1

The spec for repeat probabilities in speech recognition grammars is at http://www.w3.org/TR/speech-grammar/#S2.5.1 This includes the following example:

<-- The rule reference to digit must occur two to four times -->
<-- with 80% probability of recurrence. -->
<item repeat="2-4" repeat-prob=".8">
   <ruleref uri="#digit"/> 
</item>

I'm trying to determine just what this example means. The minimum length in-grammar utterance can contain two digits. Does it then follow that the probability of a 3-digit utterance is 0.8 and a 4-digit utterance probability is 0.64?

Or am I misinterpreting what 'repeat-prob' means?

Kurt Godden
  • 29
  • 4

1 Answers1

0

My interpretation of the specification is that there is a 20% probability that there is no repeated digits after the first two. There is an 80% probability that there will be a third or forth digit. There is no difference in probability between the chance of a third or forth digit.

Jim Rush
  • 4,143
  • 3
  • 25
  • 27