Daily Archives: July 16, 2009

Talking to a machine

Having had enough of the tedious process of typing notes for my research work, I finally gave in and purchased some voice recognition software.  Perhaps talking to the computer would be a less tedious process?  It’s a long way from Star Trek (computer do this, computer do that!) but it’s a start.

In a way it’s almost like talking to a person who is learning English and so is not so familiar with some words and sounds. The software has to learn new things as it comes across them.

The really quirky part involves the words and sounds it throws little wobblers on.  Some examples:

  • It has significant difficulty with words that have a “Tuh” sound and they end up being abbreviated to something else altogether. The word “tools” is non-existent in its vocab. It took some extended training to teach the tool the word “tool”. The word “through” poses similar problems.  You could end up with a phrase of “The tu tu the system caused tu many tu to be tu”. What does that mean? Answers on a postcard please.
  • It has a peculiar liking for the word “dissipating”. Every 3 or 4 syllable word ending on “ing” that it doesn’t understand ends up dissipating.
  • It has a number of placenames, person names, odd slang, etc that it likes to show off on occasion.  I’m not quite sure what it understood when it returned “British Tom porno” to something I said.

Correcting the mistakes is where the real fun starts.

Really, it is fun – honestly – particularly when I ask it to select the misspelled / misunderstood word and it stubbornly keeps blinking at me and doesn’t move to a pixel towards selecting said word. That’s when I have to administer the tough love. Would…. you…. pu lea ase … select … the … bleeping … word. Then, in all its intelligence, it assumes that I want it to type those words and has great fun wondering what “bleeping” is.

Once the mistake has been corrected, I often forget to tell the software to move back down to the bottom of the document to continue dictation. As a result I have sentences smack-bang in the middle of other sentences. One could weave a really interesting story there. How about: “significant shifts have occurred in flux and often distributed across geographic conceptualisation of classrooms”. Is that profound or just a lot of nonsense?

Nonetheless, I’m getting there in building a relationship with my voice software. It’s getting to know me and my “tu” pronunciations and I’m getting to know it and its dissipations. It’s the start of a wonderful friendship.

My Dragon and I