Piper - Neural TTS

I'll be giving this one a go.

I love the fact that these voices are now available on my iphone! There is no lagging between swiping and verbalization, there is no excess battery drainage that I can see, and this is just a wonderful app! Five Stars!

I’m not sure what’s going on. I’ve done everything I can within the screen that comes up after tapping the Help button. I don’t see it at all in my voiceover settings even after I have installed the Voice model. Is there anything I’m doing wrong?

Not something I'll personally use. I'm glad there is more choice for those that want it. An awesome job to the dev for bringing it to the iPhone.

I may try it out just to see what it is like. but I'll mostly stay with eloquence.

Make it so that one can play a sample without downloading and installing the voices? That would make it easier to preview and decide if one likes it or not before commiting to download and installing then beeing able to hear a sample.

so, if i install a voice, download, and install a 2nd the first voice not then available? and yes samples need to be available why i wonder are all the voices not offered as high quality?

seems good choice though of voices.

I have previously trained several piper tts voices on my computer with use with NVDA and Orca, but I am not sure what format they want the voices in to add to the app. First I tried uploading the onnx and onnx.json files seperatly, which did not work. Then I tried the NVDA way of a tar.gz with both files and the model card, then zip, then I tried tar, with no success. What am I supposed to do to get this working?

well, I hope it can have multy voices installed at once...
right nnow. it just can install one language and one voice

I installed the german Karlson voice. Then I installed something else cause I didn't like it. But no matter what voice or language I install it doesn't work, every voice is german karlson. Tried reinstalling the app and then no voice made any sound till I downloaded karlson again. And now I again only have karlson.

I've installed 3 voices but it only lists the last one in the app though the other two are there in voiceover... This is early version.

I have tried many United States voice from this app as voiceover voice
I can hardly understand them.
Here’s a screen recording which I did trying out those voices
https://share.icloud.com/photos/08axqwO4uaAPt5C6kFeTU6keg

I’m glad this is available for people who want it. But I’m going to stick with vocalizer voices of Karen and Samantha. Those have always in my two favorites since my first iPhone and I will just continue using those.

I tried one of the U.S. English voices but unfortunately there is a lot of latency when used with VoiceOver--not unlike what I saw when last I tried using my Personal Voice. I wonder if this is more of a limitation with the third-party TTS infrastructure than the voices themselves?

I Don’t think this is the case, as with other third-party synthesizers like espeak, there doesn’t seem to be any latency from what I can tell.

This version of the software seems to be riddled with the same bugs and/or limitations as the NVDA add-on version. The last time I tried to use the Piper TTS voices for NVDA, I ended up having to completely wipe and reinstall NVDA. Hopefully, this will work better on iOS. Best of luck to whomever uses it though, I'll stick with vocalizer voices for now...

As intriguing as this is, I do not think this should have been released to the appstore yet.
There isn't clear documentation on how to use it, custom voices don't reliably work, I've had the example text randomly play and of course the fact multiple voices can't be installed and used.
I do hope this improves with time though ☺️.

Hey everyone, Ihor here!

First off, thank you all for your feedback and kind words. I really appreciate it. I just want to let you know that development might be a bit slow since I’m working on this as a side project, but I’m doing my best to implement as much as I can.
Some of the issues with voice pronunciations may be out of my control, as the AI models themselves can sometimes have limited support for certain characters or languages.
For those asking about uploading your own models: to make it work, you need to select both the .onnx and .json files. The app will then parse the JSON and apply the model automatically. I’ll make sure to add clearer instructions to the Help screen in the next release.
Thanks again for all the support and suggestions - it really helps me improve the app!
- Ihor

I hope the apps can be more relyable and the next update

I am looking for more third party text to speech voices that can work with voice over

I need help with this app. Inside the app the P in the word help is very quiet or sounds like health or hell. Secondly where is the English United states model, and when will models be available?

I agree with the poster who suggested that we need to have the option to listen to samples before downloading the voices, especially since some of the voice names aren't exactly names. I've got it installed, but would imagine that I'll be sticking with my buddy Vocalizer Nathan.

greetings if you choose to download more than one voice, how do you switch between them if it don't allow you to cycle between the downloaded voices. there is a install or remove voice model but thats all and there is no way to cycle between the models you have downloaded already is there?

I downloaded the app to test it and see if it would work. Some of the voices just do not sound clear and I have trouble understanding them even the higher quality voices. And no, you cannot cycle between your voice models. It’s an interesting idea, but it has a lot of issues so I removed it shortly after testing. No I was not a beta tester, I was just testing it to see what it could do and apparently it was a total bust. I understand the developer is doing this for a sort of nighttime project which is fine. But it does need some major improvements I think. All the ideas I’ve seen here I agree with. It needs a way to preview the voices before downloading them, a way to download multiple voices and switch between them, the latency issue, all of that stuff I had issues with too.

What languages ​​does it support?

I downloaded the app but am not quite brave enough to commit to it yet. I had tried the NVDA plugin for them before but ended up switching away although can't quite remember the reason.

You can preview the voices on the github page I think but I don't know if it's exactly the same as hearing them with VoiceOver.

Some of the voices are quite nice, though, so it would be really nice if they become a viable option for VoiceOver.I'm just not quite brave enough yet.

The Polish voices in this app sound very strange and are almost unintelligible when using VoiceOver. They mispronounce words frequently and the speech patterns are unnatural, making the content difficult to understand.

I think the new update have been fixed some of the issue..
hope the voices will not disappear after restarting the phone and voice over