VoiceRecognition
Latest
Nuance buys SVOX ahead of iOS 5 release
There's a whole trail of rumors hinting at an upcoming deal between speech recognition company Nuance and Apple. For quite a while now (ever since Apple picked up personal assistant software maker Siri), the scuttlebuzz has claimed that the folks in Cupertino would make a deal with Nuance for some kind of speech recognition, most likely an iOS-level integration that would allow you to ask your iOS device for whatever you want, and get it quickly and easily. But even if that deal is on, that hasn't stopped Nuance from slowing down. The company has acquired another speech recognition firm, SVOX, the creators of high-end speech recognition and text-to-speech services. That's a natural fit for Nuance, of course, and the release says that the new deal "will advance the proliferation of voice in the automotive market, and accelerate the development of new voice capabilities that enable natural, conversational interactions between consumers and their connected cars, mobile phones, and other consumer devices." Sounds exciting to us. We didn't actually get to see either Siri or an updated voice control service show up during the iOS 5 announcement at WWDC, but that doesn't mean it's completely out of the cards yet. Maybe a deal like this is just what Nuance needs to set up the partnership that Apple's reportedly been seeking for a while.
Weekend rumor roundup: Apple Retail event, new MacBook Airs, unlocked iPhones, more
Several rumors with varying degrees of credibility came up over the weekend. According to AppleInsider, Twitter user @chronicwire (reportedly a source of past Apple leaks) reports that Apple's retail stores are setting up to launch Apple's annual Back to School promotion on Wednesday. The same source initially reported that the Back to School promo will coincide with the launch of new MacBook Airs, but he has since retracted that claim. Instead, Chronic claims the part numbers he initially thought represented new MacBook Airs indicate that Apple will start selling versions of the GSM iPhone 4 that are not carrier-locked to AT&T. Although the MacBook Air is widely expected to have a refresh soon, this is the first we've heard of unlocked iPhones being offered for sale in the U.S., and it's something we'll file under "We'll believe it when we see it." The iPhone is already sold free and clear of carrier locks in several markets, but GSM model iPhones sold in the U.S. remain carrier-locked to AT&T unless you jailbreak. Chronic has also released screenshots that supposedly come from an "internal build" of iOS 5. These screenshots show that Nuance voice recognition, expected to be integrated in iOS 5 but not discussed at WWDC, is still in development. Other sources have claimed these voice recognition features weren't ready to be shown off at WWDC but should be good to go by the time iOS 5 launches this fall. Finally, a reader has informed us that New Zealand's online Apple Store is now showing shipping times of 5-7 business days for the 1 TB Time Capsule and 1-2 weeks for the 2 TB model. These extended shipping times are also showing up in Apple's Australian and UK stores, and the Canadian Apple Store is showing a 1-2 week delay for the 1 TB Time Capsule. The U.S. store and most international stores are not showing the same delay, but they're further indicative of the Time Capsule supply constraints we reported last week, which may mean a product refresh is imminent. We'll be keeping a very close eye on Apple's online store on the Tuesday overnight shift, and we'll let you know if anything new comes up.
Russian ATM uses voice analysis to tell when you're lying
Credit card applications via automated teller are all the rage abroad these days. That's why Russia's Sberbank is using Speech Technology Center's voice recognition system in its new ATM to tell when you fudge your financials to get approved. Like a polygraph, the technology senses involuntary stress cues to ferret out fib-filled statements -- only instead of using wired sensors, it listens to your angst-ridden voice. Designed using samples from Russian police interrogation recordings where subjects were found to be lying, the system is able to detect the changes in speech patterns when a person isn't telling the truth. Of course, it's not completely accurate, so the biometric voice data is combined with credit history and other info before the ATM can crush an applicant's credit dreams. And to assuage the public's privacy concerns, patrons' voice prints will be kept on chips in their credit cards instead of a bank database. So, we don't have to worry about hackers stealing our biometric info, but we're slightly concerned that we'll no longer be able to deceive our robot overlords should the need arise.
Kinect integration in Ghost Recon: Future Soldier, hands-off (video)
Microsoft's E3 keynote may have exploded with deeper Kinect support, but nothing caught our eyes quite as sharply as Ghost Recon: Future Soldier's rifle-exploding Gunsmith demo. A Ubisoft representative showed us how it's done: separating your arms separates your deadly firearm into a gorgeous display of floating screws, components, and accessories, which can be effortlessly modified, swapped, and replaced with gesture and voice commands. Too picky to decide for yourself? Then don't: just tell Gunsmith what you're looking for. For instance, saying "Optimize for range" produces a weapon any sniper should be proud of -- even better, we found that commanding Gunsmith to "optimize for awesome" birthed a rifle (pictured above) sporting an underbarrel shotgun attachment. A gun attached to a gun? Yeah, that works. Weapons can be tested in Gunsmith's gesture-controlled firing range, an engaging shooting mode exclusive to the Gunsmith weapon editor and not usable in regular gameplay. Head past the break for a hands-on (figuratively speaking) video.
