My Transcription Process

Having the need to transcribe lengthy interviews as quickly as I could whilst achieving as much accuracy as possible led me to devise a particular way of working that I want to share with you here.

My interviews were each one hour long and consisted of me asking a series of questions, to which the participants responded at length. I had 30 to transcribe, so it was really important for me to find a way of minimising the time I spent on each one.

Before I decided to look for ways to improve my efficiency, my initial transcription process looked something like this:

• Playback a segment of the recorded interview on my dictaphone.
• Stop playback.
• Type the words into a document.
• Resume playback on my dictaphone.

Transcription Software

My first task was to find some transcription software that would allow me to control playback of an interview while I transcribed it in a Microsoft Word document. I found Express Scribe, which does just that, and it's free. It allows you to control playback of a voice recording without even leaving Word. This means that you can listen to chunks of speech, press F4 to pause it and then transcribe it in Word. Press F9 and the recorded speech continues. If you find that you need to repeat some speech that you've missed, you can just hold down F7 to rewind - all whilst remaining in Word. You'll notice that already there is a big time saving because the hands don't need to leave the keyboard to operate the dictaphone. Transcription and interview playback are both controlled by the keyboard.

I found that although the process of listen - stop playback - type words - start playback, was efficient, I wanted to be able to transcribe continuously while the interview played without pause. Unfortunately I can't type that fast. The problem was solved with Express Scribe, which allows you to slow down the voice recording. I've found that slowing it down to around 50% allows me to type continuously everything that I hear, without pausing playback. The voices sound a little funny, as though the speakers are drunk, but doing this really speeds up the process. If you can type more quickly, you'll find that you can cope with a faster playback.

Already, I'd speeded up my transcription process by using playback software that can be controlled from within Word. But then I had an idea that made use of dictation software.

Dictation Software

Dragon Naturally Speaking is voice recognition software that allows you to dictate text that your PC will transcribe. In a word, it's fantastic! After a period of "training" where you read aloud some passages into your microphone, Naturally Speaking is able to recognise your voice and transcribe your words directly into a word processor (and a variety of other applications like email, spreadsheets, etc). The most basic version is only $45 and I considering how much time and effort it saves me, I think the price is well worth it.

Ideally, I wanted voice recognition software to transcribe the whole interview into a document without any intervention from me! Wouldn't that be wonderful? The problem with this dream is that Dragon Naturally Speaking needs to be trained to recognise a particular person's voice. My participants were kind enough to give me the time to interview them, but getting them to train the software to recognise their voices was too much to ask. I knew I couldn't get the software to transcribe everything without me doing something.

Then I hit on another idea. I could listen to the interview in my headphones and repeat everything I heard into my microphone. Naturally Speaking would transcribe everything I said because it would recognise my voice. If I used a headset microphone, I would avoid any interference between the recorded interview and my repeated words. Speaking the words I'd just heard was far quicker than typing them in manually. Additionally, I found speaking the words so much more comfortable than typing them. This was the method I used to transcribe my 30 interviews.

I know a lot of people who do transcriptions use a transcription foot pedal to speed up their process. Doing this helps only if you are typing the interview yourself, as the foot pedal takes away some of the typing work your fingers would ordinarily do. In my case, I'm actually using the voice recognition software to do the transcribing, so no foot pedals are necessary.

Transcription Equipment

Surprisingly, the laptop I bought two years ago is good enough to cope with the demands of the software I'm running. Here is a list of the equipment I'm using:

• Two year old laptop with 1GB Ram, running Windows XP
• Sennheiser PC 31 Headset
• Express Scribe
• Dragon Naturally Speaking 10
• Microsoft Word 2007

Transcription Process

In a nutshell, my workflow can be described like this:

• I have Microsoft Word open.
• I'm using Express Scribe to control playback of my recorded interview in my headphones.
• I repeat what I hear in my headphones into my microphone and Dragon Naturally Speaking transcribes my words into the Word document.