Silicon Investor (SI) -- The First Internet Community

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Technology Stocks : Voice recognition... is utter nonsense in computing -- Ignore unavailable to you. Want to Upgrade?

To: stak who wrote (11)	10/19/1998 7:40:00 PM
From: Savant	Respond to of 112

Am certain Ali is smart, most engineer/comp whizes are, but perhaps a little too narrowly focused. diff. speech/voice....all you said, plus the part about ability to interpret across a wide number of individuals and differentiate between various speakers. About that refrig door.....have you tried WD-40? I forgot to add the advantage of sppech rekognishun over voice wich is speling and tiping siklls r sumtimse lakking. You might say I "recognized" the thread. I'm waiting for the comp. that can read my mind. Best, Savant

To: stak who wrote (11)	10/20/1998 1:57:00 AM
From: stak	Read Replies (2) \| Respond to of 112

Intel Katmai chip will give voice to applications By Ephraim Schwartz InfoWorld Electric Posted at 7:01 AM PT, Oct 19, 1998 If Intel has its way, every application running on its next-generation processors would automatically be speech-enabled this time next year. Intel is including the speech technology in its next processor, code-named Katmai, due out in the first quarter of 1999, and experts are predicting that by 2000 the technology will be able to go beyond simple speech-to-text input. "You will be able to say, 'What were sales for last September and how does that compare for this September?'" said Steven Rondel, president of Conversational Computing, in Redmond, Wash. Microsoft is also throwing its weight behind speech recognition, and will eventually replace the current GUI with spoken commands. The software giant will put its Whistler text-to-speech engine, for accessing data such as e-mail over voice telephone lines, in Windows NT 5.0 and its Whisper speech-to-text engine in a future OS. With OS and chip support in place, PCs will soon become personal assistants. "Users will say, 'Print two copies of my financials, in landscape mode on the network printer,'" said Bob Kutnick, chief technology officer at Lernout & Hauspie, in Burlington, Mass. Major corporate users appear interested but cautious. "It will take a better than 95 percent accuracy for a company like Mobil to use speech technology," said Cliff Walters, a laboratory supervisor at Mobil Technology, in Dallas. Walters sees the benefits of speech technology playing out in niche areas at Mobil, such as in help desk situations when users are already on the phone and talking. Intel started its project by creating a separate speech team, spun off from its original Katmai development team, to deal solely with speech technology, sources said. That group developed a new set of Katmai instructions that will work with the "hidden Markov model," algorithms commonly used by speech developers to improve speech-recognition accuracy and the speed of recognitions. The current speech algorithms are excellent, according to Raphael Wong, manager of worldwide speech programs at Intel. But the stumbling block until now has been the inadequacy of the processors. With the introduction of 450-MHz and 500-MHz processors in 1999, all of that will change. The new Katmai instruction set from Intel will also improve the capability of voice and data to work over phone lines without distortion, according to Bill Meisel, president of TMA Associates, a speech-technology consultancy in Tarzana, Calif. In 1999, voice-enabled applications used over phone lines will make major inroads into corporate services, according to David Nahamoo, senior manager of human language technologies at IBM Research, in Yorktown, N.Y. Users will have their e-mail, contact information, and datebook appointments translated into voice and spoken to them over the phone. "It will revolutionize how users in large corporations access information and services," Nahamoo said. Enterprises will be able to remotely support customers around the clock. Intel is not alone in pushing speech as the next user interface. A speech consortium headed by IBM and Sun Microsystems will ship the first Java speech API by the first quarter of 1999. "We are designing this API to accommodate the whole enterprise implementation of applications over different operating systems and different environments," Nahamoo said. Java's strength as a client/server technology is enabling developers to create applets that work inside browsers. IBM is currently developing a Java-based, speech-enabled Internet search engine.