(updated) This post previously included videos on utterz and Revver. Neither video service still exist today.
Here’s my first attempt trying to transcribe a voice file recorded on a Sony IC voice recorder ICD-UX70 with Dragon Naturally Speaking 9 Preferred.DNS actually makes a mobile edition, but I just have the Preferred edition. So I’m trying to do this basically using non-standard equipment.
That said, I found this to be relatively easy. I need to make a few minor adjustments to my technique, but all in all it worked pretty good.
The level of accuracy at the beginning of the transcription (first couple sentences) was not as high as DNS normally is when recording straight through a microphone on the computer. But towards the end of the recording, it seemed to rapidly improve.
Plus, after the transcription session, I went through the entire document and made corrections within the Dragon system so that it could learn from the mistakes made with the Sony recorder. I’m going to repeat that a few more times to see if I can continue to get better results.
The audio that goes along with this post is the audio from the recording session.
Below you will find the transribed text (Post Correction):
This is a test for Dragon Naturally Speaking a digital recorder and I’m going to be importing the voice file into my computer after I connect my digital voice recorder to my computer through its USB slot. I then have to run import function with my computer and Dragon NaturallySpeaking pulling in the voice file into Dragon NaturallySpeaking in MP3 format. In the past I’ve tried pulling in MP3 files and converting them into wave files. This time I’m going to be trying to pull it in using an MP3 file and see if Dragon NaturallySpeaking can turn transcribe it better than it can as a WAV format. This Sony recorder comes with a user’s manual that recommends the MP3 format as opposed to the wave file.
So now I will see if this version works better. I’ve tried the WAV format once and it didn’t work so great however at the time I was using the recorder with the mic sensitivity set at an incorrect setting which brought in a great deal of exterior noise.
This time I had the microphone settings correct and I am hoping that the MP3 format as prescribed by Sony will help transcription more correctly.
The other thing that I have to work on is learning how to train Dragon naturally speaking for my voice recorder. I’ve never trained Dragon NaturallySpeaking for a voice recorder before, and so this will be a new experiment for me. Currently, I am driving down the street in my car making this recording. Driving a stick shift, and it makes a voice transcription little bit tricky.
Well that’s the end of this proof of concept. I now know that I can take a simple little pocket recorder, keep a voice record, write an email, or even a book, almost anywhere from a plane to an airport to a cottage on the beach and even on my kayak.
This is just one more way to un-tether me from the wires of my computer when I need a little more space or flexibility. 🙂