Hacker News Reader

> When my girlfriend texts me when I'm out with my AirPods, I think we'd both like me to hear her message in her actual voice rather than Siri's

And then you can use voice to text to text her back, and she can hear it in your voice! It's just like a phone call from 30 years ago, but one that requires infinitely more processing power!

> It's just like a phone call from 30 years ago, but one that requires infinitely more processing power

There are folks who couldn’t, for a variety of reasons, do that thirty years ago. This feature is for them. The rest of us get to e.g. more naturally text a response to a call we’re listening into on a flight.

Can't you already do that with iMessage's voice delivery feature?

> Can't you already do that with iMessage's voice delivery feature?

Voice memos? No. I would have to, at some point, speak it. If you’re referring to text-to-speech, there is a difference between having your speech read in a different voice and your own.

It's just like a phone call from 30 years ago, but one that requires infinitely more processing power!

And latency!

On the plus side it would use less bandwidth. That phone call from 30 years ago probably used (ballpark) 64kilobits/second. This could use a lot less and have higher audio quality.

See it as data compression on the wire.

Genuinely funny reply. But (a) whether I prefer typing or speaking, and (b) whether I prefer reading or hearing, is very context-dependent and might not be the same for the person at the other end -- and so yeah I think having flexibility there is good! If I'm on AirPods and not on my phone, I'd like to hear the message. CarPlay, too. When actively on my phone, I prefer to type, even if on AirPods. CarPlay, I shouldn't probably be typing ever. So yeah, generating text and speech simultaneously and having the end result be situational is in fact a good thing.

It's worth noting this is already how iPhones work and people already love it. What I'm suggesting additionally is substituting Siri's voice for a DIFFERENT customized synthetic voice in a very specific circumstance. I'm not advocating for using synthetic voices where there currently aren't any here.

> It's just like a phone call from 30 years ago

Disagree that it's the same at all. Sending discrete messages at your leisure is quite a different experience than a real-time conversation.

> Personal Voice: When my girlfriend texts me when I'm out with my AirPods, I think we'd both like me to hear her message in her actual voice rather than Siri's. (This feature doesn't allow for that yet, but the pieces are all there.)

Neat idea, same with carplay, having it reliably imitate the voice of the person who sent the message would make it a lot nicer.

Though they would need to get all the TTS and intonation right first, which IME is not the case, I think having the right voice but the wrong intonation entirely would be one hell of an uncanny valley.

When my computer talks to me, it needs to be in Majel Barrett's voice. Anything else is just not acceptable.

Surely Shatner has sold his voice already like Bruce Willis did? I'd pay to hear my morning information read in the voice of the shat, in the format of a captain's log.

A voice changer for unknown or blocked caller id's is actually a great idea. Just looked in F-Droid and the Play Store, where there seems to be no such app. Preventing callers from sampling my voice is a concern I did not know I have.

A couple weeks ago ago, I looked into building it for iPhone, but there was no way in iOS to integrate it. So now I'm just hoping for Apple to do it. This article was what prompted the thought: https://www.washingtonpost.com/technology/2023/03/05/ai-voic...

Doesn't seem to be possible on Android either unfortunately. One might be able to do it by writing a custom dialer - scope creep x100.

>because I've started to get concerned about what can be done, fraud-wise, with even small samples of my voice.

This vector was recently highlighted as a weakness of the Australian "Voiceprint" system(1)

It's also an excellent point to make that these tools can be useful for everyone. I find myself using a number of the accessibility tools simply to speed up some of my common interactions with the phone and watch.

I've also noticed that these technologies end up in other products. For example livetext is now a standard feature on macOS/iOS yet it's an accessibility feature originating in the screenreader to deal with text flattened in images. This technology sharing also gives us a bit of a preview of what they're working on (e.g. AR)

(1) https://www.theguardian.com/technology/2023/mar/16/voice-sys...

I was thinking it would be fun just to have Siri use a Personal Voice, but your idea is better! And it could just offer to make your Personal Voice data available along with your Contact picture.

While that sounds like it would unlock some very cool experiences it also scares me to think about the potential abuses of making personalized voice models fairly easily available. It seems like the sort of thing that would need to stay secure on your own device. It would be great to see some kind of middle ground where a text to speech mechanism would generate audio output and send that, rather than make the model itself available.

> Personal Voice: When my girlfriend texts me when I'm out with my AirPods, I think we'd both like me to hear her message in her actual voice rather than Siri's.

That's genius!

Also a bit creepy.

I think I'd rather just eliminate the ringing portion of a phone call (when I have airpods in) and instead just let a trusted list of contacts talk directly into my ear (1-way) until I "answer" and open up a 2-way channel.

The "answer" button seems to serve to also say "I'm ready to listen" as much as it does to say "I'm ready to say". IMO this kind of goal is better covered by voice messages, or, at least, starting with a voice message. This allows the receiver to pick when they are ready to hear it (including immediately), replay it as needed, and choose when they respond (if at all). Many of these are benefits for the sender as much as the receiver.

Let's have both. If you have it on do-not-disturb or don't pick it, it goes to a voice message with speech-to-text.

Isn't this the walkie talkie feature built into the Apple Watch? Could be ported over to iPhone.

Yeah I want “Apple Intercom” or something where me and my girlfriend can have linked airpods and it’s like a spy movie. Perfect for communicating in loud places or crowds without requiring data.

Not sure if you ever had a Nextel phone, but they had a cellular walkie talkie feature for awhile. It was pretty popular in my high school circa 2003, but I remember kinda hating it.

It's also hard to do privacy-wise unless every text message she sends pre-generates the audio message using her on-device voice and then attaches it. That would make every message use 10x as much bandwidth, storage, and battery power. (10x is a random number but you get the point). Seems cute but really impractical.

I would think that your phone could request audio messages from the sender only when necessary. They already sync things like your DND status to show others so this would just be another flag. Messages could also then alert the sender that their message may be read aloud in their Personal Voice. Or maybe allow turning this on per conversation.

10x compared to what though? FaceTime (and similar) is already full-duplex video and audio, which I have to imagine is at least another 10x on top of what you’re describing. Are we really budgeting our computer resources so strictly that this would even show up as more than a rounding error?

6 billion texts are sent every day in America alone. What you're describing as a "rounding error" amounts to a massive change at that scale.

It would take as much data as a phone call… except you’re only sending enough audio for the message

> Finally, Apple needs to detect and tell me when I'm listening to a synthetic voice.

There's already some version of this in "Hey Siri" detection -- if I record myself with a prompt and play it back, my HomePod briefly wakes up at the "Hey Siri" but turns off mid-prompt. I guessed it was some loudspeaker detection using the microphone array, but it could be a mix of both?

> When my girlfriend texts me when I'm out with my AirPods, I think we'd both like me to hear her message in her actual voice rather than Siri's.

Other consequences of this feature:

- typos and autocorrect become even more hilarious

- you can still use her voice after you break up; heck, you probably don't ever need to date the person

- at this point: why wouldn't you pick Scarlett Johansson?

> you can still use her voice after you break up

Which is potentially quite problematic as you could abuse this for all sorts of social engineering purposes.

I was thinking this exact thought. I want Samantha reading my notifications. I desperately want an AI like her to be able to talk to. ChatGPT is getting there.

That would be great for messaging somewhere quiet where you can’t talk