Speech recognition

Using our voice to control computers has never really taken off. For many of us, using voice recognition technology wasn’t even a consideration until features like dictation and Siri arrived on our iPhones and iPads. There’s good reason too: the voice recognition features built into our devices have always had the reputation of being half-baked. They simply aren’t accurate and consistent enough to replace our tried and trusted mouse and keyboard or touchscreen. While half decent dictation features come with every Mac (and are powered by Nuance’s technology), the voice recognition features you get with latest version of Nuance’s Dragon Dictate for Mac go well beyond simply dictating speech to text. 
Expand
Expanding
Close

Nuance has just announced the next major version of its Mac dictation software, Dragon Dictate 4. The new version improves recognition accuracy as well as several new features, including the ability to machine transcribe from a pre-recorded audio file. You no longer have to be recording a voice live to get transcription.

[youtube=http://www.youtube.com/watch?v=RRfbZ0h4yBo]

We’ve had a quick play with the software and the accuracy is almost scary good compared to older versions of the software. Speaking in a normal voice gives you almost 100% accuracy and even mumbling seems to work. While Mavericks uses the same Nuance speech engine, Dictate 4 has a better interface for transcription and tons of extra features as outlined in the videos above and below.

Dragon Dictate software has also been updated for modern system architectures — it is a 64-bit app now.  This results in better performance and better memory management over its predecessors. The company says it has drastically reduced latency when interpreting speech.

[youtube=http://www.youtube.com/watch?v=YnGKb854CNI]

Expand
Expanding
Close

[youtube=http://www.youtube.com/watch?feature=player_embedded&v=e54gK0UpAxs#!]

The Google Now-like voice recognition features introduced to the Google Search for iOS app last month have been getting a lot of attention. Maps might not be the only area Apple cannot compete with Google. As highlighted in a video comparison we posted earlier this week, Google’s voice search feature has Siri beat. And many users are taking note. Today, AppAdvice pointed us to a new Jailbreak tweak called “NowNow” that allows you to call up Google voice search (from within the Google Search iOS app) by holding down the Home button—just like Siri. It appears you can also select alternate methods to invoke the feature, so you should be able to continue to use both if want. The tweak is available via the BigBoss repository now for free.

Expand
Expanding
Close

[youtube=http://www.youtube.com/watch?feature=player_embedded&v=4amvRLG4un8]

There are fewer options for speech recognition these days, and now there are even fewer with Nuance announcing they acquired Vlingo for an undisclosed figure. Following multiple lawsuits related to patent infringement, the two companies apparently came to what CEO of Vlingo Dave Grannan called  “a good outcome.” Grannan elaborated in a prepared statement (via AllThingsD):

Vlingo and Nuance have long shared a similar vision for the power and global proliferation of mobile voice and language understanding. As a result of our complementary research and development efforts, our companies are stronger together than alone. Our combined resources afford us the opportunity to better compete, and offer a powerful proposition to customers, partners and developers.

Vlingo is notably used in various voice-controlled Android apps, and it is viewed as competitors to Apple’s Siri built into the iPhone 4S. However, Siri, also used it prior to it being used by Apple, before switching to Nuance…

In an interview with 9to5Mac, Siri co-founder Norman Winarsky said Vlingo was originally used as the speech recognition component of Siri before switching to Nuance. He noted: “Theoretically, if a better speech recognition comes along (or Apple buys one), they could likely replace Nuance without too much trouble. ” The full quote is below.

9to5Mac: How important is Nuance speech recognition to the Siri technology?
Expand
Expanding
Close

.

Curious to see how Windows Phone feels, but don’t have a device around to do so? Us neither. But Microsoft has just released a new HTML 5 website that allows iPhone and Android users to get a taste of the Windows Phone 7 (Mango 7.5) operating system. The trial requires no downloads or registration, and you can try it out by just going to the webpage http://aka.ms/wpdemo on your mobile device.

While it doesn’t use any of the data on your phone like your contacts, the demo does give you a pretty comprehensive look at all of Windows Phone’s features. Microsoft uses a blue dot to guide you around the operating system, and obviously some features like voice recognition just don’t work in the browser. Drat, that’s something we really wanted to try.

Will this draw users over to the Windows Phone platform? Probably not, but it’s always cool to see what the competition is up to.

Expand
Expanding
Close

[youtube http://www.youtube.com/watch?v=MpjpVAB06O4]

Apple’s purchase of Siri in early 2010 and their partnership with Nuance in 2011 has many hoping that Apple has something like speech-to-text or voice-navigation up its sleeve for iOS 5. One of the remaining advantages of Android over iOS is its system-wide Voice Actions technology.

Unfortunately, WWDC and the iOS 5 announcement came and went and nothing related to voice-navigation had been announced. Even so, the Apple-Nuance partnership has been confirmed by way of Nuance voices in Apple’s OS X Lion and Nuance speech-to-text functionality that is referenced in Apple’s internal settings modules.

But that doesn’t mean Apple isn’t hard at work at this very moment trying to cram some native OS level voice recognition technology into iOS 5 before launch.

Coupled with Nuance speech-to-text, Apple appears to be planning to take the fruit of their Siri purchase and fully integrate it into this fall’s release of iOS 5. Because these new features have yet to appear in iOS 5 on the iPhone 4 or iPhone 3GS, Apple might be saving these new features as an iPhone 5 exclusive. This would be akin to Apple’s decision to make Voice Control and video recording exclusive features to the iPhone 3GS, even though they could technically function on earlier models of the iPhone. As you can see in Siri’s promotional video above, the company advertises itself as “your virtual personal assistant.”

As you can see in the screenshot above from an Apple iPhone test unit, Apple is currently developing and testing a new iOS feature called “Assistant.” This screenshot, from a reliable source, is corroborated with our own SDK findings (below). The source did warn, though, that development is not yet completed – and just went into testing – and may not even be finished by the time the next iPhone ships.

More info after the break…

Expand
Expanding
Close