Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

MacRumors

macrumors bot
Original poster
Apr 12, 2001
67,708
38,185


A few years ago, Apple introduced a Personal Voice feature that allows those at risk of losing their ability to speak to create a synthesized voice that sounds similar to their actual voice, so they can continue to communicate with others. The feature debuted on the iPhone with iOS 17, and it will be getting even better on iOS 19.

Personal-Voice-iPhone-Feature.jpeg

Apple today announced that Personal Voice will be faster and easier to use on iOS 19, thanks to advancements in on-device machine learning and artificial intelligence. Apple says users will be able to create a smoother, more natural-sounding voice in less than a minute, down from 15 minutes when the feature initially launched.

Personal Voice will also add support for Spanish (Mexico), according to Apple.

Personal Voice integrates with another accessibility feature called Live Speech, which lets users type what they want to say to have it be spoken aloud during in-person conversations, phone calls, and FaceTime video calls.

Personal Voice is also available on the iPad and Mac, and the enhancements to the feature will extend to iPadOS 19 and macOS 16.

Article Link: iOS 19 Will Improve iPhone Feature That Lets You Preserve Your Voice
 
  • Like
Reactions: MaximizedAction
bringing back my anecdote on this before anyone gets to being childish in the comments.

my mother had dysphonia for nearly 20 years of her life; I never heard what my mother's voice truly sounded like until a procedure in ~2014 that brought it back. it has now been nine years since I lost her and man, what I wouldn't give to have an iPhone of hers with this on it. always cool to see Apple improving the feature.
 
I think aspects like this are where AI can shine. If you have this setup or some voice recordings. If you ever lose the ability to talk, this can preserve a bit of you just incase.

The only safeguards need to be where regulation protects people from fraud and misuse.

AI will lead to great advances but people and government need to be on top of it without fear of the tech companies. And tech companies need to keep the social contract in mind. It has to be a core aspect of the tech companies moving forward. And governments should be fearless on this. After all, in a democracy the people vote in the people who are to represent us, protect us and to serve us.
 
Ok. How is this not advertised? This is amazing! Do ads about stuff like this. Theres a few nice little features, like the control center playlists, that unless you stumble across an article about it you'd never know about.
 
  • Like
Reactions: arkitect
Very cool! It was pretty cumbersome in the previous iteration. 15 minutes of recording but then a fairly significant amount of time to process on device too. This is a great step forwards.
 
bringing back my anecdote on this before anyone gets to being childish in the comments.

my mother had dysphonia for nearly 20 years of her life; I never heard what my mother's voice truly sounded like until a procedure in ~2014 that brought it back. it has now been nine years since I lost her and man, what I wouldn't give to have an iPhone of hers with this on it. always cool to see Apple improving the feature.
Ditto. My mother passed away last year. I tested this voice feature myself and found recording 15 minutes of sentences cumbersome. My mother would have struggled to do that, but I wish I could hear her voice again, even if it is a synthetic one
 
  • Like
  • Love
Reactions: phenste and idrewuk
Ditto. My mother passed away last year. I tested this voice feature myself and found recording 15 minutes of sentences cumbersome. My mother would have struggled to do that, but I wish I could hear her voice again, even if it is a synthetic one
I understand hat you’re saying but for myself, once a person is gone they’ve gone. It’s fine to have actual recordings and play them back as they are recordings of what actually happened.

But to hear your loved ones voice again, but for it not to be them putting those words together. I don’t think many would really want that. And critically that person wouldn’t have any control or consent for the continued use of their voice. Even if they gave you consent before passing away, sometimes people change their minds. That would no longer be possible and they would have no control over what others would be subsequently getting their AI voice to say. It becomes an ethical issues.

On the other hand if it’s someone who can’t speak anymore, but who is using it to communicate. Using their voice, and words that they themselves have put together using it in full knowledge is awesome.
 
Ok. How is this not advertised? This is amazing! Do ads about stuff like this. Theres a few nice little features, like the control center playlists, that unless you stumble across an article about it you'd never know about.
They did in 2023 when this feature was first announced, since removed from YouTube but copy here -
 
  • Like
Reactions: jimbobb24
This is a huge improvement. Previously it used be very difficult to create it as it took a long time to record all the phrases. It will be good if there is an option to smoothen out already recorded voices.
 
  • Like
Reactions: idrewuk and mganu
I tired this and it was only ok. I work with lots of people who lose their voice and the feature is huge. I see people use relatively light AI models to produce very good imitations with small samples so hopefully this can get improved and easier. It really is a meaningful feature for those who need it.
 
Last edited:
I understand hat you’re saying but for myself, once a person is gone they’ve gone. It’s fine to have actual recordings and play them back as they are recordings of what actually happened.

But to hear your loved ones voice again, but for it not to be them putting those words together. I don’t think many would really want that. And critically that person wouldn’t have any control or consent for the continued use of their voice. Even if they gave you consent before passing away, sometimes people change their minds. That would no longer be possible and they would have no control over what others would be subsequently getting their AI voice to say. It becomes an ethical issues.

On the other hand if it’s someone who can’t speak anymore, but who is using it to communicate. Using their voice, and words that they themselves have put together using it in full knowledge is awesome.
In 10 years we will have tailored LLMs with visual and audio avatars that your grandparents all trained to give you advice after they die. Some crazy stuff is going to happen.
 
where does the data that contains the qualities of your voice exist?
locally only ? is it backed up in iCloud backups and therefore unprotected in some situations ? etc etc.
i would want to understand how protected that data is and how for me to personally delete it before i ever authorised or used it.
 
  • Like
Reactions: alexlc1 and idrewuk
It is a great feature that I would genuinely be interested in… but at present there are only two options to create your voice:

English (United States)
Chinese (Mandarin - China mainland)

Hopefully this update will allow more for those of us who are not USA based or born and who would like to do our own versions…
As your list shows not only USA based people are supported.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.