by Vannesa Mosher
Reprinted from Usability Interface, Vol 7, No. 1, July 2000
A reader asks, “Is anyone using voice/speech recognition software? And how effective is it?”
Candace Bamber, a project manager and information products lead at Castek in Toronto, Canada replies:
“I’ve been using IBM’s ViaVoice for about a year at my job for a variety of text entry projects including manual writing, status reports, proposals, and e-mail. ViaVoice can “type” about 120 words a minute, which is a lot better than I can. However, I don’t tend to use it to edit. That is, I don’t enter updates by voice and I don’t do online reviews of other people’s materials with it, it’s just too cumbersome. If I’m dictating work-related stuff, ViaVoice is very accurate because it’s very well trained on our house style and vocabulary. However, I usually have to make corrections to e-mails because it is not as well trained to my “personal” voice.”
The basis of this discrepancy is how the system is initially set up, which can be a somewhat painstaking process. It takes a couple of hours to set up, if you’re thorough: you read a passage to ViaVoice, that takes about 40 minutes, then you feed it samples of your writing. It recognizes speech by listening to your sounds and matching what it hears against a list of the words you were most likely to have used–based on the rules of grammar and your writing style, which is analyzed from your samples.
It takes another six weeks of using the system’s “correction window” to correct any mistakes it makes. When you use the correction window, it learns the proper word. If you don’t use the correction window, you will have to keep correcting that word manually. The correction window is especially useful for adding acronyms and industry-specific words but it requires a lot of time and effort.
If the setup seems overwhelming, there are further issues that deserve mention. If you are running ViaVoice with Word, your personal voice file is very vulnerable to Word crashes and file corruption.
You must save a copy of your personal voice file to another directory every day. Also, you cannot install ViaVoice and RoboHelp or Doc-To-Help on the same machine. The Directory Link Libraries and registry changes at install interfere with each other.
Overall, I would say my experience with speech recognition software has its benefits–I had a goal to cut down my need to use the keyboard by 30 percent to 40 percent and that has certainly happened–along with a huge improvement in my hands, wrists, and shoulders. But I do suggest that at this stage of the technology, people interested in using speech recognition software have tremendous patience.·
Visit the following web sites for more information about Voice and Speech Recognition: