I see ElevenLabs now has a ‘pay-as-you-go’ TTS marketplace. As present you pay as you generate audio you can then do whatever you want with it. It can’t be too long before this technology could be packaged into a TTS engine. This raises the tantalising possibility of giving everyone a personalised individual screen reader voice!
My question is this – would you want this and would you be willing to pay for it?
I paid for additional voices for NVDA and was willing, if not exactly happy to do so. This was more because for 30 years the voice was the screen reader and there wasn’t a choice. I do remember buying various voices for various in-car GPS gizmos in the past. TomTom with Mikee the New York cab driver was my favourite.
So, what do you think? One-off purchase or rental so you can change with the seasons or your mood? And who would you choose?
Comments
Unwilling to even think of paying anything
We pay for subscriptions right and left. Internet, TV, apps, now a screen reader? Why can't we get off the soapboxes of I want a personal voice for the spring season, or I'm sad, so let's make my screen reader cry my messages out to me. Why can't we take Apple, google etc to task for dismal accessibility? why can't we be tolerant of the choices of voices we have, instead of asking for more and more every year? why can't we take these tech companies to task for providing no language support or if it is supported, the minimalist support possible? Such as say Turkish for example. apple says, we need to support this language, so let's outsource to a company that's no longer in production, is years out of date, simply to give our access to the customers. They, tech in general, are not thinking about the consumer and what they want. they are focusing on releasing things faster, newer shinier devices year after year. When is enough enough? Why do we want more and more? are we ever going to be happy? I come from computing days where you had the echo synthesizer and were able to do what was needed without sacrificing the work flow. Let's go back to what matters, software, accessibility, and increased usability. Let's leave the theatrics to people who really want to have them. I'm not about to shell out for anything like this idea.
If they can inflect, then maybe.
If it could scan a peace of text and gather the needed emotion for it, sure, but if it's just 11lab voices, nah.
They're alright but I feel like piper tts is better in some ways.
@siobhan, I do understand where you're coming from and even agree with you up to a point, but having a more emotional sounding screen reader for reading text, it'll have to scan it to get the right mood, would be nice.
I do prefer eloquence and probably always will but it is nice to hear up and coming TTs stuff to see how they've improved it.
Oh.
If you're talking about a famus person or someone like that reading to me, I don't think i'd go for that.
The only two I might consider would be Stephen fry and David Attenborough and even then that would be after being asured they can inflect correctly.
Productivity
I think it would be pretty frikkin' cool to make VoiceOver talk like Samuel L. Jackson. But this is really just ear candy, isn't it? New voices aren't going to increase my productivity. There are several far more critical ergonomic and usability issues that hardware and software developers should address.
It's a nice idea
It's a nice idea but I doubt I would go for it. I certainly don't want to make myself into a personal voice for my iPhone. Famous people? Jacob Rees Mogg would be good. Boris Johnson would be good. nigel Farage would annoy people so would be good. James O'brien would be good if Voiceover developed the ability to determine whether a document was a sanctimonious complaint letter.
Hang on hang on hang on - I tell you what would be good: suppose you were reading an article - "Mourinho hits out at disgraceful refereeing". Imagine if Voiceover were able to be clever and capable enough to read the Special One's remarks in the voice of the Special one? I suppose you'd have to have a huge library of voices to make this happen but if it were possible, you could choose someone as the narrator and then click on the 'enable character detection' setting and you'd be up and running. in fact, while writing this paragraph I have become so absurdly excited by this gem of an idea! what a clever fellow that Mr Little is.
of course, none of the above matters - not character detection, not personal voices; none of it. why not? Well, because as soon as you put the bleedin' iPhone on charge, they'd all be deleted! You'd be back to daniel Compact, samantha Compact, or whatever, and you wouldn't be able to download Nigel Farage again. Bad luck.
Siobhan, cheer up! Ireland are playing so well in the Six nations.
Charlotte, that wouldn't work
Klopp would, one day, begin the reading of an email or text by announcing to you that he was leaving your iPhone at the end of the season. Then there would be a wailing and a gnashing of teeth.
You paid for extra screen reader voices?
I feel so sorry for you, especially if your payments included Eloquence for NVDA, which can be had for free via Github, along with many other voices in the form of downloadable add-ons for NVDA if you know where to look. I've never paid for an additional screen reader voice but am also tech savvy enough to know where to find the ones I want, in fully working order, without being charged for them.
