The Next Voice You Hear Will Be Your Own

I’m in the process of getting the interview I conducted the last week in February transcribed — or, rather, I’m in the process of trying to get it transcribed.   I’ve got four hours of conversation to convert, and initially, I was planning on doing it myself.  I’ve transcribed interviews before, but nothing quite this long — and after taking 20 minutes or so to transcribe about five minutes of conversation . . . well, that’s an unimpressive effort-to-product ratio.  Clearly, I need another system.

For Plan B, I checked with a few professional transcription services, but the per-page costs of transcribing were a bit jaw-dropping.  Probably nothing your average law firm can’t soak up, but for a company of one, it was gonna leave a mark. 

That left Plan C.  I’m in the process of loading voice-recognition software onto my laptop to see how well it does. I chose the Scribe program from MacSpeech, mainly because it allows you to open a sound file directly through the program.  Now the only problem is the operating system on my MacBook.  It runs Leopard, and I need Snow Leopard.  So I had to order the upgrade from Apple (for some reason, they won’t let you download it from their site) and my package is still enroute with the FedEx man.  Or woman.

Anyhow, I’ll let you know how it works.  Anyone else have experience with a similar program?

3 responses to “The Next Voice You Hear Will Be Your Own

  1. Brian, if you’re intending on using that voice-to-speech program to transcribe your whole interview direct from a recording you made of it, you might have a bit of trouble.

    I’ve not used the programs you describe, but I have used Dragon NaturallySpeaking for years (I’m using it now!): it will only work with one voice at a time and then only for voices it’s trained to recognise; and it’ll only give an accurate transcription of a recording if you used a very high-quality digital recorder. And even if you did, you’ll still have a lot of editing to do afterwards to get it right.

    As I said, your program might well be different, but from what little I know it’s not likely to be–sorry.

    To get a reasonably accurate transcription of your interview you’re probably going to have to echo-dictate the interview into your word processing package, rather than just feeding it in direct. And that’s a whole new skill–trust me, I’ve done it.

    I hope I’m wrong and that it all goes swimmingly well for you. I’ll be interested to find out.


  2. Jane: Thanks for the message. As it turns out, you’re two for two: (1) I WAS trying to transcribe a multi-speaker interview, and (2) failed miserably. Clearly, Scribe is NOT what I need. I still require the human element, it would appear.


  3. Pingback: Credit Where Credit’s Due… « Brian Jay Jones