Let’s start with the obvious; voice interaction – despite recent developments discussed below – is nothing new. It could, however, be the next user interface.
Software that can convert spoken words into written text has been available since the early 1980s, with modern commercially available systems claiming accuracy upwards of 98% (see further reading, below). Such results are not bad, but not great either.
In almost 30 years of use and refined development, speech recognition still can’t match the recognition levels of human beings. For the most part, such issues mean most people are still confined to the traditional input methods of keyboard and mouse.
But lurking on the horizon is another upheaval and a potential boon for speech recognition. Google’s recently announced Nexus One handset includes voice recognition technology that allows the individual to control the device (see further reading).
The intuitive system – which learns in relation to individual queries – allows the user to interact with a range of services, from composing an electronic mail in Gmail to visiting the world’s site via Google Maps.
The search giant isn’t alone in developing more intuitive voice services. Microsoft’s Windows operating system uses voice control technology to help the user control Vista. The technology has been refined further in the recently released Windows 7 platform.
Such developments, however, have not been without issues. A famously unsuccessful demonstration of Vista voice recognition software in 2006 led to a simple note of “Dear Mom” being translated as “Dear Aunt, let’s set so double the killer delete select all” (see further reading).
Reuters reported that Microsoft chief executive Steve Ballmer blamed the failed speech recognition product demonstration on “a little bit of echo” in the room, which confused the speech-to-text system.
For users of voice technology, such confusion has been a common concern. But do the refinements in Windows 7, and the progress made by Google, show that we are actually getting to a point where voice recognition is usable?
Voice controlled car stereo systems – which have often suffered due to background noise – are now viewed as increasingly reliable. And early feedback from Nexus One users suggests the voice recognition technology is “amazing” (see further reading).
After three decades, then, we might finally have reached the tipping point for voice control. But as the speech interface becomes commonplace, one final word of warning: be careful where you talk.
Straining ears could pick up confidential information from spoken dictation. Worse still, confused members of the public might question your sanity! So, be careful as you embark on the road to a spoken revolution.