I'm in the process of interviewing people to collect material for a book. These interviews are recorded on a digital voice recorder, then transferred to my computer and converted into MP3 files. The length of an interview ranges from one to two hours in length. As a general rule, there are only two voices heard - mine, asking questions and probing; the subject is the other one. Generally speaking, the quality of the recording is quite high , but naturally sometimes there are indistinct passages.

At the moment, I'm transcribing them by playing back the MP3 in a program, YAMP2.3, over the computer speakers and typing them into MS Word. This requires about 8 -10 hours per one hour of interview.

Since the sound quality seems pretty good, I wonder if I could hook a mic into the computer, position it in front of the speakers, and use a voice transcription program to make an initial copy. I know that I would have to edit it since the main voice changes with every interview so that there isn't much opportunity for the program to "learn" the voice. Still, any saving in time and effort seems worthwhile to me.

I'm aware of various programs like Naturally Speaking and ViaVoice. Are these, or similar programs, able to help me? Any experience out there?

Thanks for any input.