Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
so i sent my brother a text today telling him about this. he says his bionic already does everything this is going to do and tells me apple is so far behind.

i did a google search and all that popped up were tons of posts talking about how text to speech (car dock) does not work at all on the bionic. can anyone confirm or deny? what is android offering right now?
 
I'm not really that interested in having a conversation with my phone. Seems more of a novelty feature to me.

The system will actually speak back and forth with the user to gain the most information in order to provide the best results.


In other words... it'll take 45 minutes just to order a pizza or make a google search. (And it'll be wrong half the time).

Sorry folks... but "useful" speech recognition is a long ways off.
 
so i sent my brother a text today telling him about this. he says his bionic already does everything this is going to do and tells me apple is so far behind.

i did a google search and all that popped up were tons of posts talking about how text to speech (car dock) does not work at all on the bionic. can anyone confirm or deny? what is android offering right now?

I'm not sure, but the bionic has the cool feature where it squeals at the user for no apparent reason. The iPhone can't do that.
 
The system will actually speak back and forth with the user to gain the most information in order to provide the best results.


In other words... it'll take 45 minutes just to order a pizza or make a google search. (And it'll be wrong half the time).

Sorry folks... but "useful" speech recognition is a long ways off.

i think the point right now is to keep peoples eyes on the road. as far as it taking 45 minutes to do a google search, im able to do that now on pc or iphone in less than 10 secs. and its pretty damn accurate.

i cant see apple slinging out a half ass product when it comes to their bread and butter os
 
That's one way to prevent hardware leaks.

Would it be possible to announce an iPhone with new hardware and a bigger screen on October 4th... and then have enough built for a launch 2 weeks later?

Could that be why we haven't seen any credible leaks of a newly designed body? :)

Or will the next iPhone simply be an updated iPhone 4 ? :(

I read somewhere that for FCC testing, the hardware was in locked boxes (as the design isn't required for approval). After the iPhone 4 leaks, I'd think they would be more careful this time around, so I thin a re-design is still possible.
There's only been 1 time they kept the same design. I don't think that sets a precedent.

They can produce millions of phones in a run, but two weeks is pushing it. If the phones will all be combo GSM/CDMA, that makes it much faster (and Apple doesn't brand phones with carrier crap on them which speeds things up too. Just the bar code on the box dictates which carrier it belongs to.)

Apple likes the media circus and people waiting in lines and stalking shipments at stores, so i doubt their concern is having 20 million units for launch day, or even 10 million. If that were a concern, they would ramp up production at least 2-3 weeks before announcing a release to ensure supply.

This will keep people actively running the the Apple store into the holiday season (which is very smart on their end), and then that piggy backs the Christmas traffic in the stores. Moving the iPhone release back a few months is brilliant. After the holiday spending dies, we get the new iPads.

Apple has really found a way to drive traffic in times of the year when shoppers are staying home. Brilliant.
 
Why would this require an A5 and 1GB of memory? Nuance's speech services are almost completely driven through their servers. Yes there will be some code that has to run locally, but the heavy lifting is all done remotely. Demo Dragon Go or Siri to see how Nuance's technology works."

If I stuffed pork chops with diced apples, would you try to argue it's an apple pie with a pork crust? You'd recognize it as something totally different with 1 common ingredient.

That's what this comes out to be.

The only thing this has in common with any software you mentioned is Speech recognition services done by a remote server.

Siri (now owned by Apple, and it was never a Nuance owned product FYI) was never integrated into IOS or your apps. It was basically just a web search tool with some tweaks (like it could connect to a database for taxi cabs.) the only thing it used was your GPS when locating things for you, which is basically did by searching the online yellow pages. What made it neat was the interface gave it that look and feel of interacting with you, but it really was basically googling via voice commands for all intent and purpose. Nuance speech to txt app basically just sent you a text back of what you said and was glorified text messaging.

The Assistant app appears to have access to your apps and is deeply integrated into IOS. If it's accessing your calendar, address book, to do lists (and other new apps/features coming in IOS5), it's doing some real time multitasking and it would stand to reason it would need more memory and juice to do it. The iPhone doesn't really do multitasking even though we like to think of it doing it. It really just puts apps in a suspended state unless it's your iPod or a music app of some kind.

