Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
I disagree. I would love for Apple to integrate voice to text capabilities like Android has, where you can speak on any text input screen. This would be particularly useful while driving. I can look at my phone for 2 seconds to read a text message asking "When will you be home?" or something like that. It's a lot more effort to drive and type even a simple response like "15 mins" (not to mention, downright dangerous). If I could select a little button on the text message screen and say "15 minutes," and have it type it for me. That would definitely be useful for me. I currently use the Google app to do that if I want to google something quickly without typing it, but I would much prefer to use my default browser (iCab). We aren't talking about Star Trek here; just some functionality that would help speed things up a little bit.

I would say you have no right being behind the wheel. Forget about typing - "glancing at my phone for 2 seconds" and you have traveled 200 feet at 70 mph.
 
Could the front camera be used to look at facial gestures / lip movements as a way to make the results more accurate?
Funny you should mention that. Apparently when Arthur C Clarke wrote "2001: A Space Odysey" the only human technology in the book/movie that he thought wouldn't be technologically possible by 2001 was the lipreading (the scene where the two crew go into the pod and switch off the sound so that they can discuss disabling Hal but the computer still "hears" the conversation by reading their lips through the window). Sadly it turned out that pretty much none of the computer technology that he predicted is possible even today and a lot of it is as far beyond us now as it was in the 1960s.

I'd like to have voice recognition so that I could transcribe meeting notes in business meetings.

In the old days when I had a Palm Pilot I was quick enough on Graffiti to be able to take concise notes directly into my Palm during business meetings and it was a huge timesaver because it completely avoided having to type up notes afterwards plus I always had notes from all my previous meetings with everyone I had ever had meetings with right there on my device. I'm not quick enough on the little iPhone keyboard to be able to do this on my iPhone anymore.

Being able to have the iPhone make the notes itself would be great. Unfortunately I don't think that the technology is there to do this yet with any sort of accuracy, even on the most powerful desktop Macs available let alone on an iPhone.

- Julian
 
??? :confused: You already can... from Apple's website for iPod Touch:

Want to hear some music? Just ask.

Ask and you shall hear.
Voice Control knows the music in your iPod touch. Want to hear something specific? All you have to do is ask. For instance, say “Play artist Bob Dylan,” and iPod touch does just that. Ask what song is playing and hear iPod touch answer. Tell it to play your favorite album, artist, or playlist. Speak simple commands such as “shuffle,” “next song,” and “pause.” Even have iPod touch play more songs like the one you’re listening to.​

I was refering to non music apps.
 
Could the front camera be used to look at facial gestures / lip movements as a way to make the results more accurate?

yes, that is actually possible (if I remember correctly, Microsoft has done some publications on that as well). There are mixed results and some groups reported even quite a huge improvement.

The problem with Nuance however is that their current solution for open domain speech recognition (which means free speech, not the limited "Call somebody" recognition) is all done over the internet. For them the integration of Visual information would mean more traffic and since they are mainly interested in getting your voice for improving their product, in addition to the privacy concerns when even pictures are submitted I doubt they would go after that.

To my knowledge there is currently only one company that has its own on-device "nearly" open domain free speech recognition engine, which is integrated into Jibbigo.
But even for them probably the integration of visual features coming from lip-reading would mean too much data and would slow down things.
 
Last edited:
…and people figured out a long time ago that watching videos on the computer appeared cool in movies, but not so cool in reality.

Yet here we are a bunch of years later, Youtube this, stream that, etc. Ridiculous improvements in computer performance, available bandwidth and amount of computers networked changed the game completely.

I would argue that the current state of speech control is flawed not because talking to your computer or device is inherently bad, but because of two things relating to the implementation:

1. Speech recognition accuracy is less than perfect.
2. Computers/devices are too dumb to understand what you intend.

You can control the computer in Star Trek because it rather correctly hears what you say, and in general understands what you MEAN. The second part there is the absolutely hardest one to solve. Still, there both areas are continuously evolving and advancing.

Besides, not even Star Trek uses voice control as the only means of input, there are many many buttons available...


Very good points but I disagree with one thing. I believe the voice technology has grown leaps and bounds over the last couple years. I remember trying to use it on a blackberry through a third party app...It would recognize clear words well but as soon as there was any background noise it would jumble. Since this tech was generally unreliable it became more of a gimmick than a day to day user. With my new Android phone(I know I will be Jeered) The voice tech is simply amazing. You can speak at a normal pace with background noise and it picks up everything very well. Being that texting and driving on a touch screen phone presents a bigger danger than a normal phone I hope this gets explored to the fullest extent. Could be a very cool enhancement considering how much better it is now
 
I'd say so... we figured out a long time ago that (unless you're disabled of course) talking to our computers appeared cool in Star Trek, but not so cool in reality.

Anything thats stop the torture of not being about the use a decent smart phone as i struggle with hand gestures would be great.
Simply scroll up/down expand/contract commands would be great for disabled people like me.
Till then clunky old symbian Nokia phone it is.
 
Very good points but I disagree with one thing. I believe the voice technology has grown leaps and bounds over the last couple years. I remember trying to use it on a blackberry through a third party app...It would recognize clear words well but as soon as there was any background noise it would jumble. Since this tech was generally unreliable it became more of a gimmick than a day to day user. With my new Android phone(I know I will be Jeered) The voice tech is simply amazing. You can speak at a normal pace with background noise and it picks up everything very well. Being that texting and driving on a touch screen phone presents a bigger danger than a normal phone I hope this gets explored to the fullest extent. Could be a very cool enhancement considering how much better it is now

Naturally Speaking 11 Premium try it with a decent mic it is flawless. Only problem is thats the Windows version, the Mac version is still pants but Nuance has only just accrued bought out Mac Speech so hopefully a good version will be available soon.
 
