A new thought-to-speech technology promises to give voice to the voiceless, to silence cell phone yellers, and to provide a discreet way to query search engines while in public.

At the Texas Instruments Developer Conference in Dallas last week, Michael Callahan, CEO and co-founder of Ambient Corp. of Champaign, Ill., (not to be confused with Ambient Corporation of Newton, Mass.), demonstrated Audeo, a way to convert the deliberate nerve impulses that precede vocalization into digital information that can be turned into synthesized speech.

"What we came up with was essentially a technology that captures neurological information from the brain, and translates it into speech," said Callahan in a video of the presentation.

The technology can't read minds. "It's a step above thinking but a step below speaking," Callahan explained. It requires, in other words, a deliberate effort to think about speaking words. But it can be used to provide those who have lost the ability to speak with a substitute form of vocal communication.

Later this year, Ambient plans to offer the technology to patients with amyotrophic lateral sclerosis (ALS), commonly known as Lou Gehrig's disease, a degenerative disorder that robs those afflicted of their ability to speak and of voluntary muscle control.

Ambient's technology is also being used to develop wheelchairs that can be controlled through nerve signals. It relies on TI's ultralow-power MSP430 microcontroller because medical applications are among those least able to afford battery failure and thus must be extremely frugal with power usage.

Looking ahead, Callahan also anticipates Audeo will have general-purpose applications. "One of the applications that we're developing is what we call voiceless communication," said Callahan. "Take for example when you're talking on a cell phone. Why do all of the people around you need to hear your conversation? With this technology we can allow people to speak silently, and not miss important calls."

Callahan also expects his company's technology could help people conduct searches on mobile devices without typing in text commands. In such a scenario, a cell phone user would merely have to concentrate on search terms and the appropriate keywords would be translated from nerve impulse to keyword text and sent to a mobile search engine.

"We can augment human intelligence by using all the collective knowledge on the Internet," said Callahan.

There are limitations, however. Audeo's vocabulary currently consists of only about 150 words. Callahan likens the performance of the system to the early days of speech recognition. Even so, for ALS patients, he said, the system has a huge amount of value.