21 November 2008

Voice and Gesture Mobile Phones

Five years from now, it is likely that the mobile phone you will be holding will be a smooth, sleek brick — a piece of metal and plastic with a few grooves in it and little more. Like the iPhone, it will be mostly display; unlike the iPhone, it will respond to voice commands and gestures as well as touch. You could listen to music, access the internet, use the camera and shop for gadgets by just telling your phone what you want to do, by waving your fingers at it, or by aiming its camera at an object you're interested in buying. Over the last few years, advances in display technology and processing power have turned smartphones into capable tiny computers. Mobile phones have gone beyond traditional audio communication and texting to support a wide range of multimedia and office applications. The one thing that hasn't changed, until recently, is the tiny keypad. Sure, there have been some tweaks, such as T9 predictive text input that cuts down on the time it takes to type, a QWERTY keyboard instead of a 12-key one, or the touchscreen version of a keyboard found on the iPhone. But fundamentally, the act of telling your phone what to do still involves a lot of thumb-twiddling. Experts say the industry needs a new wave of interface technologies to transform how we relate to our phones. The traditional keypads and scroll wheels will give way to haptics, advanced speech recognition and motion sensors. Until Apple's iPhone came along, keypads were a standard feature on all mobile phones. The iPhone paved the way for a range of touchscreen-based phones, including the T-Mobile G1 and the upcoming BlackBerry Storm. So far, even iPhone clones require navigation across multiple screens to complete a task. That will change as touchscreens become more sophisticated and cheaper. Instead of a single large screen that is fragile and smudged by fingerprints, phone designers could create products with multiple touch screens.

Users could also interact with their phone by simply speaking to it using technology from companies such as Cambridge, Massachusetts-based Vlingo. Vlingo's application allows users to command their phones by voice. That could enable you to speak the URLs for web pages or dictate e-mail messages. Natural speech recognition has long been challenging for human-computer interface researchers. Most devices with speech-recognition capabilities require users to speak commands in an artificially clear, stilted way. They also tend to have high error rates, leading to user disenchantment. Unlike conventional voice-recognition technologies, which require specific applications built to recognize selected language commands, Vlingo uses a more open-ended approach. User voice commands are captured as audio files and transferred over the wireless connection to a server, where they're processed. The technology personalizes itself for each individual user, recognizing and training itself based on the individual user's speech patterns. The technology has already found a major partner in Yahoo, which offers voice-enabled search on BlackBerry phones. Vlingo's completely voice-powered user interface is also available on Research In Motion phones, such as the BlackBerry Curve and Pearl. Vlingo hopes to expand its services to additional platforms such as Symbian, Android and feature phones over the next few months. Moreover, even the traditional keypad is set to get a face lift. Typing on a touchscreen keypad is slow and difficult, even for those without stubby fingers or long nails. That's where Swype comes in. It allows users to use trace a continuous motion on an onscreen QWERTY keypad instead of tapping individual characters. For instance, instead of typing the word infinity, users can just draw a line through each of the characters.

More information:

http://blog.wired.com/gadgets/2008/11/buttons-make-wa.html