IBM Touts Speech Technology Enabling Real-Life Improvements

Big Blue showed off its latest is speech technology and related plans on Tuesday, with demonstrations of applications that improve convenience for drivers, translations for the military and others interacting with foreigners in the real world, and Web translations for news readers. The company cheerfully announced it has cleared several hurdles that have been hampering development of speech technology, including background noise, regional dialects, and accents.

The technologies also add emotional cues to automated voice-response systems and interpret different dialects and phrasing, while allowing many realistic command and response scenarios for mobile voice technologies. Most work in several languages, with Arabic, English, Mandarin, and English being the most popular among IBM's applications because of business potential and government demands.

During demonstrations, an IBM employee spoke Baghdadi Arabic into a microphone to show how a laptop with IBM technology could translate and repeat his statements in English, while also displaying text. American military forces in Iraq are testing prototypes of rugged laptops and PDAs with the Multilingual Automatic Speech-to-Speech Translator -- or Mastor -- IBM said. Down the line, company executives noted the technology could be used for travel, telecommunications, and medicine.

"We're breaking the translation barriers," David Nahamoo, CTO for speech technology, IBM Research, said during an interview Tuesday.

Sponsored post

Nahamoo said that the technology could improve communication for Internet communities, which are now limited according to the languages that participants speak.

"I think the impact of translation technology -- when we've made enough progress so it's pervasive -- is going to be big," he said, adding that the impact could be akin to major breakthroughs in transportation and telephony.

Other partnering companies showed how IBM's Embedded ViaVoice speech recognition technology supports Pioneer Electronics' AVIC Z2 system. The system, with a suggested retail price of $2,200, includes hands-free voice recognition, with conversational commands for navigation, video, and audio functions.

IBM said it is also working with Avoca Semiconductor and All Media Guide to allow music listeners to select songs from large libraries by stating a variety of requests.

"You can say, 'Play me the Godfather of Soul, Mr. Please, Please, the Hardest -Working Man in Show Business," AMG product manager Zac Johnson said.