TBRG 2005

You Talkin’ To Me?
Voice Recognition Software Quickly Making a Name for Itself

by Dwight M. Owsen and Kent N. Schneider, CPA

In the “Star Trek” series, Captain Kirk often would deliver commands to his computer, which would then perform the requested task. Or, he would verbalize a problem and the computer would respond with an answer.

While voice recognition technology is not at that level of sophistication, it is becoming a useful tool that is finding its way onto accounting professionals’ desks.

What is Voice Recognition Software?
Simply, voice recognition software allows computer users to speak commands into their computers—such as “open file”—via a microphone to complete the task.

Voice recognition systems range from high-end applications for medical professionals to low-end programs that allow users to navigate the internet.

For most CPAs, neither extreme has much appeal due to cost and functionality issues. However, there are mid-range programs—such as Dragon NaturallySpeaking, IBM ViaVoice and even Microsoft Office XP—that are available at reasonable costs.

Many users already have the computing power to run voice recognition programs, so the decision becomes which software to purchase. Prices start at less than $100 and increase depending on functionality.

Microsoft Office XP, office.microsoft.com/en-us/assistance/HA010565111033.aspx, possesses basic speech recognition capabilities, which can be an inexpensive way to experiment with the technology.

Want greater capabilities? Dragon NaturallySpeaking, www.scansoft.com/naturallyspeaking, has a range of programs for Windows users and offers versions that recognize several languages in addition to English.

Those using alternative operating systems can use IBM’s ViaVoice software, www.3.ibm.com/software/speech, which comes in versions for Windows, Macintosh and Linux.

It’s worth taking the time to review these and other software packages on the market to see what meets your needs.

Getting Started
The accounting profession is often a leader in adapting new technology. For example, it was one of the first to include spreadsheets and word processing software in its business toolkit when they became available in the early 1980s.

So why hasn’t the profession incorporated voice recognition? The software is easily available; it’s affordable, as many packages are available for less than $100; it’s simple to install; easy to learn; and compatible with newer computers and the latest office suite software packages.

While installing voice recognition software is easy enough, users must train the software—and themselves.

The training process requires the user to dictate text that appears on the screen into the software. The software then links the digitized voice input to the words in its vocabulary files. To further enhance accuracy, some voice recognition programs will compare the supplied vocabulary files with existing documents prepared by the user. Unique words found in the user’s documents are added to the vocabulary files, and then the program asks the user to dictate these new words.

Today’s voice recognition software has reduced training requirements from several hours to roughly 30 minutes.

Although the software training process is relatively quick and painless, users must train the software to recognize their particular speech patterns.

In addition to speaking distinctly, users must learn program-specific voice commands for turning the microphone on and off, editing text and correcting errors. Typically, users can quickly master the most frequently used commands.

Those accustomed to using keyboard shortcuts and entering commands by mouse often find voice-operated editing to be slow and inefficient. If so, they may choose to use the software for dictation, but not for editing.

Voice recognition programs don’t require users to abandon the keyboard and mouse. After experimenting with the software, users find the particular blend of microphone, keyboard and mouse use that is optimum for their needs.

Keys to Success
There are four keys to using voice recognition software effectively and productively.
 
Reasonable Expectations—Voice recognition, like any software, is not perfect, so you must have reasonable expectations regarding the level of accuracy that can be achieved. Voice recognition has an accuracy level of 90 percent to 95 percent for dictation.

Operability —Some applications are more easily operated via voice recognition software than others. For dictating text into a word processing program, the products available at the local computer store are sufficient since the word processor does not have to shift frequently between accepting data and interpreting program commands. However, to use voice recognition software with a spreadsheet or database, adjustments must be made to ensure success.

Some users will continue to use the keyboard and mouse, and use the microphone for data entry. On the other hand, users desiring to use the microphone exclusively for spreadsheet or database applications likely will need to purchase a more sophisticated and expensive voice recognition program.

Training —The more time spent training the software and modifying your work habits, the higher the level of accuracy. In older versions of voice recognition software, users had to speak clearly and enunciate with a small amount of silence between the words. Today’s versions learn best if you speak at a normal speed and read text material with breaks only to catch your breath. The software learns by associating what you say with a specific word or phrase, as well as the way you say it. With practice, most users intuitively adapt their dictation cadence to maximize accuracy and productivity.

The Right Equipment —Before spending a dime, determine your needs—and budget—so you buy the right hardware and software for the job.

With computer prices lower than ever, computing power is not likely to be a problem. Instead, for those dissatisfied with their voice recognition results, the culprit can often be a cheap or improperly adjusted microphone. Throw away the microphone that came with your voice recognition software and buy a good quality replacement. If your computer came with a high quality sound card, then a moderately priced headset with RCA plugs will be fine. However, if your computer has a free USB slot, you will be much happier spending a bit more for a higher quality USB headset.

Applications and Benefits
CPAs will find voice recognition software to be appropriate in word processing needs, such as letters to clients, supplemental notes to financial statements and any narratives or memos in the work papers; office memos; and e-mails.

In general, voice recognition software can lead to:

• Greater Productivity —One can usually dictate faster than one can type.
• Carpal Tunnel Syndrome Relief—It is projected with the increased use of keyboards that more people (including CPAs) will suffer from carpal tunnel syndrome. Replacing the keyboard with a microphone will allow these people to be functional during the periods of recovery.
• Pre-recording—Using a digital recorder, one can dictate on the run and then convert to text for memo later using the voice recognition software.
• Automation—Voice control repetitive tasks on the computer.

Voice recognition software represents today’s computing and to exclude this handy tool from an accountant’s toolkit is to limit the use of technology—and miss out on the benefits it can provide.

Dwight M. Owsen is a professor in the department of accounting, taxation and business law, at the Brooklyn Campus of Long Island University.
Kent N. Schneider, CPA is a professor in the department of accountancy at East Tennessee State University. You can reach them at owsend@yahoo.com and kent@etsu.edu, respectively.

©2005 California Society of Certified Public Accountants. For reprint permission, contact Aldo Maragoni, managing editor.

 

Highlights
What is Voice Recognition Software?
Getting Started
Keys to Success
Applications and Benefits