Reduction
of Repetitive Motion Injuries
It’s Time toTalk
to Your Computer
Note: The entire draft of these materials was dictated to a computer using Dragon NaturallySpeaking Deluxe, a voice recognition program from Dragon Systems. These materials were orignially prepared for the 1998 Small Firm Success Conference on April 30 and May 1 in LaCrosse, Wisconsin. The bar associations of Illinois, Indiana, Iowa, Michigan, Minnesota, Nebraska, and Wisconsin sponsored the conference. For more information see:
http://www.wisbar.org/convention/smallfirm/ or call: 800-728-7788
Table of Contents
Note: To jump to a topic, click the * after it.
Speaker Profile *
Overview *
Voice Recognition Essentials *
Benefits of Voice Recognition *
Comparison of Voice Recognition Products *
Features Chart – Dragon Naturally Speaking*
Features Chart – IBM ViaVoice *
Tips for Using Dragon NaturallySpeaking *
Speaker Profile
Wells Anderson is president of Wells Anderson Legal Tech Services in Minneapolis. His firm works with attorneys who want to upgrade their computer systems and with law offices that want better ways to manage information.
Emphasizing practical solutions for busy attorneys, Mr. Anderson works with law departments and law firms to improve profitability, satisfy clients, and manage information more effectively. He speaks on legal technology topics and has served as a technology mentor at legal conferences, including the American Bar Association Techshow.
Beginning his career as a litigator, Mr. Anderson handled jury trials and appeals, including arguing to the Minnesota Supreme Court, before becoming a corporate counsel. He is experienced in the areas of litigation management, employment law, insurance, family law and corporate law department management. Working exclusively with the legal profession, Mr. Anderson understands the needs of attorneys who want to make more profitable use of technology.
Member of the American, Minnesota State, Hennepin County, and Ramsey County Bar Associations. 1978 graduate of the University of Minnesota Law School and a member of Phi Beta Kappa. Member of the bar of Minnesota and of the U.S. District Court, Minnesota District.
|
Wells Anderson LEGAL TECH SERVICES 527 Marquette Ave., Ste. 1350 Minneapolis, MN 55402 |
phone 612-791-0471 fax 612-204-0008 e-mail WA@WellsLegalTech.com web WWW.WellsLegalTech.com |
|
Helping lawyers leverage technology |
Ó Copyright 1998 Wells Anderson |
Is It Really Time to Talk to My Computer?
Yes. The technology is finally here. If you are willing to invest in voice recognition, you will reap the benefits. Some people say that voice recognition is not really here yet. That view arises out of the frustration and disappointment that many experience when first trying the new voice recognition technology. Unfortunately, our expectations of how a computer will respond tend to be unrealistic. Perhaps we've watched too much Star Trek.
Only when both you and your computer have been adequately trained will you be ready to really use voice recognition. Perhaps because the concept of voice recognition is so simple to understand, people new to the technology expect that it will work smoothly right out of the box. We all know how to use a microphone. We all know how to talk. With the latest software, shouldn't the computer be able to write down what we say right away? No. It is worth repeating: Both you and your computer need thorough training.
Once you and your computer have been trained, you can expect accuracy ranging from 95 to 98 percent. That translates to between four and 10 errors per page, well below the accuracy of an experienced transcriptionist. But the computer offers immediate, real-time transcription at zero dollars per hour. It really is time to talk to your computer.
What Do I Need to Make Voice Recognition Work?
Whatever software package you buy, the following are essential to your success:
- High-quality sound card (one recommended by the software company is best)
- Computer with 64 MB RAM (96 MB is better) and 200+ MHz Pentium chip
- Training by experienced professionals
- Patience to make it through the training period
- An understanding of potential and pitfalls of the technology
Voice recognition is finally ready to work for you because the power and storage capacity of personal computers have reached the high levels demanded by this technology. In order for you to take advantage of the power of voice recognition, you also need the knowledge essential for exploiting it.
Processor Your computer needs power to match digitized versions of phrases you dictate with voice prints in its dictionary. At a minimum, use a Pentium 200 MHz processor with MMX capability. Because the speed of the processor affects the amount of time you will need to wait while using the program, you are better off with faster processors. If your budget allows, use a 233 MHz, 266 MHz or faster chip.
Notebook Computer vs. Hand-Held Dictation Unit
Dictating away from the office can be a very effective way of using your time. In addition, it allows you to capture thoughts when they are fresh in your mind. A notebook computer along with voice recognition software might seem to be an obvious solution.
Notebook computers now are available with processors fast enough for voice recognition. But they tend to be either very expensive or very heavy. One of the barriers to using a notebook computer for dictation is the time it takes to start it up and to shut it down. For dictation, you may want to consider purchasing a hand dictation unit that is compatible with your voice recognition software. Dragon Systems offers this combination. For working on the go, you may prefer the portability and instant-on capacities of a hand-held dictating unit and a palmtop electronic organizer. Both units can link up to your office computer to transfer dictation from the dictating unit and refresh the names and dates in the organizer.
For successful voice recognition the single most important component in your computer is the sound card. Make sure that you specify one of the high quality sound cards listed by the voice recognition software manufacturer before purchasing a new computer.
Software packages for voice recognition all seem to include a headset microphone. These inexpensive units often require the use of a small, battery operated amplifier. Some experienced voice recognition software users insist that a $100 headset offers significant advantages. Others say that their inexpensive microphones work just fine.
Since you can use voice commands or the keyboard to turn the microphone off and on, there is no need for a switch on the microphone itself. If you use a telephone headset, you will need a special switch box in order to alternate between dictating and telephone use.
User Training It is worth your time to seek out a company experienced in training people to use voice recognition software. You might think you can learn voice recognition by working with the manual and using the trial and error approach. That would be a mistake. Voice recognition offers the opportunity to effect a major change in the way documents are prepared in your office. A bad start likely will bring disillusionment and a quick end to your efforts. An experienced trainer can tactfully guide you in directions you would not find on your own.
When learning any new skill it is all too easy to form bad habits that become ingrained. With voice recognition, two training processes occur together. The computer must make mistakes in order to learn your voice. When you see mistakes on the screen, your natural tendency is to change your way of speaking in an effort to reduce the mistakes. Some of your changes may be appropriate but others will not be.
Inappropriate changes you may make in the way you talk to your computer may coincide with changes that are effective. For example: (1) the computer makes a mistake on a phrase; (2) you adjust by dictating the phrase again more slowly and clearly with a flatter tone; (3) the computer now responds correctly.
The effect will be to condition you to speak unnaturally. But slowing down is not helpful for the computer and makes you less efficient. Using a flatter tone was actually reduced accuracy, but was offset by the improvement that resulted from your speaking more clearly. It is important for you to learn to adjust your voice only in ways that are effective. In the presence of a trainer with a practiced ear, you will be encouraged to speak rapidly and distinctly. And you will be corrected when you inevitably wander into counterproductive habits.
An experienced trainer can also be invaluable in assisting you to do a good job of training your computer. You need to correct the computer’s mistakes in order to make the program work more efficiently.
Mastering the shortcuts and techniques for correcting recognition errors is a key to success. Correcting mistakes may be the biggest barrier to successfully following through with the implementation of voice recognition. If it feels like too much of an inconvenience, you will abandon using this new tool when your initial enthusiasm wears off. A considerable number of new users give up during the period that the computer is getting used to them. By mastering the techniques and shortcuts, you arm yourself with the tools you need for success.
Breaking in a person new to your office takes patience. For the office to become more efficient, productive members must spend unbillable time training in the new person. We go through this stage because the new person will add enough value to the office to compensate for productive time lost to the training process. The same is true of working with your voice recognition program. If you lack the patience both to go through effective training yourself and to train your computer, you will end up having spent a number of hours unproductively and will have nothing to show for it but unused software and a headset.
Voice recognition programs work by finding the closest match between what you say and a word that exists in the program's dictionary. The dictionary will not contain any misspelled words unless you put them into it. Therefore, the computer literally cannot misspell words. It can make mistakes in choosing words, but the words will never be misspelled. Despite this characteristic, careful proofreading is required for all but the most informal communications. Voice recognition programs make embarrassing errors that are not caught by the spelling feature of your word processor. Homonyms are words that sound alike but are spelled differently. Voice recognition programs contain sophisticated logic for differentiating among homonyms. Usually, but not always, they choose the word most appropriate in the context of the sentence. In addition to occasionally using an inappropriate homonym, the computer will substitute a phrase that sounds something like what you dictated but has an entirely different meaning. Here are some examples of errors that occurred in the dictation of these materials:
What the computer wrote:
What the author meant:
beer years
barriers
deficiencies
efficiencies
If you whack the patients
If you lack the patience
defective
effective
Proofreading by Another Person
As the author of a document, your eye will tend to see what you intended to write and not what the computer has actually written for you. You are virtually certain to overlook some recognition errors. They tend to be much farther from what you intended than garden-variety typographical errors made either by you or your assistant using the keyboard. So the stakes and the risk of embarrassment are much higher when you use voice recognition software than when you use traditional dictation.
