Voice recognition
| Voice recognition | |
12 March 2004
Normally the fingers of both hands come into play on the keyboard. But not now, thanks to literally falling off the back of a lorry and breaking not only an arm but the wrist as well, just for good measure. (By way of explanation, the author was helping to unload a kit house from a lorry trailer when he invented the new extreme sport of trailer diving. In another cruel twist of fate, who was on duty in A & E that day? The author’s wife!) Now, anyone who has ever said you can type with a knuckles-to-elbow cast on your arm has either A) never had a cast on their arm, or B) never used a keyboard. Short of cutting the keyboard in two, it can’t be done. The point to this anecdote is a salutary lesson: always have a contingency plan that you’ve actually put through its paces in advance. Having some time before pondered the question, “How would I be able to carry on working if I ever broke one of my arms?” the author was confident it would be business as usual. After all, there was that microphone headset kicking about in the desk drawer, and there was that software, Voice wotsit on a CD somewhere. Having concluded that keying in with the ‘bad arm’ was a non-starter, the contingency plan swung into action. Sort of. The headset was retrieved from the back of the drawer, Voice wotsit, “enables you to take voicemail messages on your modem…” X”!&%~#!! It was time to find out about voice recognition software. Some online research narrowed the contenders down to two brand names. Before wielding the plastic, though, it was decided to cast around, pardon the pun, friends and colleagues for some real life experience of using voice recognition software. What was most pressing was, “Will voice recognition software input text any faster than I can type with one hand?” The other issue was, “How long does it take to train the software so every second word isn’t a typo?” The consensus was that voice recognition has come on in leaps and bounds and would input faster than a one-armed touch typist. Training the software to faultlessly interpret yours truly’s dulcet tones would take about a month. Since the cast is only going to be on for six weeks, the conclusion was that the author remains to be convinced he can be outgunned on the keyboard by a piece of software. The compromise was to download an evaluation version of a voice recognition software package and try it out for himself. Still to come in this voice recognition spotlight series:
| |
