Monday, January 13, 2014

Using speech recognition (dragon naturally speaking) effectively

I work upwards of 10 hours a day on the computer and mobile phone. Needless to say, this is a potential to affect my health adversely. In order to reduce the strain on the fingers because of typing, I have been wanting to use speech recognition since about 2007. However, I could not use it as part of a daily routine because of the poor quality of speech recognition. 

I experimented with a variety of microphones, including those that were recommended by the vendors and popular sites such as this , but I would still not get an environment where I could just dictate sentences and not have to correct them after every few words. There was an additional problem of the program taking up a huge amount of memory. As a developer, I had to run other development packages on my computer simultaneously which caused the computer to slow down significantly.

A friend of mine told me that Dragon NaturallySpeaking version 12  has the ability of being able to dictate content wirelessly. This turned out to be a very useful feature. In fact, for me, it turned out to be a feature that helped me use voice recognition in my daily life instead of it just been a cool new software to try at times. I use the Dragon naturally speaking with the following arrangement:
  1. I have installed Dragon NaturallySpeaking on an Windows XP old laptop with 2G RAM. I connect to this wirelessly from  a dragon naturally speaking  app on my android mobile phone. Separately, on my main development machine, I login to the Windows XP laptop through remote desktop. I then dictate to Dragon NaturallySpeaking through the mobile phone and view the results on the Windows XP laptop using remote desktop.
  2. The superior quality of the microphone in the phone has dramatically help reduce voice recognition errors. Since the software is running on a separate machine, it does not slow down my main development machine.
Because of this feature, I have abandoned trying windows speech recognition, although I really liked it better while using browsers.

No comments:

Post a Comment