Apple's New Data Center Focus of Nuance Voice Recognition Partnership?
Late last week, we reported on claims that Apple is in discussions with voice recognition firm Nuance Communications regarding some sort of partnership presumably linked to rumors that Apple is integrating significant voice capabilities into iOS 5.
TechCrunch now follows up to report that the deal seems to revolve around Apple utilizing Nuance's technology in its new North Carolina data center to drive centrally-hosted voice services. The partnership is said to likely be introduced at Apple's Worldwide Developers Conference (WWDC) early next month.
In digging into the information about the relationship between the two companies, we had heard that Apple might actually already be using Nuance technology in their new (but yet to be officially opened) massive data center in North Carolina. Since then, we've gotten multiple independent confirmations that this is indeed the case. And yes, this is said to be the keystone of a partnership that Apple is likely to announce with Nuance at WWDC next month.
More specifically, we're hearing that Apple is running Nuance software - and possibly some of their hardware - in this new data center. Why? A few reasons. First, Apple will be able to process this voice information for iOS users faster. Second, it will prevent this data from going through third-party servers. And third, by running it on their own stack, Apple can build on top of the technology, and improve upon it as they see fit.
Top Rated Comments
(View all)
voice recognition is so useless and dumb, can u imagine all those ppl talking to their gadget on the tram
Short sighted people are always amusing to point at when everybody's adopted the new technology they said was useless and dumb just months prior to it becoming a hit.
Voice recognition hasn't taken off for a couple of reasons:
1 - It's not quite accurate enough so you have to speak at an unnatural speed and tone.
2 - Current technology is mostly one way so you still have to view the screen to communicate properly because you can't yet trust that the system will be correct within an acceptable ratio.
Apple's acquisition of Siri solves both of these.
Imagine having a conversation with your iPhone without thinking in pre-determined commands and being confident that it understands you and will act appropriately.
Instead of:
- Click and hold a button
- Wait for prompt
- "Play Song 'Beat it' "
- Verify that it got it right
- Begin next command...
You'll be able to do this:
"iPhone, play 'Beat it', make an appointment for lunch with Wendy on Tuesday at noon. Oh, also I need to get groceries today. I need potatoes, pasta, steaks and apples and send this to my wife too."
iPhone replies: Playing 'Beat It' by Michael Jackson. I've created an iCal appointment for 12pm this Tuesday for 'Lunch with Wendy'. I've set up a Task named "Groceries" and listed "potatoes, pasta, steaks and apples" and have shared it with Debbie.
This kind of real time interpretation requires a lot of power and constant refinement not possible in a battery powered mobile device. Instead, a small file containing a mono audio snipit can be sent to Apple's servers, interpreted and commands sent back within a few seconds, making this possible.
Look to iOS5 as being the beginning of a true personal assistant that you can speak to but that is also able to speak with you in a natural conversation.
Apple was indeed years ahead of the competition with the introduction of the iPhone and I doubt they've been sitting on their hands waiting for Android to catch up.
They've been working on the next killer feature that will once again push the boundaries and propel Apple way ahead of the pack. This type of conversational speech recognition definitely has the ability to do this.
Spock: "Computer, structural analysis, and make up of object to port of Enterprise".
Computer: "Computing".
Computer: "Object appears to contain early 21st century remains of Humanoids."
"There appears to be a digital recording encoded on silicone." "Mossberg, our friend, is no longer writing good things about us."
Spock: Looks at Kirk, and raises 1 eyebrow. :apple:
Yeah, I couldn't even imagine people talking to their PHONE in public.... :rolleyes:
:D
Hopefully my phone will only respond to my voice, or if I walk past ipedro on the street my phone might start playing Michael Jackson's Beat It.
Uh, are YOU going to give me 2TB of cloud storage? I'm at 1TB, now, haven't got all the video and photos on disk, yet.
Ok, I'll have to stop using the current vr, then. Thanks for the tip. I'll also tell my coworker to delete Dragon from his phone, obviously it couldn't actually be working for him. :rolleyes:
Dragon works over the network. The speech recognition engine is not running on the iPhone. It is just sending audio to the server for processing.
The Voice Control built into iOS is similar to what is running on the Mac. It is not very sophisticated and can't handle dictation or the kinds of things being discussed here.
Nuance is big software - in terms of disk space required, CPU utilization, and memory requirements. Ask anyone who has integrated it with telephony systems (I have). You wouldn't want this running on the iPhone. It would have to be running at all times to be responsive. It would use a lot if not all of the iPhone's resources and it would kill the battery.
Speech recognition is not some web 2.0 mashup that some aspiring hacker can whip up in a few lines of PHP code. It is many decades and billions of dollars of research. It is one of the hardest problems in computer science and Nuance is practically the only player left in the game.
(too soon?) :rolleyes:
Short sighted people are always amusing to point at when everybody's adopted the new technology they said was useless and dumb just months prior to it becoming a hit.
Voice recognition hasn't taken off for a couple of reasons:
1 - It's not quite accurate enough so you have to speak at an unnatural speed and tone.
2 - Current technology is mostly one way so you still have to view the screen to communicate properly because you can't yet trust that the system will be correct within an acceptable ratio.
Apple's acquisition of Siri solves both of these.
Imagine having a conversation with your iPhone without thinking in pre-determined commands and being confident that it understands you and will act appropriately.
Instead of:
- Click and hold a button
- Wait for prompt
- "Play Song 'Beat it' "
- Verify that it got it right
- Begin next command...
You'll be able to do this:
"iPhone, play 'Beat it', make an appointment for lunch with Wendy on Tuesday at noon. Oh, also I need to get groceries today. I need potatoes, pasta, steaks and apples and send this to my wife too."
iPhone replies: Playing 'Beat It' by Michael Jackson. I've created an iCal appointment for 12pm this Tuesday for 'Lunch with Wendy'. I've set up a Task named "Groceries" and listed "potatoes, pasta, steaks and apples" and have shared it with Debbie.
This kind of real time interpretation requires a lot of power and constant refinement not possible in a battery powered mobile device. Instead, a small file containing a mono audio snipit can be sent to Apple's servers, interpreted and commands sent back within a few seconds, making this possible.
Look to iOS5 as being the beginning of a true personal assistant that you can speak to but that is also able to speak with you in a natural conversation.
Apple was indeed years ahead of the competition with the introduction of the iPhone and I doubt they've been sitting on their hands waiting for Android to catch up.
They've been working on the next killer feature that will once again push the boundaries and propel Apple way ahead of the pack. This type of conversational speech recognition definitely has the ability to do this.
100% right
[ Read All Comments ]
Even as Apple is preparing to open its first Dutch retail store in Amsterdam on March 3, the company is moving closer to expanding its international reach even further as it has updated its Swedish...
Following reports yesterday that Apple would open its first Dutch retail store in Amsterdam on March 3, the company has confirmed that date today with emails to customers and a new dedicated store...
The Charlotte Observer reports that a total of 25 iPhones valued at over $16,000 have been reported stolen from Apple's Northlake Mall retail store in Charlotte, North Carolina. While...
German site iFun.de reports [Google translation] that has it has received information from a source "to be taken seriously" claiming that the iPad 3 will debut there on Friday, March 23.
...
AFP reports that Proview Technology has expressed a willingness to work toward a settlement with Apple in the "iPad" trademark dispute in China, even as the company continues to press forward...
