Wednesday, June 3, 2009

822nd Post - A New Approach

People have been asking me when the next interviews will be up. I can tell you that 2 weeks ago I sent a transcription of an interview with the subject; he hasn't got back to me yet, and I have been too busy to follow up.

Four more interviews are still in the transcription stage. It's a bottleneck.

I'll complain for one paragraph only. My frig, this is a lot of typing! I am wearing my fingerprints off.

I have been researching a new approach to producing text and think I have found one. There is software called Dragon Naturally Speaking from Nuance. There is now a version for Vista, 64 bit. The standard version is a hundred dollars U.S. I might invest in that.

However, Vista itself has voice recognition software built into it. I must confess, being fairly new to Vista, that I didn't know that.

As I have explained before, there is no voice recognition software on this planet that will take the spoken words of more than one person and produce any kind of reasonable text from them. The technology isn't there yet, and won't be for a long time to come. Since I have these audio files, dozens of them, to transcribe, I thought VR software wouldn't fly.

But I have done a bit of research over the past 24 hours. I like what has been suggested to me. One can fire up this Vista VR software, "train" it to become accustomed to one's voice over the space of an hour or so, and then do something that someone called "hyper echo dictation" to transcribe these audio files.

Let me explain.

Wearing headphones and playing my digital voice recorder, I would listen to an audio file. After a 2 second-or-so delay, I would simply repeat what I had just heard into a microphone that I'll purchase this evening, and which is hooked up to a usb port in my computer. The result, I hope and pray, will be a fairly clean text file that will require minimal editing before going on to the next audio file and repeating this process.

If this works, I should be able to blaze through this transcription process from now on. It will mean a much faster turn around for these interviews. Much less drudgery for Bevboy. Which means a better end product for you guys to read.

I'll let you know how this works out. There will be a learning curve for me and the VR software. I am willing to make this time investment.

Can hardly wait to get home and try this out!

Bevboy

No comments: