Apple Highlights iPhone Feature That Lets You Preserve Your Own Voice

[MacRumors]

Apple today shared a heartwarming ad for its new Personal Voice accessibility feature, available on the iPhone, iPad, and Mac.

Introduced with iOS 17, iPadOS 17, and macOS Sonoma, Personal Voice 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.

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.

In the video, called "The Lost Voice," physician and disability advocate Tristram Ingham uses Personal Voice and Live Speech to read aloud a bedtime story to a young girl. More details about Ingham's life can be found in an Apple Stories article shared today.

To create a Personal Voice, users read aloud a randomized set of text prompts until 15 minutes of audio has been recorded. The feature can be found in the Settings app under Accessibility → Personal Voice, and it is currently available for English only. Apple says the feature uses on-device machine learning for privacy and security.

Article Link: Apple Highlights iPhone Feature That Lets You Preserve Your Own Voice

 

[surfzen21]

Do you still miss that girl that got away?

Put her voice into your new AI robot and she will be with you forever. 😳

What a scary future we are heading towards.

 

[antiprotest]

I was excited about this feature until I tried it and the quality is really bad. It will improve in the future I'm sure.

Right now if you need to use it as an accessibility feature, it's better just to use one of the included voices, such as the Siri voices.

 

[aries81]

Hmmm…pass.

 

[swm]

this is the first relatable use of generative AI known to me that helps everyday people.
yes. it can be misused, and most probably will. but maybe apple has built in some recognisable pattern to easily distinguish between the actual person and the generated audio.

 

[ELman]

Does anyone know of a service like this that allows you to preserve an aging family member or loved ones voice?

 

[antiprotest]

Do you still miss that girl that got away?

Put her voice into your new AI robot and she will be with you forever. 😳

What a scary future we are heading towards.

But if you can force her to say all those specific phrases to train the voice, she literally could not get away. And that's the more scary part.

 

[surfzen21]

But if you can force her to say all those specific phrases to train the voice, she literally could not get away. And that's the more scary part.

Currently yes. But as the technology progresses I bet a few voicemails that the person left would be enough to recreate their voice.

 

[jdavid_rp]

I was excited about this feature until I tried it and the quality is really bad. It will improve in the future I'm sure.

Right now if you need to use it as an accessibility feature, it's better just to use one of the included voices, such as the Siri voices.

I think after the initial set up, you can still train it right? Maybe yours needed some extra voice data?

 

[antiprotest]

Currently yes. But as the technology progresses I bet a few voicemails that the person left would be enough to recreate their voice.

Actually I think you can already take someone's voice with a large enough sample. Like, it's easy to get Trump or Obama AI voices. It's just that you cannot do it with Apple's implementation.

 

[iLoveDeveloping]

Do you still miss that girl that got away?

Put her voice into your new AI robot and she will be with you forever. 😳

What a scary future we are heading towards.

Asked my wife to do it on my phone - her iPhone was too old. Now I get her to say all the things she wouldn’t say before! 🤣😉

 

[macfacts]

So gimicky

 

[antiprotest]

I think after the initial set up, you can still train it right? Maybe yours needed some extra voice data?

Perhaps. But I have already listened to a number of YouTuber's examples and they all sound bad -- the voices are only half similar to the humans' voices, robotic when speaking, and the sound quality is also bad, like AM radio, while the included premium voices (such as Siri) all sound more natural and with better sound quality.

I am sure things will improve though. I look forward to the day when I can write an entire speech or podcast, click a button, and broadcast it with my own voice without having to read it myself, and people won't know the difference. If I made a mistake in the script or just want to change something. I can just retype that part and click a button, and it's instantly updated.

 

[antiprotest]

Asked my wife to do it on my phone - her iPhone was too old. Now I get her to say all the things she wouldn’t say before! 🤣😉

You dirty, dirty, clever dog.

 

[dz5b609]

Heartwarming is 1 word for it, creepy is another.

 

[supremedesigner]

I tried it out and it's pretty accurate. What's even more amazing is that this personal voice pronounces a dozen words perfectly. I'm even shocked!

 

[jimbobb24]

I was excited about this feature until I tried it and the quality is really bad. It will improve in the future I'm sure.

Right now if you need to use it as an accessibility feature, it's better just to use one of the included voices, such as the Siri voices.

Exact same. I spent the time and created the voice and its not very smooth. Other AI algorithms have surpassed this but are pretty data intensive. But its a good start especially for people who may lose voice in future.

 

[iLoveDeveloping]

You dirty, dirty, clever dog.

🤣🤫 it’s ok, she had a good laugh about it.

 

[antiprotest]

Does anyone know of a service like this that allows you to preserve an aging family member or loved ones voice?

Free Text to Speech & AI Voice Generator | ElevenLabs

Create the most realistic speech with our AI audio tools in 1000s of voices and 32 languages. Easy to use API's and SDK's. Scalable, secure, and customizable voice solutions tailored for enterprise needs. Pioneering research in Text to Speech and AI Voice Generation.

elevenlabs.io

There are others. Google for more.

 

[jimbobb24]

Modern technology makes hilarious a lot of old futuristic spy and sci-fi stuff. Using your voice as the security layer to your evil lair/military compound is apparently the worst security ever in the future.

 

[winxmac]

Does anyone know of a service like this that allows you to preserve an aging family member or loved ones voice?

Why not record a video of candid or special moments and replay it?

 

[supremedesigner]

Why not record a video of candid or special moments and replay it?

You will need at least 15 minutes + 150 phrases to do that. Otherwise, it won't work.

 

[MrENGLISH]

Do you still miss that girl that got away?

Put her voice into your new AI robot and she will be with you forever. 😳

What a scary future we are heading towards.

“Evelyn…”

 

[sracer]

Sounds like a great way to advance DeepFake technology beyond detection.

 

[ELman]

Why not record a video of candid or special moments and replay it?

It would be nice to have the ability to have the voice say anything you want.
For instance you record a grandparents voice that's gone after a grandchild is born. The grandchild could hear her grandmothers voice saying whatever you want.