gene ----- Original Message -----
From: Brian Vogel
Sent: Saturday, April 20, 2019 12:06 PM
Subject: Re: [chat] The voice behind text to
speach -- Brian - Windows
10 Home, 64-Bit, Version 1809, Build 17763 A learning experience is one of those things that say, "You know that thing you just did? Don't do that." ~ Douglas Adams (1952-2001) |
|
|