Dictation software helps those with injuries continue to work

At a time when speech recognition technology has improved in recent years to include increased vocabularies and natural dialogue capabilities, the newspaper industry has realized the benefits of voice recognition to help prevent repetitive stress injuries.

In the past, the systems required key words to be spoken along with a pronounced enunciation of each word; today, users can speak in a more normal manner.

Newspaper reporters and editors have been utilizing computer keyboards for decades to file and edit stories. As a result, this computer-intensive business has been difficult for people with physical limitations. For example, an editor who suffered from a repetitive stress injury, like carpal tunnel syndrome, would be unable to use the keyboard to perform the simple task of editing, using the keys.

In addition, people with severe nerve damage -- resulting in no control of their hands -- were unable to seek or continue their jobs in newspapers. As a result, employers have sought to assist workers by finding an alternate solution to performing their daily computer edits and transcription without the use of their hands.

"A few years ago, we had a copy editor, who worked with us and he had a voice activated system at home," said Dennis Ryerson, editor of Iowa's 163,000-circulation Des Moines Register. "He had a progressively degenerating disease, which did not allow him the use of the keyboard. He was a wonderful copy editor, who used the voice activated system, which ultimately contributed to our success."

Just how exactly does this system work?

Ryerson said that Larry Vint, the copy editor, "was able to tell the computer to turn itself on, call a story up by name, move the cursor up and down and even save and store files. He had some use of his hands but the voice activation system helped him a lot."

Voice activation has recently become an information technology tool to many newspaper executives. These systems utilize real dialogue technology to provide interaction with a software system to enable these functions. Because people speak nearly seven times faster than they write or type, these systems are ideal methods for people in publishing.

"Our copy editor used the DragonDictate system, which was able to extend his work life," said Diane Graham, assistant editor of the Des Moines Register.

'The system is amazing'
Dragon's general-purpose dictation system offers a large vocabulary with continuous speech and its newest version is the Naturally Speaking 4.0 package, which contains an active vocabulary of more than 160,000 words. The system even allows a user to browse the Web by voice, using Microsoft Internet Explorer. The computer must be trained to understand a user's voice, which can be achieved in just five minutes.

"The Dragon systems are most popular among people with disabilities in typing, having a cumulative trauma disorder," said Donald Hensel, environmental services manager for the Newspaper Association of America (NAA) of Vienna, Va. "Many people were unable to do their jobs because of their injuries. As a result, they were unable to function as a reporter or editor. The DragonDictate system is amazing. You can watch a person dictate into the machine and just see the words appear of the screen as they were spoken. People now can speak in a regular voice at a regular pace, which is the most significant difference from other systems."

Dragon now allows easier access through the use of handheld and universal serial bus microphones to assist the users.

"When I worked at the Des Moines Register, the last three or four years I had to work from my home because of mobility difficulties," said Vint, the retired copy editor. "The company introduced voice recognition programs to use with editing systems for people with various physical limitations, like carpal tunnel syndrome. The DragonDictate was a great system for me because it gave me a lot of flexibility. I didn't throw away my keyboard, though, but in terms of editing procedures, it made it a lot easier for me."

To the average person, it may seem like these programs are time consuming to perform a simple task but since the spoken voice is actually faster than the written word, it actually provides a quicker result.

"I was able to tell the computer to just delete an entire paragraph by saying, 'delete paragraph,'" Vint added. "I used the system for a few years and it worked very well for me. Since I retired, Dragon has come out with a more advanced system. ... If I was dictating text into the old system, I had to speak with a break between the individual words. The new system is very accurate and is very easy to make corrections with Naturally Speaking."

Speech and voice recognition technology is now increasing much of the innovation in user interfaces. Today, there are even voice-automated directory assistance services that are so advanced callers may believe they are speaking with a live operator instead of a computerized system. These systems are also able to recognize proper names and even handle pronunciation difficulties or accents.

Voice networks
The newspaper industry is not the only field to utilize these new systems. Many companies now provide services like call forwarding, call screening and even personalized messaging. Other multipurpose voice networks offer voice-requested services, including access to airline flight schedules, exact time reports and even weather forecasts. Most companies that utilize these systems believe they help to improve customer service because they operate nonstop and changes can be easily made when needed.

"About five or six years ago, we had difficulties with voice recognition," said Dorothy Wyatt, environmental safety manager for the 759,000-circulation Washington Post.

"I have seen the Dragon program used on workstations here and they were very helpful to the user. The quickness came from having a large capacity of RAM [random access memory] on their computer. The system really seemed to make the person's life a lot easier."

Mobile telephones, Web browsers and personal computers are all now able to accept voice-written text. The modern digital dictation systems, like Dragon, result in nearly 40 percent time savings over analog dictation and transcription.

"The best part about the Dragon systems is the continuous speech feature, which allows the user to speak into a laptop or PC with under five minutes of training required," said Renee Blodgett, director of corporate communications for Dragon Systems. "For anyone who needs to get text entered quickly and for people with repetitive stress injuries and carpal tunnel syndrome, it allows them to do things faster than they were ever able to do before. The success in the product is in the productivity gain. Dragon takes it beyond just dictation as the person can use it in Excel or any other Windows application."

As the move toward voice recognition systems increases, many people are now realizing the simplicity and ease that goes along with its use.

"Voice activated computers are a very specific area of interest for us," said Larkie Gildersleeve, director of research and technology for the Newspaper Guild, based in Washington, D.C. "We are in the process now of finding out how many people utilize this system so we just sent out a survey. So far, the response has been that most people have used an older system and are now trying the current ones. Many say these new systems have really improved over the years."

Last year Philips joined together with IBM to share technological resources and devise various voice-controlled work flow applications. This year, Philips collaborated with the Suntek Group in an effort to design the first natural speech recognition system in China. As a result, Suntek will soon form several pilot projects that are able to recognize Cantonese languages and even regional dialects.

Speech recognition technologies in the newspaper industry have made such vast advances over the years due to improved systems with higher accuracy rates. Even though voice recognition has improved dramatically since its introduction to the marketplace, it is not flawless. But companies like IBM, Philips, and Dragon are working to perfect these systems.

Longtime user
"I have been using voice recognition since September 1997," said Diana Henriques, a financial investigative reporter for the New York Times. "I was experiencing a repetitive stress injury and got Dragon's Naturally Speaking. It has been very effective for me as I use it for all my writing here and I also used it to write a book."

As Henriques has seen, voice recognition systems made a breakthrough when they went from isolated speech to continuous speech, enabling the user to speak directly into the system without delays. The older systems required a user to speak slowly and clearly so as to provide the computer with enunciated voices for recognition.

"There are many different programs for many different users," Henriques said. "I principally use voice recognition for my computer notebook, e-mail and filing stories. The Dragon system is most intuitive for writing stories, especially if you are using different applications. When typing with the older system, you had to allow time for corrections but now, it is at least two times as fast. I can turn around a story in half the time I used to be able."

Each version of voice recognition systems has progressively evolved in the wake of new research. When it first emerged in the late 1970s, speaker verification and identification was the major milestone. Since then, technology has rapidly improved to include the automatic process of natural speech and speech-to-text applications.

"I have noticed a gain in accuracy with the newer versions of voice recognition systems," Henriques said. "The improvement curve was much higher with the older versions. I have noticed an improved gain in accuracy."

Companies like Dragon, IBM, and Philips have introduced a mixture of natural speech recognition applications that provide the user and an automated system with the ability to command the controversial flow of words. Their biggest challenge now is to encourage the user to reduce any background noise, which may cause some errors to occur.

"People first encounter with the current version is its easy adaptability," Henriques added. "The real issues in applying voice recognition to a newsroom is motivation and support. The newest versions require a lot of memory but newsrooms should realize that software is much cheaper than worker's compensation. The technology department here at the New York Times has been very supportive in getting the right software, the right headset and microphone for me. There are less than 10 people here who use the system."

Even though most people who purchase the system have suffered some type of repetitive stress injury, Henriques believes it should be used widespread throughout the newsroom in an effort to prevent an injury.

"Many people view its use as a disabling stigmata but it should be viewed as being on the cutting edge instead of fear of being disabled," she said. "I started to use the system when I suffered severe pain when I typed more than one hour. There are a number of people who are not injured who are using the system right now. It's a very real solution for people who can't or don't want to type."

-- Kellie K. Speed

See also Most popular voice recognition systems

From THE COLE PAPERS, April 2000, Copyright © 2000, All Rights Reserved.