It's a lot more effort to drive and type even a simple response like "15 mins" (not to mention, downright dangerous)

Which is why you never do this, right? People who use cell phnoes while driving, or worse... actually text while driving, should be treated no different than drunk drivers... fines... suspensions... jail time...
 
Could the front camera be used to look at facial gestures / lip movements as a way to make the results more accurate?

I was going to suggest that perhaps they're developing tech to decipher your phone conversations and serve you ads based on keywords.

I don't really believe that, but since Full of Win isn't in a cynical mood today I might as well say it.
 
It's illegal to drive in Canada when using a cell phone or texting. Must use a hands free set. You would not believe how many accidents are caused by drivers on a cell phone.

I like how nuance has their ad at the bottom of the page.
 
HELL YES! Speech-to-text is the one thing I really wish iOS had, especially in regards to texting.

You guys do realise the iPhone already has direct speech-to-soundwave technology built in, which allows the other person to actually hear your voice. It's quite amazing.
 
Really pointless.

The iPhone has voice recognition - it never gets used. Same goes for the Mac. Its not a killer feature and I'd confidently say that over 95% of people never use it. Its a gimmick more than anything.
 
Nuance to supply additional voices for Lion

Geez, so much speculation from such a small piece of (old) news.

Nuance will supply additional voices for Lion. Lots of different international accents will be added. This has been reported on various sites a month or so back.

This negotiation with Apple, is most likely tying the knot for that.

http://www.macnn.com/articles/11/03/02/apple.ramps.up.support.for.disabilities/
 
What I really wanted was an API to be able to use TTS and speech recognition in iPhone/iPad apps but I guess that's still a long time away...
 
Geez, so much speculation from such a small piece of (old) news.

Nuance will supply additional voices for Lion. Lots of different international accents will be added. This has been reported on various sites a month or so back.

This negotiation with Apple, is most likely tying the knot for that.

http://www.macnn.com/articles/11/03/02/apple.ramps.up.support.for.disabilities/

Yeah but for me this would be much more useful on iOS... And especially if developers could build tops on top of it.
 
For me, speech recognition is just one of these "because we can"-things which is nice to have but you won't use on a regular basis. I have speech recognition on my computer and I never use it, because I can type/correct/click a whole lot more faster than talking to my computer. I often change my text, add a sentence here and there and Speech recognition does not really do the job for me there.

Plus, I feel quite uncomfortable talking to my computer, even alone, much less with people around me.

There are a few scenarios for an iPhone to use speech recognition, but let's face it most of the time you don't want to talk to an electronic device.
 
I disagree. I would love for Apple to integrate voice to text capabilities like Android has, where you can speak on any text input screen. This would be particularly useful while driving. I can look at my phone for 2 seconds to read a text message asking "When will you be home?" or something like that. It's a lot more effort to drive and type even a simple response like "15 mins" (not to mention, downright dangerous). If I could select a little button on the text message screen and say "15 minutes," and have it type it for me. That would definitely be useful for me. I currently use the Google app to do that if I want to google something quickly without typing it, but I would much prefer to use my default browser (iCab). We aren't talking about Star Trek here; just some functionality that would help speed things up a little bit.

Or you could just make phone call and not text while driving at all.
 
I use it every day at work to generate anywhere from 25-40 reports. It does a passable job. You have to remember that 95% recognition still spells 1/20 words wrong, and you really have to hunt to find them.
 
I just see a picture in my head of two people on iPhones "talking" to each other, as they are obviously speaking into their phones. However, they are speaking into their iPhones to produce text messages to each other though. I'm sure this day of irony is coming.
 
I use dictation software a lot and always wonder why it's not used more w/ the ios devices.

It'd be so nice to be able to hit a button in any app to dictate. Especially with the ipad where typing lots of stuff out isn't ideal.
 
I think you have two options:

1) Relax. The people waiting for you to arrive should be patient enough for them not to require constant updating.

2) If 1) cannot be achieved, then let them GPS track your every move and then you'll never need to tell them anything.

Or you know :

3) You're holding a god damn phone in your hand. Call the person and talk to them on handsfree by using the already existing voice control of the iPhone and actually talk to them.

Seriously, texting is so dumb. Are people that afraid of actually talking in 2011 ? Now they want to talk out their text messages ? Think about this : Talk out your text message so the device can write it for you, send it to another person who'll what, use the phone's ability to turn the text messages into a synthesized voice ?

People you're holding a god damn phone.

That is why SMS is blocked at the carrier on my phone. I don't want none of this non-sense. You want to tell me something, call me.
 
Or you know :

3) You're holding a god damn phone in your hand. Call the person and talk to them on handsfree by using the already existing voice control of the iPhone and actually talk to them.

Seriously, texting is so dumb. Are people that afraid of actually talking in 2011 ? Now they want to talk out their text messages ? Think about this : Talk out your text message so the device can write it for you, send it to another person who'll what, use the phone's ability to turn the text messages into a synthesized voice ?

People you're holding a god damn phone.

That is why SMS is blocked at the carrier on my phone. I don't want none of this non-sense. You want to tell me something, call me.

Texting is cheaper and more convenient down under.

Calling in NZ (Australian rates are similar but cheaper) can be 44~80c a minute while texting is only 9~20c (This is normal prepay rates without plans or packs) its easy to see why people might not call over texting.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.