Computerworld -
You arrive at your desk in the morning and sit down in front of your computer. Instead of issuing a voice command to your PC, or reciting an email, or dictating a memo to your boss, you start typing and clicking. In the environs of the office, where speech technology could save us time and make us more productive, most of us are still stuck with keyboards and mice.
Yet once we're away from the office, many of us don't think twice about issuing voice commands to our smartphones -- whether that means voice-dialing the phone, speaking a search term to Google or asking Siri what today's weather will be like.
Companies that provide speech technology have invested heavily in the concept of "personal digital assistants" like Apple's Siri and Google Voice Actions (available on many Android phones) that can understand natural-language commands, says Dan Miller, the senior analyst and founder of Opus Research. In fact, most recent breakthroughs for speech-recognition technology have come in the area of cloud-based natural language searches made from mobile devices, he says.
The main advancement is that the speech tools are now closer to the user -- on our phones and tablets as we go about our day -- and many run in the cloud, which provides immediate processing and a constantly expanding language database. Unlike older desktop-based software, these new tools do not require speech training, thanks to improvements in the algorithms. "We can be pretty imprecise in what we say," Miller says.
To continue reading, register here to become an Insider
It's FREE to join
Computerworld - You arrive at your desk in the morning and sit down in front of your computer. Instead of issuing a voice command to your PC, or reciting an email, or dictating a memo to your boss, you start typing and clicking. In the environs of the office, where speech technology could save us time and make us more productive, most of us are still stuck with keyboards and mice.
Yet once we're away from the office, many of us don't think twice about issuing voice commands to our smartphones -- whether that means voice-dialing the phone, speaking a search term to Google or asking Siri what today's weather will be like.
Companies that provide speech technology have invested heavily in the concept of "personal digital assistants" like Apple's Siri and Google Voice Actions (available on many Android phones) that can understand natural-language commands, says Dan Miller, the senior analyst and founder of Opus Research. In fact, most recent breakthroughs for speech-recognition technology have come in the area of cloud-based natural language searches made from mobile devices, he says.
The main advancement is that the speech tools are now closer to the user -- on our phones and tablets as we go about our day -- and many run in the cloud, which provides immediate processing and a constantly expanding language database. Unlike older desktop-based software, these new tools do not require speech training, thanks to improvements in the algorithms. "We can be pretty imprecise in what we say," Miller says.
Of course, today's speech technology isn't perfect. Even on mobile devices, voice-recognition software doesn't work well for everyone. And some natural-language sentences like "Let my boss know I am running late for my meeting" still confound even the most intelligent voice-controlled systems. Companies like Nuance, Microsoft and Apple have built vast language databases for their speech-recognition products to tap into, but even today the software sometimes has trouble knowing the difference between "hamburger," the beef product, and "Hamburg," the city in Germany.
Miller says there is a need for even more artificial intelligence that operates in the cloud. "We're getting closer," he says. "There are challenges at every level, but progress is being made -- though it may never be perfect."
But while ever-improving results are driving speech adoption on mobile devices, speech technologies have not made much of an impact in the workplace: Walk into any corporate campus in America, and you'll be hard-pressed to find workers issuing voice commands. In this story we'll explore what's holding it back and look at some ways speech technologies could benefit workers, both at their desks and around the office.
Speech at the workstation
For visually impaired users or those who suffer from repetitive stress injuries such as carpal tunnel syndrome, controlling a computer by voice makes good sense for navigating the interface and applications, conducting searches and dictating long emails and business documents. There are benefits for other users as well, especially slow typists, mobile users and meeting attendees who want a record of the proceedings.
The technology is available, Miller says. Basic voice-control options have been built into Mac OS X and Windows for years, and this summer Apple added speech-to-text dictation capabilities to its OS X Mountain Lion release. Last year Google brought voice-initiated search, which debuted on mobile devices, to desktop and laptop computers in its Chrome browser.
For more advanced dictation and PC control capabilities, dedicated voice-recognition software such as Nuance's Dragon NaturallySpeaking has steadily improved over the years. In Computerworld reviewer Lamont Wood's tests, the current version of NaturallySpeaking Premium scored higher than 99% in speech-to-text accuracy.
Wood says that for him, composing via speech is about twice as fast as typing; other users' results will differ depending on their typing speed and comfort with speech software. He also notes that using the newest noise-canceling headsets means the software doesn't get confused by background sounds, which was often a problem in the past.
You are previewing premium content. 
Free Insider Guide
