Voice/Speech Recognition Software

I work for the Department of Social Security in Australia, I am a claim assessor; however I am also the Depts. Voice Activated Computer (VAC) Trainer I have had experience with Dragon NaturallySpeaking Home, Version 11.  For it to work you  must NOT use the keyboard “just this time, because I’m in a hurry”. People who have been professional Typists or are good mature typists are the Major offenders. None of them in my experience are as fast or accurate as a typist. This is not a reflection on the programs but on the level of Technology the human race has attained and all of the programs is getting better each upgrade!

Question: Lots have said about voice/speech recognition software as a substitute for the keyboard and mouse. Is there anybody out there who actually has “hands off” experience with Dragon?
I would be interested to hear user reactions on:
1. Continuous speech recognition systems (i.e. replacement of a secretary for dictation);
2. Any completely hands free means of software development.

Answer: In my experience, consistency is important. Accents are another matter. I know some people say regional accents cause problems. They shouldn’t. When this last came up, I tested it by retraining some words using the French pronunciation, and completely altering others. (Be careful if you try this, for the reason I’m now going to give.) Dragon coped fine as, I’d think, would the others, as it knows the pattern of my voice. I pronounce Rousseau in an unusual way. The first time I said it. Dragon got it.But that is the reason not to repeat my experiment. Changing horse to banana, or whatever, unless you put it right straight away. My system tends to get better trained than many because I have a soft voice; I compound that by training it to know me when I slur my words. It does not take that much longer to do a thorough job of this. I have been amply repaid. So, don’t be perturbed.

Leave a Reply

Your email address will not be published.

Notify me of followup comments via e-mail.