Beyond that, though, remember that a pay-as-you-go, subscription-based screen reader model would get quite expensive quite quickly. Just consider how many words your screen reader speaks in the course of a single day, or even a single hour or minute. Even if you set something like VoiceOver to use the absolute minimum verbosity possible, and to speak as little as possible, you would still be looking more than likely at an outrageously expensive bill at the end of the month, unless you're someone who only uses their device for five minutes per day. This is because screen readers, by their very nature, are designed to speak a lot. If they weren't, you wouldn't be able to get all of the information you need to reliably understand the screen. And given the many complaints from users on this site about the affordability of even Apple products that would be considered relatively inexpensive by most other standards, I doubt many people would be able to put forward the high cost of a subscription-based screen reader without significant cutbacks on spoken content.
Of course not
Screen readers should stay as screen readers. Yes, I like natural sounding voices but paying for sinthesised voices? No, of course not, thank you very much.
Why would I want my screen reader to yell angreely for example? Or why would I want my screen reader to laugh a funny comment? All a screen reader have to do is reading what is on the screen properly, with out bugs.
The problem is response
Screen reader responsivity is the main issue.
I mean: you can have a 100% customized voice with all parameters expressiveness, prosody, tone, style, whatever you want about human sounding.
But while it's pleasant to hear when reading a text -book or similar-, what about when you have to perform commands? Basic screen reader commands but even in games or whatever. Yes you can have per-app voiceover activities, but you should set them up everywhere.
As soon as Eloquence became available in iOS, I quickly switched to it because -especially with wireless earphones- when writing long portions of text or performing many commands in short time, human-sounding voices gave me those responsiveness issues.
So, if I have to pay for a quickly responding voice I'd do it, but not for a slow one.
I am a happy ElevenLabs subscriber -creator plan- as I am preparing some audio stuff, but one thing is a recording, another is screen reader.
Commands?
what ho all,
Oh come on, where's your imagination? Imagine Kieth from the Prodigy voicing your screen reader: "Press the left-hand cursor! till it won't go any further! I'm a firestarter! a twisted firestarter!"
Oliver, don't give me all that hogwash about reading. You're a student. You won't do a tenth part of the reading you're supposed to have done, just like all the others.
I take it Bingo's idea of character detection has hit the buffers, then? You don't want nigel farage's smoker's cough casually interwoven into your emails?
And, Neosonic, I am very impressed by your fiscal responsibility. I wish you were the Chancellor of the Exchequer. perhaps I wouldn't have to pay out £22,250 stamp duty if you were.
I see siobhan hasn't checked in re the Six Nations. I wanted to know whom she would pick as the Irish hooker.
and on that point, sir, you're just being silly...
Depends on how much
I think there are 2 elements to a voice. Firstly, the practicality - is it easy to understand what it is saying or does it mispronounce, try to be too clever (e.g. pronouncing version numbers as a date) or do they sometimes pronounce two things in a way that's hard to differentiate.
And then there's just how nice the voices sound. Is it actually pleasant listening to them. The Microsoft Natural voices and even Siri voices work here for me. Siri is definitely no good on the Mac unless bugs can be fixed (ha ha). The MS voices sound amazing though.
I'd probably pay for improvements to either. But the problem with a voice is that you don't really know how well it works until you've used it for a while. I thought Jamie was decent until it said mark and then it had to go in the bin. Sadly all voices on the Mac have quirks that I struggle with, so if they could be fixed then I would consider shelling out, sure.
But surely Apple will add some more realistic voices soon given how tts is improving right now.
This would be Uber Cool!
Hi everyone. I just got my second iPhone earlier this week. I am on the series 14 and while I'm happy with what we currently have, it'd be so cool to have a celebrity reading things and echoing back commands, and the list goes on. I think I read awhile back that Cereproc has a Donald Trump voice. I don't want to get into politics on here because that's off topic, but while I strongly disapprove of Mr. Trump I wonder if Apple is eventually going to implement his raspy voice. There are of course others which I'd like to see implemented, but I'm not so sure I'd pay for a voice. Perhaps that's why Infovox iVox is gone? I actually tried out a couple of their voices and was quite impressed.
I am more than happy to pay
I am more than happy to pay for a voice for screen reading. On Android, I paid for lots. Voxygen, some sirious voices and novalty, like helium and Witch, Seroprock, though the issue with making TalkBack unresponsive is even werse with VoiceOver, to the point that the phone gets intensely hot, and this inspires me to write a one star app store review. I all ready told Seroprock support what I think, and that's why the voices are so cheep. Elooquence for Android was expensive, but was, and still is, werth every penney, Accapela voices, I paid for lots of them, and they are amazing! I do not condone cracked voices, unlike some on here. fact is, these developers of screen readers and voices do it for money, amd they work hard, giving us a quolity product. so if a voice became available, and I liked it, sure, I would pay! but Seroproke do not deserve our support and money! although their voices are just as good as Ivona were, they suffer from the same issues!
OP asks, "My question is…
OP asks, "My question is this – would you want this and would you be willing to pay for it?"