For documents that you file with the court or send to a client, you should consider proof reading by another person mandatory. On the other hand, you may find that you can do a good enough job of proofing your email and intra-office communications that you will find voice recognition a welcome relief for these communications.
While a computerized spelling checker it is of no help with voice recognition software, a relatively new developments in grammar checking is especially useful. The computer can now put a wavy green underline under words or phrases suspected of violating rules of grammar or usage. This feature is especially useful for spotting a number of small errors that can creep into a document dictated with voice recognition.
Basic to the concept of voice recognition is that you dictate and the computer responds. If you have to alternate between using the microphone and keyboard to use a voice recognition program, a basic purpose of using voice recognition is defeated. Users new to voice recognition often complain that it is distracting and counterproductive for them to switch among the keyboard, the mouse, and the microphone. Like all Microsoft Windows programs, voice recognition programs allow you to perform the same action in more than one way. If your voice recognition program does not give you a convenient way to perform actions with the microphone, you may find it frustrating to keep switching from microphone to the mouse or keyboard. The same frustration can arise if you are unaware of the appropriate microphone commands.
An important goal for a voice recognition user is to gain total voice control over the computer. To do that, the user must master the vocabulary of commands that can be given using the microphone. Initial training courses provide important experience with this new language of commands.
Transfer of Busywork to the Computer
Your time and the time of the people you work with is valuable. Clearly there are benefits to transferring busywork from people to computers if it can be done efficiently. The most obvious benefit is saving money. For example, let us assume that an efficient administrative assistant spends 25 percent of his or her time transcribing dictation. Part of that time is spent correcting the attorney’s dictation errors. If we assume that the assistant will still need to devote 10 percent of his or her time correcting voice recognition errors, we are left with a savings of 15 percent. If we assume the cost of compensation and benefits is $40,000 per year, implementing voice recognition will result in cost savings of $6,000 dollars annually per assistant.
Increased Productivity of Staff
Support Staff Though support staff may take pride in producing well-formatted, error free legal documents, few people relish the task of straight transcription. By delegating this part of the process to the computer, support staff are freed up to focus on the content and format of documents. If the attorney becomes proficient at creating and correcting documents using voice recognition software, then the support staff can spend less time on the mechanical steps of creating documents and more time on higher level tasks. Both the firm and support staff benefit from transferring a mundane transcription task to the computer.
Voice recognition software can increase attorney productivity. Initially, there is a productivity penalty while the attorney masters this new tool. Over time, the attorney’s dictation speed can increase significantly and the computer will become more accurate as corrections are recorded in the attorney's speech files.
An additional benefit for the attorney is the capability of viewing previously dictated text immediately. With traditional dictation, the attorney dictates without being able to see any of his or her work. Voice recognition eliminates the delay that occurs when the document goes to the transcriptionist and then back to the attorney.
When the hours of the attorney do not coincide with the hours of the transcriptionist, the attorney can still dictate and get immediate turnaround.
For some attorneys, voice recognition software can allow them to work without incurring the overhead of an assistant. This is a step that should not be undertaken lightly, because there is clearly a risk of a false economy. It does not make financial sense for a $175 per hour attorney to be spending a significant amount of time performing functions more suited to an administrative assistant.
Faster, More Cost Effective Client Service
By eliminating the extra hand-offs between lawyer and assistant, voice recognition technology can reduce the time between dictation and delivery of work product. By reducing time spent on transcription, the cost of producing a document can be reduced. In addition to these basic savings, advanced features in voice recognition software allow for more sophisticated automation. Here is a time saving example. Writing instructors recommend that an author begin a writing project by a engaging in a brainstorming session, quickly writing out the stream of thoughts that the author has on the subject. Based on the brainstorming session, the author can then identify topics and subtopics. Using a computer and voice recognition software, an author can move directly from brainstorming into the process of organizing and grouping his or her thoughts.
