Talking to your computer is hardly an anomaly -- everyone does it when the machine crashes.
But the companies displaying their wares at Comdex are hoping that natural language commands -- not swear words -- will become the method of the future for speaking to devices large and small.
Speech-recognition heavyweights IBM and Lernout & Hauspie (L&H) are displaying a wide range of new voice products -- and a wide variety of uses for them.
David Barnes, senior product evangelist for IBM's speech systems, sees voice technology as the next new interface for desktop PCs, phones and embedded devices.
"We want to be the engine underneath it all," Barnes said.
ViaVoice showing up across Comdex
IBM ViaVoice technology is showing up in products across the Comdex floor, from banking to the wearable computer. No longer is the technology itself a gee-whiz feature of the show. Voice technology vendors are exhibiting more real-life and consumer uses for the technology than ever before.
"It's here. It's not science fiction," Barnes said. "Voice technology is a front for e-business."
To that end, IBM is at the Citibank booth, displaying an upcoming product that will allow bank users to perform tasks such as checking their account balance by talking to the computer.
The company also provides the underlying speech technology for an Olympus voice-recorder at the show. Users can dictate their document into the recorder and then transfer it to a PC by inserting a chip from the recorder into the PC/MCIA slot in the computer. ViaVoice translates the recording into text.
And in perhaps the funkiest use of the technology, ViaVoice appears in Xybernaut's new wearable computer -- a palm-sized device containing a 200-233 MHz processor and a head-mounted, silver dollar-sized monitor. Designed for people in the field who might need to keep their hands free, Xybernaut lets workers such as telephone pole repairmen and remote sales people dictate their findings into the computer. "Speech technology is really a natural for this type of application," Michael Reagan, Xybernaut's director of marketing, said. "It's only when you get a wearable type of computer that this can be done."
Accuracy hurdle yet to be cleared
Speech recognition technology still has several accuracy hurdles to clear, but it's come a long way since Comdex last year. Anne-Marie Derouault, director of IBM's global speech marketing and sales, expects that speech technology will be ubiquitous within 2-5 years and within 5-10 years will become extremely user friendly -- to the extent that people can ask natural-language questions in a wide variety of accents.
"You have to get it into people's minds that it really is here," Derouault said.
The technology's momentum will be fueled not only by greater accuracy in the future, but also by faster processors, more precise microphones and a wider variety of applications. IBM is working on natural language features that it plans to release to the financial sector early next year. It's also working on tool kits and technology for embedded devices that don't require keyboards, such as car computers and home alarms.
Two new products launched
Lernout & Hauspie is also eyeing the consumer market, and has introduced two new products at Comdex. "Now You're Talking" lets users incorporate voice technology into office applications, allowing people to dictate directly into their Microsoft Office documents, navigate by voice through digital calendars, and search the Web by voice query.
The company also unveiled Talking Max, the company's first voice-activated game, which lets a user raise, feed, and interact with a surly parrot named Max. Eventually, the company plans to make Max an intelligent agent who'll check your e-mail and perform other tasks.
L&H spokesman James Williams said the company is embracing consumer products as part of its campaign to make speech technology more readily available to the average user. "Speech recognition technology is at a point right now when it's ready to go mainstream," Williams said.
The company also unveiled text-to-speech technology called RealSpeak. It offers companies that want to incorporate speech into their products a more natural-sounding voice.
To Anne-Marie Derouault
Director of Global Speech & Marketing & Sales
Thank-you for taking the time to read this email.
My name is Andrew Blode CEO of the Jack & Robert Smorgon Families Foundation. Our Foundation has been working very closely with those that suffer with Multiple Sclerosis (MS).
What we have discovered is that through speech recognition software we are able to open up a whole new world for these MS sufferers. And in fact make an enourmous impact to the lives to the all sufferers of any disability which retards the movement of the uppper limbs.
By request is there anyway IBM can assist, or work in partnership with us, purchasing software, training and installing such software.
Yous sincerely,
Andrew Blode
CEO The Jack & Robert Smorgon Families Foundation
61 03 9533-6666