Will Voice Commands Silence the Click?
|
While it's still far too early to trash your mouse, speech-activated access to the Internet is making giant strides
|
Web surfers be warned: Your clicking days may be numbered. Index-finger clicks on a gliding mouse could soon be replaced by an effortless voice command. Whether it be voice recognition via vocal portals for use with a telephone, or plug-ins for use with a computer or cell phone, Web navigation as we know it today will never be the same.
Interactive Speech Technologies, a French startup created last February, has just completed a plug-in for Web surfing on a computer. The program, which uses only 150K, is embedded in the browser and activated when a Web page using the technology is called up. Web users then pronounce phrases that appear as choices on the menu bar. The plug-ins give added value to a site and make Web surfing a playful experience, says the company.
But it's telephony, whether through a cell or a traditional phone, that will change the cyber click as we know it today. The real novelty of the French company's program, which received an award this September from Anvar, the French agency for innovation and from the Education, Research & Technology Ministry, is that it will be used with mobile-phone technologies like WAP (wireless application protocol) and UMTS (universal mobile telecommunications system). In the U.S., where the mobile Internet has been greeted with less enthusiasm, "vocal portals" already connect traditional phone users to Internet sites via voice. "On telephones, which have a limited number of keys, the voice offers new possibilities for navigation," says Interactive Speech Technologies co-founder Anselme Dewavrin.
According to the French company, adapting voice technology to mobile phones will be relatively simple. As HTML pages evolve into WML (wireless markup language) and become accessible via mobile phones, all it will take to navigate via voice are a few extra lines of code and the installation of a plug-in. Programmers have yet to come up with a plug-in small enough for use with a mobile phone. Other obstacles to overcome involve the ability of the system to prioritize sounds and disregard background noises, as well recognizing the same word pronounced in different accents. However, "the system is interesting," says Interactive Speech Technologies, "because vocal information can be processed inside [the phone] and doesn't congest the network."
VOCAL BROKER. Vocal portals, available in the U.S. since this summer, use a more complex technology capable of recognizing entire phrases. Traditional phone users dial up a number and ask questions, which then send a computer on an Internet hunt for information about, say, weather, traffic, or movie programs. The choice of possibilities, though, remains limited. However, the beauty of this technology is that it actually "understands" sentences. If someone says, "I want to sit on a chair," the program, if it hesitates between the word "chair" and "hair," will choose the former since it had previously identified the word "sit." Responses to queries are then given vocally to the user, eliminating the need for screens, keyboards, and clicks.
In this way, the meshing of telephone and Internet looks imminent -- at least in the U.S. America Online has recently bought out Quack, which, along with BeVocal, Shoutmail, and Tellme, is one of the leading vocal portals in the U.S. "Quack's technology will help us expand our services beyond the computer and will give our clients the possibility to transform their telephones into AOL access terminals," says Ted Leonsis, president of AOL's Interactive Properties Group.
Things are likely to evolve differently in France. Although French online brokerage Filmatex uses vocal portals as in the U.S., the service is limited to share orders via voice. The future fusion of telephone and Internet, however, seems unlikely. Why? Given the high level of mobile-phone penetration in France and the arrival of wireless Internet technologies, Internet via a regular phone will likely prove unnecessary. Either way, Web navigation is at a turning point. The next challenge? Finding a vocal signal that's just as quick and efficient as the good old click.
By Stéphane Foucart
|

|