I’ve been playing around with Dragon NaturallySpeaking. It is really quite remarkable. After only about 5 minutes of training, it is able to track with my voice extremely accurately. I’m trying to learn to use a mix of my mouse and voice to create text very rapidly. The only trick is learning all the voice commands. Though they’re very intuitive, there are about a million of them and it is taking some time.
As you might have guessed, I “wrote” this entry with a microphone.
5 Replies to “Speech!”
How long did it take, Jay?
Longer than it should have because I was playing around with some of the control commands i.e. moving the insertion point around, highlighting text, etc.
From what I can tell, if I had both hands free to perform corrections (it provides a very fast method of making corrections using the keyboard), it looks like it would have a very high top end. I’m betting I could approach or surpass 100 wpm on a large chunk of text.
do you know, or could you try, using a cassette
recording of your voice and let me know how successful it is (if I may impose ).
I’m curious about it because our minister writes
out his sermons in a long-hand verbose outline style (using a water-based ink pen). We would
like to convert tapes to print and I’m gathering info. I have the version that is at least 3 years old. I also assume the software would have to be ‘trained’ for his voice (can it be trained for more than one voice?).
Here’s what I know off the top of my head. It includes a specialized Transcribe function meant to convert recordings to text. It can be trained to different voices (you set up user profiles). So, if you had him train the software to his voice, I’m willing to bet you’d have decent success with it, though I have not yet tried working with a recording yet.
I worked with a pediatric radiologist at Wash U who wanted to use Dragon on his computer. He was the ONLY one in the entire medical school using it. This was about 3 years ago, however, so I’m sure the software has been improved greatly.
I can’t stand working with voice recognition software. They were toying with the idea of moving the entire radiology department to VR but eventually canned it for a while. Too many foreign doctors with accents the software just could not map.