Apple Reported to be Exploring Ways to Let Developers Provide Custom Text-to-Speech Synthesizers on iOS

Member of the AppleVis Editorial Team

Over the past few days, 9to5Mac and MacRumors have posted details of several iOS features that are reportedly in development at Apple.

Based upon findings in supposed unreleased iOS 14 code, one of these reports is of a new developer framework named VoiceProvider, which could let developers replace built-in speech voices on iOS with custom text-to-speech synthesizers.

Information on how this might be used is speculative, with MacRumors saying:

Voice synthesizer extensions would be bundled with apps from the App Store, and then could be offered to the system to replace the default speech voice, which could be important for languages and dialects that Apple does not support. This could potentially extend to an API for the HomePod, although this is not for certain.

iDownloadBlog.com takes the speculation a step further by suggesting that custom synthesizers could “range from Space Odyssey’s HAL 9000 computer to the Majel Barrett-Roddenberry computer voice from Star Trek”.

The leaked code is currently in the early stages of development, so much of what's been found might not make it into the first iOS 14 beta, or even into iOS 14 at all. Accordingly, all that we can reasonably say for now about what this finding might mean for VoiceOver users is to “watch this space”, but it does certainly appear that Apple has something in mind.

If you think you know what that might be - or have a wish - we would love to hear in the comments to this post.

In its exploration of the leaked iOS code, 9to5Mac has discovered what might be some major new accessibility improvements coming in iOS 14:

Our code findings uncovered a new feature that will be able to detect important sounds like fire alarms, sirens, door knocks, doorbells, and crying babies. Presumably, iOS will translate these alerts into haptics for people who have hearing loss.

More new accessibility features would bring support for the camera to detect hand gestures while “Audio Accommodations can improve audio tuning over AirPods or EarPods for people with mild to moderate hearing loss”. Further iOS 14 should offer the ability to run an audiogram on-device to help the users tune the audio.

Other findings from the supposed iOS 14 code include:

Blog Tags: 

19 Comments

well,

hope voice over can do something in gaming
like narator the mainmenu of the game and stuff
or when we use controler. we can access the game menu via speech or voice over

Bugs

I prefer that they focus on fixing the many bugs and them do so. Bugs, bugs are not fun. Specially with unlucky 13 which is a major pain.

I can't wiat fo rhtis, and as

I can't wiat fo rhtis, and as fo rthe bugs in iOs 13 I can finally close the other haof of the bugs I've reported. It si rock solid and stable.

Question

I think you stated the iOS is more stable, I hope that is the beta of 13.4 because it is not for most people. FYI. iOS 13.4 beta 5 is out.

Staying on topic

While I perfectly understand your frustrations with IOS 13, this post is not talking about that and you can comment on one post without saying unlucky 13, 13 has bugs, and repeating it over and over again. Ultimately we here can do nothing and it is only Apple who can fix them. Understand however that at present some people cannot use even the buggy 13. Why? Because they do not speak English and there is no TTS available for their language. Serbian is a great example. We have to use Czech vocalizer because it is closest when it comes to reading of our texts correctly. Even then, there are many limitations. You cannot read the Serbian Cyrillic script, and some things are obviously pronounced incorrectly and don't sound so good. If Apple does this, this will be the biggest and most positive change for our country and will correct one of the major flaws of IOS for us. SO please stop dismissing things so easily. I realize that for people in US it might not matter that much, you can use a few extra voices that might be developed, but you already have high quality voices. We don't however and this would be an essential feature for some people. Thanks for understanding.

New leaks

Club AppleVis Member

A new leak reports that VoiceOver will gain the ability to communicate text formatting changes through speech, sound, or braille, using the reader's preferred braille code's formatting symbols.

This "leak" may also be wishful thinking on the part of the author, who really wants text formatting, animations, and other such visual information, to be communicated through VoiceOver, and may not be entirely accurate. However, sources say that this leak is real, and that iOS 14 will have many more VoiceOver features for public beta testers to hunt down.

