When training or correcting words using the Vocabulary Editor or Train Words features is there any advantage to training the words multiple times - or is it sufficient to train the word or words a single time? I have submitted this question to several different voice recognition forums over the last several years and have received answers that run the gamut from training the word or words only one time to achieve the best accuracy to some people who claim that training the word multiple times gives the program more information enabling it to ascertain the most accurate enunciation for the word. Any help with this perplexing problem would be greatly appreciated.

Hello Marty,

Fortunately I do train words in context to the surrounding words. What I am trying to ascertain is whether or not it is advantageous to train the words multiple times. For example, when training the words "church steeple" would it help the program to recognize the words in the future if I said "click Go" and then said "church steeple" and then said "click Go" again and then pronounced "church steeple" again, and so on. Or is it sufficient to pronounce "church steeple" once and then click "Done"?

Robert

Marty pretty well nailed it but going back to Marty's example about training words in context, we recommend training words in the actual context they are used in. Let's go back to Marty's example where he dictated “I'm going to see the church steeple”. Rather than training “church steeple” we recommend placing the cursor anywhere within the phrase and saying train that, clicking the Go button, reading the text and then clicking the Done button. The reason why it is better to train the entire phrase (in this case all 6 words) rather than part of the phrase is because NaturallySpeaking has made a temporary recording of that phrase to use for comparison. Training a partial phrase is a waste of time because NaturallySpeaking won't know what to do with it.

We like this feature so well that we automated it in KnowBrainer 2008 with a Quick Train command which opens the training dialogue, clicks the Go button, verbally prompts you to begin speaking (training), (you have up to 3 seconds to begin speaking but can take as long as you want) and KnowBrainer will click the Done button for you to put you back in your document. This is one of the 11,000 commands KnowBrainer uses to automate these types of mundane chores.