Both of the leading voice recognition products now offer voice macros. Macros operate like little computer programs to automate frequently performed tasks. A simple example is a voice macro that enters a standard closing and signature line. More complex applications can be developed to insert boilerplate language and present pop-up windows offering the user choices among multiple options.
Reduction
of Repetitive Motion Injuries
By transferring the task of raw transcription from the administrative assistant to personal computer, voice recognition technology can dramatically reduce the amount of repetitive motion performed by the assistant. The same applies to attorneys who draft on their computers. As we move deeper into the Information Age, the risk of repetitive motion injuries (RMI) grows. Much can be done to prevent these injuries through the use of ergonomic chairs, keyboards, palm rests, stretching and exercises. Yet voice recognition technology offers the most dramatic form of prevention.
Some have speculated that dictation to computers may lead to a new category of injuries involving strained vocal cords. We are unaware of any completed research of this subject. Still it makes sense to exercise caution. By working with experienced trainers, people new to voice recognition can avoid falling into bad dictation habits that are so easy to form when the computer is untrained and prone to make more frequent errors.
A Change of Pace from the Keyboard and Mouse
Composing even that driest legal prose is a creative process. As the body becomes fatigued, the mind becomes distracted or dulled. A change of pace may be all that is needed to reinvigorate the writer's mind. For lawyers who compose rapidly and easily with keyboard and mouse, the microphone and voice recognition program may offer a refreshing change. In addition, there is something inspiring about having an obedient, uncomplaining assistant hanging on your every word. It can be inspiring to work while words literally fly out of your mouth and onto the screen. Voice recognition offers more freedom of movement than the keyboard and mouse. You are free to look out the window while you compose your documents. That is hard to do with hands on the keyboard and the screen right in front of your face. You can leaf through documents and dictate quotations without cranking your head back and forth between the screen and a book that stubbornly resists your efforts to keep it propped open.
Comparison of Voice Recognition Products
Dragon NaturallySpeaking vs. IBM ViaVoice Gold
Though there are a number of competitors in the voice recognition field, two companies have released breakthrough products at affordable prices. Dragon Systems was first with Dragon NaturallySpeaking. What was so startling about this new product was its ability to recognize continuous speech accurately. Previously, companies had offered voice recognition products that recognize discrete speech, requiring brief pauses separating all of the dictated words. For people unable to type, these products opened new doors. But for others not so motivated, the staccato speaking style required by these products made them unacceptable. IBM was quick to follow suit, releasing its continuous speech recognition product, ViaVoice, at a dramatically lower price, $149. Dragon Systems countered by lowering its price and differentiating its products into consumer and professional versions.
From the outset, Dragon NaturallySpeaking offered a powerful set of spoken commands for formatting and correcting dictated text. The IBM product required extensive use of the mouse and the keyboard to make corrections, but was the first to allow the user to dictate directly into Microsoft Word and other products.
Dragon has added the ability to dictate directly into Microsoft Word. Unfortunately, and this addition appears to be more an effort to match IBM on the features chart than a real improvement. Voice recognition was unacceptably slow using the Dragon Systems add-in for Microsoft Word on a computer with a 266 MHz chip and 96 MB of RAM. Most users will waste little switching back to dictating inside the Dragon NaturallySpeaking window.
Major Differences between NaturallySpeaking and ViaVoice
Though the features battle continues, the two products each have their own area of strength. NaturallySpeaking has the upper hand in the area of voice commands and voice macros. ViaVoice does a better job of integrating into other applications. Stay tuned for the newest offerings from both companies as they strive to better the competition.
Dragon Systems now offers a connection to a hand-held dictating unit produced by Norcom. With this unit, you can dictate anywhere. When you return to the office, you or your assistant can connect the dictating unit to the computer running Dragon NaturallySpeaking. Either one of you can then proofread the drafts transcribed by the computer, making corrections that will increase the voice recognition accuracy.
Microsoft understands that voice recognition features are extremely important to the future of computer software. In their usual style, they have snapped up one of the better companies in this niche marketplace, Lernout and Hauspie, and plan to offer voice control of the computer as an integral part of their future products. Windows 98 will not have the new capabilities but watch for them in future operating system releases. On the shelves of computer superstores are any number of products that contain voice recognition features. Most of these add voice control as a substitute for the use of the keyboard and the mouse. Their quality varies. You may be better off using the voice control features available in the products from Dragon Systems and IBM.