Chaufr lets you shout searches, yell URLs at Chrome
Generally, shouting commands at the internet isn't going to get you very far but, if you're just yelling a few destinations and search terms, Chrome extension Chaufr can take you where you need to go. A previous add-on, Speechify, let you speak to fill input fields, but couldn't help you actually navigate the web. Chaufr, on the other hand, lets you simply say the magic word -- "Engadget" -- and it drops you right at our online doorstep. You can also use it to perform searches by saying Wikipedia, Google, Amazon, YouTube, or Yahoo followed by whatever it is you're looking for. It worked well enough in our brief hands-on, but we do have one nit to pick -- activating voice input requires you click on an icon in the tool bar then click on a microphone in the drop down menu. (Can't a brother get a keyboard shortcut?) You can try it out for yourself by clicking on the source link.
Garmin nuLink! 2390 torn apart by FCC, put back together again on US site
Last week Garmin announced the latest member of its high-end GPS navigator family, the nuLink! 2390. Sadly, it was a Europe only affair, leaving American consumers wondering why the company was giving us the cold shoulder. (Whatever it was baby, we're sorry, come back.) Then we spotted an unnamed 4.3-inch Garmin making its way through the FCC that matches up quite nicely, size- and feature-wise, with the 2390. The newest nuLink-enabled device is even showing its face over at the company's US website (you really do love us!), though it's not available to order and you'll have to do some serious digging to unearth it. Whenever it does hit American shores you'll be able to pull in 3D traffic data and search Google thanks to its GSM radio and tether your phone to it using Bluetooth for hands-free calls. It also has voice recognition software so you can furiously bark commands at it when you miss a turn and a tracking feature for keeping tabs on unruly teens. If you're into seeing gadgets splayed open like an organ transplant patient check out the gallery below. %Gallery-123787%
TomTom sends HD Traffic update to all Live models, extends Traffic Manifesto to US (video)
TomTom CEO Harold Goddijn announced at a NYC event last night that the company's HD Traffic service, previously only included with the Go 2535 M Live, would be available on all U.S. Live models, including the Go 740 Live and XL 340 Live. Traffic updates will be one component of the subscription-based Live, which will also see a 50 percent price drop, to $60 per year. This is all part of TomTom's grand Traffic Manifesto, which aims to cut traffic by five percent overall. Achieving this rather lofty goal in the U.S. would require 10 percent of the country's drivers to be using Live, which transmits real-time traffic data using a dedicated AT&T SIM. The company says drivers using the service themselves can expect to see travel times reduced by up to 15 percent. Our commute often involves a pajama-clad hike from the bed to the desk, so if you're currently a subscriber who drives to work, let us know if Traffic is making a dent in your travels.
Windows Phone 7 updates Bing to find music and barcodes, provide turn-by-turn directions and send speech-to-text SMS?
Developers are getting plenty of toys alongside Windows Phone 7's "Mango" release, but there may be extra baubles for regular users, too -- Microsoft will reportedly add a few features to Bing in the near future which could prove particularly useful. According to the latest episode of the Windows Phone Dev Podcast -- which hosted Microsoft's Brandon Watson as a guest -- a new function called Bing Audio will act like a Shazam for recognizing music (and will sell you Zune tracks) while Bing Vision will use your smartphone's camera to read barcodes and do optical character recognition, plus potentially provide support for augmented reality apps. There's also allegedly turn-by-turn voice directions for Bing Maps and a native podcast player, and one more potentially exciting thing -- voice-to-text for sending SMS messages without lifting a finger. Hear all about the rumor at our source link, at just about the 40-minute mark. [Thanks to everyone who sent this in]
Is a Nuance and Apple deal in the works?
TechCrunch is reporting that Apple is in the process of some sort of deal with Nuance Communications, one of the leading companies in the field of speech recognition. Many readers may be familiar with Nuance's Dragon NaturallySpeaking software, however the Dragon speech engine is also licensed and used in a number of apps for Windows, OS X, iOS, and Android. What could the deal be? The most obvious choice is an acquisition, but as TC points out, it would cost Apple at least US$6 billion to buy the company. Apple's got the cash, but even for them that would be quite a purchase. TechCrunch thinks it's most likely the two companies are entering into some sort of partnership "that will be vital to both companies and could shape the future of iOS." Speech recognition has been rumored to be a big part of the future of iOS. Last year, Apple bought another speech recognition company, Siri, which itself is powered by Nuance technology. Perhaps with the release of iOS 5 we'll be talking to our phones more than using them to talk to people.
