People here have been complaining about the lack of so-called neural voices and updates to RHVoice. Don't remember who made the add-on and posted about it here, but I greatly appreciate his/her efforts and would like to ask him/her whether (s)he would be willing to (help) make the voices available on iOS.
Comments
if you'll give me piper tts on I devices, you are king!
really, I would like to get this on I devices.
eSpeak/eloquence
eSpeak or eloquence for me. I've always thought those human sounding voices sounded stilted, more like poor imitations. I prefer the blank slate of eloquence where I can project my imagination of what its suppose to sound like.
It's not as easy as all that.
I wish it was but, the person would have to pay $100 per year, I believe where as on windows it's free.
Also, it's good at reading but typing is terrible for now, depending on what voice/accent you use.
I can't use robotic tts, I just can't.
I just freakin can't.
This thread is supposed to be specifically about Piper.
I created another thread for general discussion earlier but people have been making their own separate threads even to repeat a question that has already been asked over and over. So this is why I finally decided to make a separate thread for Piper so that it might draw more attention and perhaps enable us to actually take steps to make it available on Apple devices, but then you began posting your thoughts on different TTS engines and your preferences. Head over to the general discussion thread if you're willing to discuss whether so-called humanlike voices or robotic ones should be preferred. I can already list several well-known and repeatedly mentioned arguments for and against both, like the former being more natural and the latter being more responsive, so I request you to not deviate from the topic. Anyway, I also do use Eloquence and eSpeak-NG but not in all languages and situations do they meet my needs. I would use them to browse the web, read e-mails or play games, but not to listen to a book. I am totally aware of how even AI voices remain far from realistic particularly in certain contexts and languages, but what's wrong with wanting to have yet another (free) option besides the currently-available ones, many of which also have their own problems? Whoever's been messing around the dictionary entries, have been making it more difficult and unpleasant to use the Turkish Vocalizer voices for me. The Arabic voices hardly ever read anything correctly; eSpeak-NG's support for Asian dialects of Turkish is broken, and I doubt we'll ever have RHVoice on iOS with all the languages and dialects that it supports. I'm almost sure there are other people who speak other languages and don't want to rely on eSpeak-NG and Vocalizer, as Eloquence already supports only a few languages. Piper's support for new languages and voices is expanding and I wish to have it not only because it sounds more natural but mainly because I can at least understand content that I hear in languages poorly supported by eSpeak-NG from Piper voices. Oh one more thing: Vocalizer voices have been just committing suicide and acting like mad. It's even quite difficult to be able to download and use some of them.
Re: Unless you’re somebody who just asked for moderator for
Hey Singer Girl,
We understand your perspective, but our posting guidelines do explicitly ask that people keep their comments on the original topic. The guidelines state: “Please stay on topic. Always stick to the original topic that was stated by the person who started the forum thread. If you have a suggestion or comment that is on a different topic, please start a new thread.”
The original poster here asked specifically about Piper voices, so comments about other voices, while potentially interesting, divert things away from the original topic. We kindly ask that we respect the intent of the original poster and stick to discussing Piper voices here. If you'd like to discuss other voice preferences, feel free to start a new thread on that. Our guidelines are designed to facilitate more meaningful and organized conversations - not to boss anyone around.
Hey, calm down!
You've been posting over and over about your favs that I've memorized (Samantha and Karen) in a thread made for another purpose; others have been posting about why they don't like "humanlike" voices, and all I do is tell you there should be nothing wrong with wanting yet another option. So it's already me who made that other thread for general discussion; it's already none other than me who encourages discussion and tries to talk about the possibility of having Piper on Apple devices, but then it's also me who gets such a response saying there's nothing wrong with sharing your opinion. Sure, I don't disagree with you on that, nor do I assert I'm a moderator and can force others what (not) to do. Just do bear in mind that I'll make sure to remind you of this reaction of yours if you ever happen to find irrelevant posts in a thread you make or follow. So go make your own threads, post your own stuff but do it appropriately.
Having piper would be interesting
The only thing I can see which would be tricky, though not impossible, would be the $99 per year, although if this is open source and people have the ability to donate through the tts app, in theory it shouldn't be a problem. But the more voices we have and the more choice we have, the better. Maybe a funding effort could be started by @op or the entire applevis community to get the ball rolling? Although before that, we should find out if the developer would be open to bring this to the apple ecosystem before looking at options like fundraising etc. If the dev is open to it... Then why not?
I can't see this now
These voices currently have a problem. They are not intended for use with screen readers. They would be terribly slow. It already happens with NVDA on Windows, they are not usable for reading text. It would take a good developer and I'm not sure it would be worth the effort.
The x-low-to-medium-quality voices work smoothly.
Has anyone used Microsoft's so-called neural voices on a Windows 11 computer, or the new Siri voices on a Macbook? The Siri voices are responsive enough on the iPhone, and so should they be on the Macbook. So if Microsoft's "neural" voices are unresponsive on Windows like Piper voices with NVDA, then this might have something to do with the operating system itself rather than the voices, as I doubt the same voice would be as responsive on the iPhone as it is on Windows. So if the Piper voices are unresponsive on Windows and the Siri voices are quite responsive on iOS, then the voices might be made more responsive in some way or you might just stick with lower qualities. Well, I'm just making guesses and assumptions.
Fundrasing/Crowdfunding
As the app is downloaded by more users, the revenue will probably help compensate for the expenses. And you can always ask some global giant to help with funding, as the case with NVAccess.
If you use the british voices it won't work.
It might in the future but for now you really can't type with these voices, at least on windows you can't.
Samples
On the Piper GitHub website, there is a Voices page with pop-up menus to choose the language and the voice as well as the quality, although there was only one quality for the voices that I tried. There is also a button that says Speaker, but that didn't do anything. Then there is a play button that will play a sample sentence. The Piper GitHub page: https://rhasspy.github.io/piper-samples/
Most of the YouTube videos that I found when I looked it up appeared to be coding tutorials.
Opened an issue on GitHub.
Let's find out what happens... Those interested in having the voices on their devices, might have to contribute to funding the development.
r:Disgusted
I doubt I will change anytime soon. What the voices lack is context, and until the machines develop far enough to understand what they are saying, they won't be able to deal with context. I'm not very familiar with piper though I've tested it, but I'll give an example from Vocaliser. Vocaliser danyel is a very upbeat voice. That works very well if your reading an e-mail wishing you happy birthday, or if if a business is using this tts to greet their customers. But if you're reading an e-book of the call of Cthulhu, that voice isn't going to cut it.
Time to revive this thing!
Well, iOS now even lets you use personal voices. And now that Piper is faster thanks to the new Sonata add-on for NVDA and "fast" variants, why not organize some crowdfunding thing to have it on iOS? eSpeak-NG is lightweight and responsive but its voices are robotic. RHVoice haven't been updated for ages and I doubt it will ever be. So we should probably focus on using Piper voices on iOS with all the different options like different qualities and variants for different models and hardware requirements.
Android version?
Is there an android version of this?
Thanks
Yes
Possibly unofficial though; don't know.