I would not want to use this, and would not be willing to pay for this. I'm not "tech literate," so I use the default voices of the devices, Samantha Compact, in the case of the iPhone, and espeak English, in the case of my desktop computers. If the default voice is one I can't understand. even after a time of getting used to it, then I would probably just avoid that device or software because at some point, I would have to set it up or reset everything, starting from the default voice again. I do allow myself to adjust the pitch and speed of the default voice.
I went through a faze of trying to find the perfect voice back when I used JAWS, then NVDA, and never found it. I even got yelled at, long, long ago, by a resource tech teacher at a blind school for adjusting the pitch of a Braille & Speak device to a higher pitch.
OP asks, "So, what do you think? One-off purchase or rental so you can change with the seasons or your mood? And who would you choose?"
I can't answer that question, but rental is something I avoid.
what would br nice
what would be nice, is if I could use my personal voice I created in the iOS 17 beta cycle, as a VoiceOver voice. that would be amazing! as far as other voices though, Freddey HIber, the acter who plays Shawn on The Good Docter would be nice. even David Woodbridge would be nice, too.
Re: own voice
I can't think of anything worse than using my Mac and hearing my own tedious little whiny voice coming out of the speakers. Anything but that please!!
I can't help but feel celebrity voices would be a bit of a novelty. The Sonos has that actor from Breaking Bad as its voice assistant and it is awesome but not sure it would work as a screen reader. Sure I'd try them if they were free, but if I'm going to pay I want something that makes me more productive and not just for the fun of it.
I also wouldn't pay a subscription for a voice, or even for access to a range of voices.
I think it would increase my productivity
I think having a voice like some of my favorite people would increase my productivity better, I think I'd focus better, having a voice I enjoy hearing, and am comfortable with. I'd pay for this.
But that's just me
@Charlotte
wow, i'd love these voices to be a part of my screen reader. I love fan fics,, so can't wait to see where this goes in the future.
I love that when it mentioned whispering it did it, the pronounciation of French words was nice too and I can imagine that helping me to restart my French learning.
It really is a great time to be alive .
Are there any more voice clips than these 3?
I'd like to hear more if posssible, here's the article I read: https://techcrunch.com/2024/02/14/largest-text-to-speech-ai-model-yet-shows-emergent-abilities/
William Daniels
A little backstory. . .
In the 1980s there was an amazing action/adventure/sci-fi tv series about a vigilante who would act as a fixer/private investigator/body guard who was in possession of a heavily modified automobile that, even with today's advances in technology, still does not exist. . . yet.
The series was called "Knight Rider", and the automobile was the "Knight Industries Two-Thousand", or K.I.T.T. for short.
This automobile had everything one could ever want in personal transportation; it had onboard GPS, wireless capabilities including telecommunications, an onboard PC, a fully sentient (and snarky) AI, was self driving, and was nigh bullet proof.
The main thing was KITT's snark. It was legendary for its time. I will provide a link below that will highlight that and some other interesting and comical facts about the series.
Now that the backstory is out of the way, I would absolutely pay for an AI that was as sentient and snarky as KITT. I would even be pleased if it used the original voice actor, William Daniels, as his voice just worked for the series. 😀
https://www.youtube.com/watch?v=tyntGDmxXxU
Vaguely Remember That...
One of my family members--I forget which one--had a toy car like that. There was a button at the bottom that when pressed, said: "Call me Kitt for short." It was pretty awesome.
1980s ftw
Well, that is the voice I would gladly pay for. Also, the 80s had some of the best action tv shows, ever!
i am dating myself here. . . 😐
Agree with Siobhan on this one
I respect those of you who think this is a neat idea, but I agree with Siobhan. I'm already having to pay for way too many subscriptions, so paying for a voice on my phone holds absolutely no appeal. Even more to the point, though, I would never, ever want to hear my own voice as TTS. I wish I could find a voice that I liked as well as Alex, just for variet's sake, but paying for one is most definitely not an option for my budget.
Sounds Intweresting
This idea sounds interesting but like others stated I would not use my own voice as a voice over voice. I also wouldn't pay a subscription for a voice because that $10 every month adds up... Now, If we could use existing voices from voice dream or the MS platform that would be cool. Go to preferred voice dream voice, and there would be a setting to "use as voiceover voice" I would do that. Let's see what happens...
Unwilling to pay.
I have to agree with the person who said that we have enough to pay for already. However, I would be willing to do a 1-off purchase if it was the voice of one of my loved ones. I'd be more than willing to hear the voice of my partner. As I ponder the daunting thought of knowing that her health conditions will most likely be the death of her, I would love to have her voice captured forever in a screen reader. Creepy? Perhaps. But that's how I cope, friends.