Note: Don't trust articles with uncited sources.

eSpeak

hope ios will get ESpeak, it's only one tts whitch supports georgian

eloquence

well,
eloquence is also very fast as well.

anything is possible

well,
some of the speech synthesizer like like neo speech can add to the app too.
I am looking forward to it!
and hope voice over can read the game menu when plug in the controlor

Awesome!

This is a very interesting article. I know these are rumors but I hope they become true. If I could get the Heather voice that's in VDR that will be cool.

Great news, but one possible problem

Member of the AppleVis Blog Team

I’m glad this possibility is being explored, and I hope it does happen. Maybe all the people who keep requesting Eloquence will finally get their wish. I do have one worry though. If VO crashes while using a 3rd party voice, or something goes wrong with that voice, will VO be able to recover and revert to one of the default voices if necessary? I used to use Eloquence with NVDA and if, for whatever reason, it couldn’t check that the licence was still valid, it would keep repeating “please register this product.” And there was no shortcut to get back to one of the default voices. That particular issue probably wouldn’t happen on iOS, because of the way Apple handles purchases, but they’ll still need to make sure any stability issues with 3rd party voices don’t make VO unusable. So long as Apple can be sure to iron out any problems like that before release, this will be a very positive step.

OCR may be on the way too

I read on Mac rumors, i'll find the link, that possibly OCR will convert handwritten text to typed.
Link:
https://www.macrumors.com/2020/03/09/ios-14-ocr-apple-pencil-text-conver...
This is more about the apple pencil, but keep reading. If they take that one step farther, we might have OCR for inaccessible dialogs, menus and the like. My only other worry, besides what's been posted is the size of the OS might increase. I know we can download individual voices, i also know that the small phones are way bigger than my 64gb model. Maybe there is some truth to it. Although Voice over needs a lot of work, I'm afraid it will become more bloated then it already is in some aspects. I'm not putting anyone's wish for Espeak and Eloquence down, but remember at least with Eloquence there is no more support. So even if it was supported by IOS14 would it be in Apple's best interest to release a modern OS with a voice so outdated?

Well, being that eloquence

Well, being that eloquence is not a 64 bit app, and just got kicked off the android store cause code factory just didn't want to take the time to make it a 64 bit app, I don't see this happening any time soon ish.

Espeak would be good and maybe RHVoice for those of us with slower phones etc.

Voiceover TTS

I'm really hoping this means we could see Google TTS being used with Voiceover. It's one of the few things that I love from them and am quite partial to the UK female voice. Alex and Daniel just aren't the same for me.

It'll be useful for other languages

It'll really be useful for other languages. I absolutely love some voices that exist for Afrikaans.
I love reading books in Afrikaans but right now it's a 2 step process. I have to convert the book in mp3 using my windows PC and text aloud (since it only works on windows), then import that MP3 file into voice dream. Seems a bit silly but it's the only way I can think of doing it. It'd be nice if that could be synthesized on the phone itself.

google. text to speech

one of the google text to speech language (cantonese) is already in IOS 13... one of the siri voices.

Re OCR

This is a little off topic but as one of the other posters mentioned, it looks like the Apple Pencil will gain handwriting OCR which is an interesting concept for those of us who might want to physically take notes with a pencil. How effective it is for someone who hardly writes and doesn't have the closed loop of refining motor skills for such things, who knows. interesting though.

And, yes yes, to new voices. I'd guess this is going to be more app based rather than necessarily system wide as, another poster pointed out, apple might not be happy handing over voice feedback of VoiceOver to an outside party. As buggy as it is they'd not want to risk breaking it further. To be honest, functionality is more important than the voice itself. I'd like eloquence even though it's a dreadful imitation of human speech, but because I'm just familiar with it.

Hello ,

Hello ,
I want e-speak as well