Let's assume that you have decided to take the plunge and commit to giving voice recognition software a fair trial. What is riding on your decision? To get to the point where your software works well for you, you will have invested a considerable amount of time in your own training and in training your computer to recognize your own speech. That investment of time will be preserved in speech files customized to your voice by the program. Those files, and thus your investment, will work only with the brand of software you have chosen. And they will work only with the microphone and computer on which they were created.
Your initial time investment will be partially productive, because you create real documents as you train the computer. However, at first you will not create them as quickly as you could using your old methods.
The work you do to train your computer is retained in speech files. At the beginning, you go through a number of exercises. One of them sets the correct volume levels of your equipment. Another does preliminary voice recognition testing. A longer exercise does initial customization of the software's speech files to match your voice. Then as you continue to use your software and correct its mistakes, your speech files are refined and their accuracy improves.
Since there are no standards in the area of voice recognition, the speech files you create in one software package cannot be transferred to a competing company's products. If you start with Dragon NaturallySpeaking, you will have a considerable investment tied up in this product. You can take your speech files with you as you upgrade to new versions, but you will lose your major time investment if, for example, you switch to IBM ViaVoice.
Your investment is also tied directly to the sound card, computer and microphone that you use in training your speech recognition program. So it is very important to purchase a powerful computer system. If you buy an underpowered computer, not only will your voice recognition program work less efficiently, but you may also face an early loss of your investment in speech files. You may need to junk your underpowered computer, along with your speech files, in order to upgrade to new software versions that offer new efficiencies.
Because of your large investment of time, the stakes are high in choosing between Dragon NaturallySpeaking and IBM ViaVoice.
Allow us to offer our very subjective opinion on the choice between Dragon NaturallySpeaking and IBM ViaVoice. Bear in mind the essentials for voice recognition discussed above:
- Powerful computer
- Training, including you and the computer
- Patience
- Proof reading
- Total voice control
Both companies offer products with comparable voice recognition accuracy. Each product has some features that its competitor does not. In reaching our conclusion, we worked on a variety of documents from email, to letters, to long documents. Superior voice control over editing and correcting text emerged as the deciding factor.
Recommended Product: Dragon NaturallySpeaking
We recommend Dragon NaturallySpeaking because it offers the best methods for controlling the cursor, making corrections, and commanding the computer to do what you want it do. In addition, it offers a working connection to a hand-held dictating unit, freeing busy attorneys to make productive use of their time in a variety of situations.
Recommended Version: Dragon NaturallySpeaking Deluxe
Dragon NaturallySpeaking Deluxe is the best choice if you are committed to taking full advantage of voice recognition technology to make better use of your time, whether you are an administrative assistant or practicing attorney. It offers the most powerful features and, significantly, it is sold by experts who offer experienced training and valuable support after the sale.
If you are more tentative about adopting voice recognition, Dragon NaturallySpeaking Preferred includes important features omitted from less expensive versions and is available for $149.
To compare the various versions of Dragon NaturallySpeaking and IBM ViaVoice, refer to the tables on the following pages. In addition, there is a list of tips for using Dragon NaturallySpeaking software.
|
Wells Anderson LEGAL TECH SERVICES 527 Marquette Ave., Ste. 1350 Minneapolis, MN 55402 |
phone 612-791-0471 fax 612-204-0008 e-mail WA@WellsLegalTech.com web www.wellslegaltech.com |
|
Helping lawyers leverage technology |
Ó Copyright 1998 Wells Anderson |
Features Chart – Dragon Naturally Speaking
|
|
|
||||
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
30 minutes training |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Features Chart – IBM ViaVoice
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Tips for Using Dragon NaturallySpeaking
|
Wells Anderson LEGAL TECH SERVICES 527 Marquette Ave., Ste. 1350 Minneapolis, MN 55402 |
phone 612-791-0471 fax 612-204-0008 e-mail WA@WellsLegalTech.com web WWW.WellsLegalTech.com |
|
Helping lawyers leverage technology |
Ó Copyright 1998 Wells Anderson |
Return to Top of Article
|
|
|