I would bet money they will allow developers to integrate the functionality into their own apps as well or it would be a bit of a crippled feature for some people. (And no, they didn't need to let developers have it early. Just like Air Play, third party came later.) So in the example in the article, it mentions a calendar event is made within the Assistant app with a animation showing your calendar updating. So somehwere a background process is running (thus multitasking). Now imagine Facebook adding the ability to use assistant to create a Facebook post. Now somewhere in the background Facebook is using your internet connection, creating a post while the Assisant runs in the app, and the Nuance integration is transcribing your speech over the internet, and then posting it. In some cases it would be conceivable the feature might need acces to multiple apps at one time. (If say the assistant could upload a photo and post to Facebook). That's a lot going on.

Don't forget in the mix of all this you have the phone doing standard background process like searching for signals and wi-fi, running through push notifications, driving the display and graphics, etc.

I see people using apps on a 3GS that run slower and not nearly as well as they do on my iPhone 4. I'll be bummed out if the iPhone 4 can't support the feature, but I'd rather not have a feature and have good performance on my device that have a feature that boggles things down to the point I want to throw my device at the wall. Like I do with my iMac since I upgraded to Lion.
 
Newton redux?

Let's hope that this is so good that Gary Trudeau can do nothing with it!! :)

I agree with the early poster who wrote that this could be BIG!

Imagine voice recognition so accurate that you do all your Google searches via voice--no typing.

Imagine it being so accurate that you could actually type your essays, e-mails, letters by speaking.

Imagine... how John Sculley must feel? Vindicated! They should send him a complimentary device.
 
I'm not really that interested in having a conversation with my phone. Seems more of a novelty feature to me.

Drunk and needing a cab, I found Siri very helpful. Or you're driving and can't remember the time a movie starter, perfect. If it keeps what was good about Siri as well, this would be my favorite feature.

I can think of lots of uses. You get a text message about an impromptu meeting on weds and you're driving and don't want to risk forgetting. Click, speak, done.

You are about to run to the grocery store and want to add four things to a list real quick. I never use digital lists for these things because I can write it down faster, but if I could click and speak, "butter, american cheese, and vodka" and it actually work I'd be very happy.

The problem I find (even with Nuance's app) is that it never seems to get that you said, "butter, american cheese, and vodka." You get something goofy like "mother, asian cockroaches, and voodoo."

Apples current voice feature sucks, and not in the good way. It works like 20% of the time. By the time you cancel and start over 3 times, you could of just done it manually and with less chance of wrecking the car.
 
There are some standard messages I'd love to send with a few voice commands such as

- I'm on my way home.

- Do we need anything?

- I'm running late.
 
I'm sure their will be a way to access the new features that aren't available on the iPhone 4 by jailbreaking.
 
While this is great and all, but it just adds to the fire that all Apple is doing with this release is playing some serious catchup. First the Notification Drawer (taken directly from Android), then Quick Camera (taken from Windows Phone 7), and now this - which Android has had for at the very least a year now. My father uses this daily and it comes in incredibly handy for him (he hates touchscreen keyboards), so I can see the appeal, but it's hardly a big feature.

What I am curious about is why the strict hardware requirements? Android offers this feature to bottom of the barrel phones, and yet a processor as advanced as the A4 can't handle this? Hell, a 3GS should be able to do this no problem. Sounds like an artificial limitation to me, just like multitasking on the 3G.
 
If I stuffed pork chops with diced apples, would you try to argue it's an apple pie with a pork crust? You'd recognize it as something totally different with 1 common ingredient.
No but if you cooked a pie using an apple pie recipe I might be tempted to think you baked an Apple pie.

That's what this comes out to be.

The only thing this has in common with any software you mentioned is Speech recognition services done by a remote server.

Siri (now owned by Apple, and it was never a Nuance owned product FYI) was never integrated into IOS or your apps. It was basically just a web search tool with some tweaks (like it could connect to a database for taxi cabs.) the only thing it used was your GPS when locating things for you, which is basically did by searching the online yellow pages. What made it neat was the interface gave it that look and feel of interacting with you, but it really was basically googling via voice commands for all intent and purpose. Nuance speech to txt app basically just sent you a text back of what you said and was glorified text messaging.
FYI I never claimed siri was owned by Nuance. FYI siri used Nuance technology. FYI Apple is licensing that same technology from Nuance. The only thing this has incommon is it is using Nuance technology.

As for your description of Nuance's app, I am guessing you have never user Dragon Go. It does far more than "basically just sent you a text back of what you said and was glorified text messaging". There is a reason Apple is going to use Nuance tech to support all of this.

