Re-Dictation: Harnessing the Power of Front-End Speech Recognition Technology
As all long-time transcriptionists know, years of high-level production transcription inevitably take a toll on the fingers and wrists. What many MTs may not know, however, is that modern consumer speech recognition technology (SRT) offers a possible alternative to all that keyboarding. I call this option "re-dictation," which is the process of using a front-end speech recognition (SR) application such as Dragon Naturally Speaking or Windows Speech Recognition to "re-dictate" transcripts rather than type them on a keyboard. It takes some getting used to, but I've personally found it to be a useful alternative when I want to give my hands a rest.
The technical requirements are not particularly complex. In addition to the SR software, I recommend using a PC headset with one earphone and a boom microphone. For me, at least, having one ear "open" to hear myself speaking while using the other ear to listen to the dictation works best, but your mileage may very. Once you have the hardware and software components assembled, you're ready to begin the re-dictation process.
If you're new to using front-end SRT, the first step will be to train the software to recognize your voice. Dragon Naturally Speaking, the program I use, has an extensive array of training tools available. The first step in training is to read aloud from prepared samples of text; later you can also add additional words to the application's vocabulary, either manually or by scanning documents on your hard drive, and then train the program by reading the new words aloud. Some versions of Dragon come with an extensive medical vocabulary included, but are quite a bit more expensive. I've had good success with off-the-shelf versions of the program and have not seen the need to spend the extra few hundred dollars to get the medical version.
Once you've gone through the initial training process, it's time to start re-dictating. The process is quite simple: you play a short snippet of the dictation file, stop playback, dictate that phrase into your headset microphone and watch the words appear on your computer monitor. As is the case whenever you use front-end speech recognition software, you'll need to make corrections from time to time. In many cases you can do this using voice commands, with little or no keyboarding.
As you get used to the re-dictation process, you may find you're able to "overlap" at times, that is, listen to the dictation and re-dictate at the same time.
The question often arises as to how re-dictation compares to keyboarding in terms of speed. For many high-speed typists, re-dictating will be slower, there's no question about it. But if it comes down to a choice between not working at all (or working in pain) versus working a little slower using SRT, re-dictation is an option not to be hastily overlooked.
An important caveat to using SRT, whether of the front-end or back-end variety, is that meticulous review of the words on the screen is a must. When you switch from keyboarding to speech recognition, it's amazing how easily your brain can be tricked into thinking your eyes are seeing something that really isn't there. Speech recognition is still an imperfect technology, and you can never assume that the program got everything right. Whether you edit your work in real time as you go, or re-read the entire document after you've finished, or both, using speech recognition makes careful review even more critical to accurate transcription.
For many years, speech recognition technology has been a boogey man to many MTs, a mysterious force to be feared and resisted. Re-dictation is an example of harnessing the power of this technology to work for us, not against us.