Voice-controlled Japanese robot assists with eating, makes veggies more fun (video)
Isao Wakabayashi, a student at Chukyo University in Japan, seems to have made the arduous chore of eating easier. Using a customized version of a Robix robot kit, Wakabayashi coded a program that makes the feeder recognize individual food items and feed them to you. The meal-assistant features two arms, dexterous enough to handle utensils, and can be controlled using your voice. In theory, this system would be ideal for the elderly, folks that currently have trouble eating by themselves, or you know -- for those that may or may not be too lazy to bring food to their face.
Japanese elevators get voice recognition, Japanese elevator rides get even more awkward
We here at Engadget are all about helping the less fortunate, so Mitsubishi Electric's latest innovation in elevator tech has us all warm and fuzzy. The new interface allows for blind users -- and presumably lazy users -- to select their destination floor by voice, with a subsequent announcement when they arrive. Additionally, the system kicks in whenever it detects a wheelchair, replacing the potentially difficult process of reaching high buttons with the simple act of speaking. No word on whether the system works in English just yet or if it'll make it to the States, but you might want to brush up on your Japanese either way.
TomTom's GO 2435 / 2535 PNDs get quiet teaser, we're left wondering what's new
The very busy folks over at TomTom have just squeezed out two new sets of PNDs sporting touchscreens, voice recognition, and a "new, intuitive user interface," but despite the company's high profile on the GPS market, the GO 2435, which works a 4.3-inch screen, and the GO 2535, a 5-inch iteration, slipped out without much ado. Both PNDs come in three versions: the "T" series supports lifetime traffic updates, the "M" line offers lifetime map updates, and the "MT" edition features -- you guessed it -- lifetime traffic and map updates. Thus far, the basic specs resemble those of previous GO PNDs -- both tout Bluetooth calling, 4GB flash storage, and 3 hours of battery life -- leaving us to wonder what's up with this "new, intuitive user interface?" Among other things, TomTom is still mum on price and availability, which means we'll have to wait until they speak up to give you all the dirty details.
Nuance opens Dragon Mobile SDK to app developers, we see end to embarrassing dictation
There are some messages that are just too embarrassing to dictate to a human being. Lucky for us and the retired circus contortionist we hired to type up our missives, Nuance is expanding the reach of its transcription software by making its Dragon Mobile SDK available to developers for use in iOS and Android applications. The SDK, which is free to members of the Nuance Mobile Developer Program, sports speech-to-text capabilities in eight languages and text-to-speech in 35. There are already apps out there that can do the job, including Nuance's own Dragon Dictation, but we welcome new advances in automated transcription. You know, it's not exactly a walk in the park dictating an entire Clay Aiken Fan Club newsletter to a guy named Sid the Human Pretzel.
Apple looking to hire voice technology, speech recognition specialists
Apple is looking to hire several voice and speech recognition experts for iOS, according to four new postings on Apple's job board. The new hires would join the iOS Application Framework team and would be working on "speech-related development activities." These multiple listings come hot on the heels of the public release of several patents detailing contextual voice commands for the iPhone. If you add in Apple's acquisition of Siri, you have the possible beginning of a robust voice control system for the iPhone and other iOS devices. Improved speech recognition would be a welcome addition to the iOS platform. In my experience, the current implementation of voice command on the iPhone is mediocre. While third-party applications like Dragon Dictation are superb, native speech control is wrong as often as it is right. Apple needs to refresh this portion of its mobile OS as it is falling behind Android in voice control technology. Google recently added support for personalized recognition to its popular Voice Search application for Android. In its current form, the app can be fine tuned to your voice so that you can dictate a text, navigate to a destination or place a call with increasing accuracy. [via 9to5 & CNET]
Google Voice Search update helps you personalize your results, helps Google build another database to take over the world
Google Voice Actions was the first step towards our Star Trek dreams of lassoing the world with naught but vocal cords, and today Google's taken a second hop towards that inevitable future by letting Android devices record our every utterance. Yes, if you've got a handset running Froyo or better, you can download an update for Google Voice Search right now, which will let your phone dynamically personalize its speech-to-text engine to better recognize your voice most every time you use it. Of course, by so doing you're giving Google permission to record your sentences -- anonymously, of course -- to use in future products, but whether that's a problem or just a happy coincidence depends on whether you take Google at its word. We hit the "yes" button, in case you're curious. Find it on Android Market, or just use the handy-dandy QR code below.
LG's Optimus 7 gets previewed by Korean newspaper, has voice to text feature?
