Dictating on the go with a headset can be problematic, since most headsets don’t fit well in to a compact computer bag. Thankfully Nuance created Dragon Recorder, which works along with Dragon MacSpeech Scribe to make mobile dictation easy enough that I actually enjoy doing so.
I’ve previously reviewed Dragon Dictate for Mac, which works, but not as well as I’d like. However, using the Recorder and MacSpeech Scribe made me more productive.
[Also check out our Notebooks.com overview of Dragon Express from the Mac App Store]
Dragon Recorder will record audio using your iPhone and then let you export it to your computer for transcription using Nuance MacSpeech Scribe. You upload the resulting WAV file, created in the app, to a computer using your web browser. MacSpeech Scribe transcribes the audio into editable text.
MacSpeech works best if you record a full segment of audio and then upload it. I wanted to give it a shot and see how well it works. So, I am writing this review on my iPhone.
The process is simple enough. First, you record audio on your iPhone. Then, using your browser to download the WAV file, you open it and MacSpeech Scribe will transcribe the audio into editable text. The iPhone app provides the IP address which you enter into the browser address bar. Your computer and iPhone will have to be on the same network.
Before you can actually use the program it must go through a training process. When you use dictation software it normally asks you to read prewritten text. Instead MacSpeech Scribe uses the Recorder app on your iPhone to record any audio you choose simply by speaking into your iPhone. It then attempts to read the audio and turn it into text. The resulting audio was pretty garbled, but that’s because it hadn’t yet learned my speech patterns or voice inflections.
The program learns how to recognize your voice by breaking up your recorded audio into small bites. The training screen shows you the resulting text and you click on the phrase one at a time. Type out what the correct words should be and MacSpeech teaches itself to properly recognize your speech.
The process was simple. It automatically advanced to the next snippet of audio once I corrected it with my keyboard. When it finished with the first 15 second segment, it re-recognized the entire thing.
MacSpeech Scribe did a good job. It wasn’t perfect, but close. I was able to transcribe the introductory paragraphs of this review using the training files. Now I’m writing the rest of the review using the app.
I like that you can stop the recording on the iPhone app if your phone rings or you have go to a meeting. After taking a break from dictation you can then simply hit the record button on the app again and it picks up where you left off, adding to the end of the previous audio. This way you can upload a full segment in one file even though it was recorded at different times.
On my first attempt to train MacSpeech Scribe the audio file recorded on my iPhone was only about 60 seconds long. In order to train the computer app, you have to upload a file that is at least 120 seconds long. I just started over and this paragraph of the review was my second attempt.
Being able to record audio on your iPhone and then upload it to your computer for recognition could potentially make you a lot more productive, even while away from your computer. You could write a letter while driving down the road or as you wait in a line somewhere. Since most people speak faster than they type I can see that this would be a huge benefit even if you’re sitting at your desk.
MacSpeech Scribe is a simple and easy to use program. The training could use some more hand-holding to teach the user how to get started. But after that, it works well. The transcription part of the program works much like Dragon Dictate for Mac. You can speak text and punctuation saying things like “period” for a period at the end of a sentence. The vocabulary editor lets you add your own words as well. To help MacSpeech Scribe understand your writing style you can upload some of your documents. It analyzes them and finds words that it doesn’t understand so you can train the program to work with vocabulary unique to your work.