The Assistant app appears to have access to your apps and is deeply integrated into IOS. If it's accessing your calendar, address book, to do lists (and other new apps/features coming in IOS5), it's doing some real time multitasking and it would stand to reason it would need more memory and juice to do it. The iPhone doesn't really do multitasking even though we like to think of it doing it. It really just puts apps in a suspended state unless it's your iPod or a music app of some kind.
The iPhone absolutely does multitasking. It is UNIX FFS. Userland apps are limited to single tasking, for good reason. An OS process, such as Assistant, would already be able to run in the BG. Just like phone.app, mail.app. Lots of OS level tasks are running in the background at all times, many with RT access to your contacts, calendar, etc.


The assistant running in the BG itself would consume more resources. Being location aware, at all time, would require more resources. The real processor intensive stuff will be the voice recognition. Language analysis, intent analysis, pattern analysis etc is all very CPU heavy. That's why the Nuance API, which Apple is going to use, offloads the work to the network.

I would bet money they will allow developers to integrate the functionality into their own apps as well or it would be a bit of a crippled feature for some people. (And no, they didn't need to let developers have it early. Just like Air Play, third party came later.) So in the example in the article, it mentions a calendar event is made within the Assistant app with a animation showing your calendar updating. So somehwere a background process is running (thus multitasking). Now imagine Facebook adding the ability to use assistant to create a Facebook post. Now somewhere in the background Facebook is using your internet connection, creating a post while the Assisant runs in the app, and the Nuance integration is transcribing your speech over the internet, and then posting it. In some cases it would be conceivable the feature might need acces to multiple apps at one time. (If say the assistant could upload a photo and post to Facebook). That's a lot going on.
Good bet on developers being given access. None of that sounds particularly processor intensive. If they are gong to be doing much of the voice analysis locally, that would require lots of CPU and memory. That is exactly why Nuance offloads it to the network. Orders of magnitude more compute power in a data centre than on device. But, that also introduces delay, which may be enough for Apple to do much of it locally.

And that's really my question. How much are they going to try to do locally that would require an A5 over an A4. The A5 will be a big improvement over the A4, no doubt, but the level of voice recognition they are going to be throwing at it would be best served from a data centre for accuracy.
 
While this is great and all, but it just adds to the fire that all Apple is doing with this release is playing some serious catchup. First the Notification Drawer (taken directly from Android), then Quick Camera (taken from Windows Phone 7), and now this - which Android has had for at the very least a year now. My father uses this daily and it comes in incredibly handy for him (he hates touchscreen keyboards), so I can see the appeal, but it's hardly a big feature.

What I am curious about is why the strict hardware requirements? Android offers this feature to bottom of the barrel phones, and yet a processor as advanced as the A4 can't handle this? Hell, a 3GS should be able to do this no problem. Sounds like an artificial limitation to me, just like multitasking on the 3G.

i agree, but you know how finicky apple is with battery life and processor degradation (flash). even though the a4 was reported to degrade faster than any other chipset when it was first released. it doesnt surprise me that with as late as they are to the game... they would want to act like we needed this configuration.
 
Sorry if this was posted before...but once they release the GM, unless this is an iPhone 4S/5 (whatever it is) and iPad 3 specific feature, this should be clear...right?
 
For the first time the rumors are starting to make sense. This is good old Apple delighting customers through "magical" integration of hardware and software, if the article is to believed.

I'm looking at it this way: Apple wants us to buy more handsets. An A5 speed bump simply isn't enough. A new form factor and/or NFC and/or 64GB and/or better camera might be enough for die-hards like me, but not for the majority of iPhone 4 owners.

"Assistant" would do the trick. They can call it "revolutionary" without getting laughed at, and it's understandable that it would require the extra RAM And CPU if it does everything described.

And there are some bonuses in the article that would be really nifty: Wolfram Alpha integration? Find my friends? Sign me up. I hope the rumors in the article pan out.
 
People still sound weird when I very occasionally see them walking around talking through a bluetooth headset. They are very rarely used, with the main focus for use while driving.

You must not live in California, I see a lot of people try to look important with bluetooth headsets. I still it think it looks stupid but if the technology is there, people will walk around talking to themselves. I think it will be cool for some things when I'm driving or by myself.:cool::apple:
 
I agree 110%. I have never gotten voice control on my 3Gs to accurately execute a command.

I say "Play Green Day".
It says "Calling Jason".

Doh!

I have found part of it is user errors. The screen tells you what to say (and I'm guilty of not paying attention, but it does scroll a bit fast.) You also have to wait for that damned beep which seem inconsistent with how fast it beeps for you to talk.

But music you can't say "play Green Day" You have to say, "play artist Green Day." Or "Play song Blvd. Of Broken Dreams."

I also learned that saying "call john doe" for some reason gets taken as something else. BUT you can say "Dial John Doe" instead, and using the word dial seems to have better results.

