At 10:06 PM -0800 3/1/01, Kristin Luker wrote:
> <snip> How about voice recognition software? I suspect it can't work
>off a tape, but how about a person who has "trained" the software repeating
>a tape into the voice recognition mike? How about digital recorders that
>have voice recognition built in? (I read in the New York Times about an
>Olympus digital tape recorder that has this.)
I use voice recognition software to transcribe, but it won't work the way
you want it to. I have trained the software to my voice, so I listen to the
tape and simultaneously speak the words into the mike. This works a little
better (but not much) than just typing the words in. But, then, I am a very
slow typist. I can also stop the tape recorder and dictate memos or
annotations. I have found that after the first pass, I must go back and
listen to the tape again and correct mistakes.
The digital recorder also works reasonably well, but only for one voice.
You cannot automatically transcribe an interview because the software will
not recognize multiple voices at the same time.
I had to upgrade to a pretty high end computer system to get what I
considered satisfactory performance from the voice recognition software. I
was using a 250 mhz AMD K6 computer with 96 mbytes of RAM. It was not up to
the task. I am now using a 1000 mhz AMD Athlon (Thunderbird) processor with
512 mbytes of RAM and it works pretty good.
The software I am using is DragonSystems Naturally Speaking Mobile Edition
(which came with a recorder). I expect IBM's ViaVoice would work about as
well and have the same requirements.
Hoping this helps, I remain electronically yours,
Elliot
Elliot Richmond
PhD candidate in science education
University of Texas at Austin
[log in to unmask]
http://www.geocities.com/Athens/4758/
|