Voice recognition technology goes mainstream
Fifty years ago, scientists predicted and movies portrayed the 21st century as a time when all of us would casually speak to our computer and get a reply.
Although reality didn’t quite turn out that way, once your expectations are adjusted, you might be impressed with the expanding technologies for voice-recognition technology.
An ever-developing application is “voice-recognition typing,” a sophisticated technology that has been available to PC owners for the past 10 years.
This technology strives to take the hands off the keyboard and instead listen to and convert the user’s voice to text. Voice-recognition typing can be quite impressive, with claims of 99 percent accuracy and speed.
Candidates for voice-recognition typing are workers who type high volumes of routine text. Anyone considering these products should realize that the program will need to be trained and may or may not meet expectations.
The best strategy is to experiment by purchasing from a vendor who will assure customer satisfaction.
The trouble with voice
Users of voice recognition can experience puzzling errors for several reasons.
The computer algorithms behind most voice-recognition engines do not have the benefit of observing facial expressions or gestures that people use when communicating. Also, if anything alters the human voice (like a cold) voice recognition can fall short.
The most popular voice-recognition program is Dragon NaturallySpeaking (www.scansoft.com) version 8.
Dragon has a standard version, along with robust legal and medical versions, with retail pricing from $99 to $1,050, depending on the version.
An important and often overlooked aspect for decent voice-recognition typing is a good microphone. The best kind of microphone is one that is powered by and uses a USB connector, easily available at most computer stores.
Owners of both Windows XP and Office XP have the advantage of a built-in voice-recognition application known as SRE. (SRE may need initial setup.) However, SRE is just a basic “plain Jane” version of voice recognition and may not be suitable for more demanding applications.
Voice recognition is far from being limited to PC applications. Handheld computing devices, small gadgets and special services are using voice recognition for neat, helpful applications. Examples include voice command systems in cars and many devices where a keyboard cannot be placed.
Particularly useful for the business traveler are small handheld translation devices such as the Ectaco Universal Translator UT-103 (www.universal-translator.net).
For around $300, this little wonder can translate aloud an English phrase into equivalent French, German, or Spanish. Devices like this may not be perfect, but are certainly worth trying.
Many businesses are heavily dependent on an Microsoft’s e-mail product, Outlook. At many large businesses, the server back end of Outlook e-mail clients is known as Exchange server.
Exchange enables large enterprise e-mail systems that include corporate-wide contact lists, calendaring, and e-mail. Microsoft indicated recently at the SpeechTEK Exposition and Educational Conference in New York that it plans to integrate technology from its Microsoft Speech Server into Exchange, possibly as early as 2006. However, it should be noted that there are third-party vendors that already offer some degree of voice integration with Exchange today. One of leaders of this technology is Adomo (www.adomo.com), Cupertino, Calif.
In July, Adomo released a new product known as Voice Messaging for Exchange, which alters Exchange from just as a place to store text messages to a communications backbone. Adomo’s solution would integrate phone voice mail messages right along side text messages in Exchange/Outlook systems. According to Tony Romanek, senior systems engineer for Next Step Systems Integrations (www.nextstepsys.com), Scranton, which specializes in networking, e-mail solutions and hosting, “If Microsoft says they will they will implement voice-enabled Exchange, it’s just a matter of time. One feature to expect will be the ability to read back text e-mail over the phone, along with phone voice mail integration into inboxes.”
Romanek warns that these more resource-intense voice applications will take time to implement. Though it may be quite common to listen to your e-mail in the near future, creating new text based e-mail from the phone will be a different and more challenging situation. Servers that house these new voice enabled systems will definitely need careful evaluation to assure reasonable performance.
Article provided by Northeast PA Business Journal