PS. I learned this today when writing another post, and wanted to make sure I didn't say something in error. I was going to say it never works, and damned if it didn't start working.

Anytime I've tried to show the feature to someone, the phone just does whatever it feels like

----------

It's not a novelty, it's magical.

But wait... there's one more thing!

The iPhone 4ME (Magic Edition) with a host of great new features with Apple's fool proof speech recognition technology.

Coming soon, the iPsychic app. Why waste time on magic 8 ball apps? With voice options such as Miss Cleo or Dion Warwick you can get a reading from the stars. Interact like never before.


Just intime for the holidays, the iVooDooDoll. Powered by the Assistant app, you customize your dolls with pictures from the camera roll, and your iPhone Magic edition will inflict your virutal voo doo doll with anything you like, and then post your custom made voo doo curse on the victim's Face Book chage all with the sound of your voice! Responds to over 50 commands. "STAB eyes... No, no, I staid, STAB eyes. Grrr. OKay, CHOKE her. CHOKE her..." (not recommended for use in public subways or near insane asylyumns.)

Currently doesn't support witch craft, spell casting, necromancy, or the iPhone 3GS and lower. Not responsible for curses cast from the phone on people, family pets, or the republican presidential hopefuls.

----------

Sorry if this was posted before...but once they release the GM, unless this is an iPhone 4S/5 (whatever it is) and iPad 3 specific feature, this should be clear...right?

Actually no. Apple has in the past launched an update (Was it IOS 3?) and announced features coming in a point update "this fall" or "this winter" or "by the end of the year." MMS was one I think, maybe folders? You can google it yourself, I'm too lazy, but it's been done before.

Developer testing probably was not needed with the feature to launch it. That would be important if they give access to third party apps. The rumors were contract issues with Nuance, so it could be a last minute addition. We'll see though.

But in any case, the GM isn't going live before the media event so in that way, yeah, it would answer things.
 
gimme a 4 inch screen and this awesome personal assistant feature and it's bye bye iPhone 4!
 
Why would this require an A5 and 1GB of memory? Nuance's speech services are almost completely driven through their servers. Yes there will be some code that has to run locally, but the heavy lifting is all done remotely.

I think you're correct. It's likely the iPhone 4 will be able to do this, but it may be one of those things where it will be such a hog that the user may have to dial back how many apps, docs or songs they store on their phone as well as cause mild degraded performance.

The iPhone doesn't really do multitasking even though we like to think of it doing it. It really just puts apps in a suspended state unless it's your iPod or a music app of some kind.

While the iPhone does do multitasking with music currently, keep in mind that anything dealing with voice could use that channel because you would shut down music to talk or listen anyway.

I would bet money they will allow developers to integrate the functionality into their own apps as well or it would be a bit of a crippled feature for some people.

I think you'd win that bet. It is likely that some developers have been invited to show their integration at the iOS 5 intro.

-------------------------------------------

While this is great and all, but it just adds to the fire that all Apple is doing with this release is playing some serious catchup. First the Notification Drawer (taken directly from Android), then Quick Camera (taken from Windows Phone 7), and now this - which Android has had for at the very least a year now. My father uses this daily and it comes in incredibly handy for him (he hates touchscreen keyboards), so I can see the appeal, but it's hardly a big feature.

Apple is always jumping ahead and then playing catch up because they have ONE iPhone release a year. This gives the other 40 players a year of product releases to pull ahead. Give Apple credit. The competition still hasn't caught up with things Apple released 16 months ago.

What I am curious about is why the strict hardware requirements? Android offers this feature to bottom of the barrel phones, and yet a processor as advanced as the A4 can't handle this? Hell, a 3GS should be able to do this no problem. Sounds like an artificial limitation to me, just like multitasking on the 3G.

As yet we don't know that the iPhone 4 can't do the assistant thing...it's only a rumor, not a fact. Jezz!
 
Doesn't seem like a key feature to me or one I will use. For me, a key feature would be new iPhone 'available now' or 'in stock' ;)
 
I'm not really that interested in having a conversation with my phone. Seems more of a novelty feature to me.

Maybe it's just me, but I can talk quicker than I can type on my phone, and I can't walk and type on my phone at all, so usually have to stop if I want to type something meaningful. This would be awesome for me.

Weather it's a feature worth upgrading my iPhone 4 for is another matter.
 
Great. Yet another "article" that refers to a rumour from another site as though it's fact.

Let me remind you of the fact that this blog is called: "macRUMORS". Basically all articles shown on this blog should be considered a rumor unless proven otherwise. Spreading rumors is their main conceirn. It's what they do!
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.