You know how we abhor machine translation, but this rumor was too juicy to pass up -- the Korea Economic Daily reportedly got hands-on with LG's Optimus 7 (aka E900) way ahead of release, and if we're reading this right, the Windows Phone 7 device will be capable of writing your text messages, emails and status updates just by hearing you speak. The publication also reports it's got a 3.8-inch, 800 x 480 screen (rather than the 3.5 or 3.7 inches we've heard before), a 1500 mAh battery, 16GB of built-in storage and a 1GHz processor. There's also apparently "automatic panorama" feature where you simply pan the camera to take stills and stitch them together, which sounds a lot like the Sweep Panorama dealie Sony recently added to its Cyber-Shot lineup. Can we expect a US version to have these features? Hard to say. Even should this preview be wholly legit, speech-to-text would probably need quite the overhaul to tell English from Korean -- and let's not even get started on Engrish.
Kinect won't support voice recognition in some countries until 2011, putting on airs won't help
"If you see it, just say it," says Microsoft's Kinect, but only in the US, UK, Mexico and Japan to start -- those are the only four countries that will support Kinect voice control by the motion-sensing peripheral's November launch. So said Microsoft PR manager Lidia Pitzalis in an interview with Eurogamer Spain, adding that Germany, Italy, France and even Canada and Spain would have to wait for a Spring 2011 update for additional language support. Microsoft claims it's a matter of accents throwing off the voice recognition, as the company's already had to build separate software versions for US and UK English, but if you're an Italian who can fake a good Cockney, you're still out of luck, as Microsoft's reportedly disabling voice control entirely in said regions until it can formally support it. And game companies wonder why we try so hard to hack those DVD drives... Update: Canada will in fact have voice support, according to Microsoft. So long as you speak English and not French.
Vlingo bows to Google Voice Actions, makes Android version free of charge
When Google Navigation hit the scene, it sounded the death knell for paid GPS on Android, so you can imagine the doom and gloom at Vlingo HQ last week when Google released the similar Voice Actions for free. As it turns out, however, Vlingo's not going to give up that easily; Vlingo for Android, once a $10 download, is now free as well. In a surprisingly gracious blog post that genuinely congratulates Google on the accomplishment, CEO Dave Grannan explains that he wants Android users to be able to freely compare the services as Vlingo adds features further down the road, and makes one valid point in his firm's favor -- you can try Vlingo now if you've got Android 2.0 or above, but Google's service only runs on Froyo. Interestingly enough, Vlingo on Android was the only version that actually charged; on Nokia, Blackberry and Windows Mobile, however, you could purchase a "Plus" license. Perhaps the company's not quite as generous as we thought, but there's still no arguing with a free voice command service that also reads your email aloud -- go ahead and give it a try.
Ford's SYNC learns 100x more voice commands, integrates Nuance technology
We'll be honest -- we weren't the biggest fans of SYNC from the onset, but that had more to do with its insistence on playing Run DMC on our request for Naughty By Nature during a cross-town CES run than anything else. Now, Ford remains one of the few automakers that actually bothers to update their navigation systems on a consistent basis, and while SYNC is still far from perfect, it's getting a rather significant update today. The introduction of MyFord Touch brings the amount of commands that SYNC understands up to 10,000 -- that's a pretty big jump from 100, where it began. Following in BMW's footsteps, Ford is also integrating voice control technology from Nuance, which allows drivers to speak "more naturally" to the system rather than having to memorize a few hundreds first-level snippets. Unfortunately, we still think voice control within vehicles is too slow; much like Windows Vista, there are simply too many confirmation prompts, too many forks in the decision tree and not enough of an overall benefit to chose it over handling things with your digits. Feel free to disagree and pick this up on your 2011 Ford Edge, but be sure to view the video past the jump before getting all hot and bothered.
Kinect-optimized Xbox 360 Dashboard preview
We just got a look at the new Kinect Dashboard-lite for the Xbox 360. In some ways, it's pretty charming, with fun, jazzed up icons (when you hover over them they tilt and show off depth), a simplistic layout, and some great voice controls. The downside is this all comes at the cost of a brand new, fairly redundant interface for accessing functions that are already available with your Xbox 360 controller in the regular Dashboard. Still, there's no denying the joy of waving a hand to log in, hovering over icons to select channels (though the wait-to-click mechanism strikes us as eventually frustrating), and scrubbing through media with very intuitive gestures. The best part is the voice control, however, allowing you to speak "Xbox" and then a command like play, pause, back, forward, stop, Zune, etc. It's ripe for abuse (pausing your S.O.'s seventh Netflix screening of The Constant Gardner ad nauseam), but Microsoft says it's working on eliminating random commands that might seep into conversation. Our biggest worry is that we're going to see confusion and fragmentation of functionality with the divergent UI requirements that Kinect apps and regular Dashboard apps require -- we can't imagine the Facebook or Last.fm folks being super thrilled at creating an all-new app just to support Microsoft's whims, but we're told just such apps are in the works. Is the existing Dashboard really so complicated that Kinect couldn't control it? Check out a video of this puzzling Kinectasticness after the break. %Gallery-95214%
