I did this with all of the transcripts for my PhD - some 40 interviews.
I found it really useful, and unless you are a touch-typist would
recommend persisting with it to see if it will work for you.
My recordings were done digitally and on cassette tape. I used the
cassettes for this exercise. I had the cassette playing into an
earpiece, and then spoke the words I heard into the headset linked to
Dragon. It took only a little while to get used to it. I found that it
was producing transcripts that were about 70-80% accurate - which is
much better than I could have got in that time by typing.
I also found that when I bought a better quality headset, the accuracy
of the transcript improved a lot.
My best time for a transcript was with one short interview with a CEO
who spoke very slowly. The 20-minute interview was fully transcribed
and corrected and printed off within 40 minutes of starting! It took
much longer with interviewees who spoke quickly, but then I guess that
would be true no matter what way I chose to transcribe it.
Regards
Ruth
Dr Ruth Bender
Cranfield School of Management
[log in to unmask]
-----Original Message-----
From: qual-software [mailto:[log in to unmask]] On Behalf Of I
Jones
Sent: 18 July 2005 08:33
To: [log in to unmask]
Subject: Voice Recognition Software and transcription
Hello.
Recently I tried using Voice Recognition Software (Dragon Naturally
Speaking
7) to transcribe.
VRS lets you speak into the microphone or playing an audio file and the
words appear in a text editor (or other application of your choice). You
have to train it to your voice for about 30 mins before first using it.
Obviously it is not practical to use for transcription recordings
because you cannot get every participant on the recording to spend an
hour training the software to their voices. However, a colleague
suggested that what you can do is listen to the recording and instead of
typing it into the computer you can say it into the computer. In theory
this should save time.
I have tried it with one 45-minute audio file. It was slow and fumbling
at first and took just as long as typing. But towards the end I was
getting the hang of it (so was the software - it continuously learns). I
suspect if I try it 2 or 3 more times things will really start to pick
up and my transcribing could become pretty fast, perhaps done in less
than half the time of typing.
Has anyone else who often needs to transcribe tried this method, or
would be interested in trying it?
Ian Jones
|