Limits the number of pronunciations for phrases in user dictionaries.
|
Value |
Boolean |
|
Default |
0 (pronunciations for whole phrases and their individual words) |
|
How to set |
In a grammar using a <meta> tag. See the text on compilation-time parameters in Parameters set in grammar files. |
|
Usage |
This is an advanced tuning feature for speech grammar developers to improve accuracy. |
|
Performance |
Use with caution. Can add CPU load with no benefit. |
In a user dictionary, you can define pronunciations for individual words and word phrases. By default, the recognition uses the phrase pronunciations that you provide plus the pronunciations of the phrase’s individuals words (which are provided automatically by the system). In other words, the phrase can be recognized as a whole or as the sum of its parts.
This parameter forces phrase pronunciations to be the only pronunciations used when the phrases appear in a given grammar. In most situations, this is not needed nor recommended (because it adds CPU load). To use this feature, set swirec_multiword_replace as a <meta> in the grammar. As a result, whenever the grammar combines words into a multi-word dictionary phrase, the Recognizer uses the phrase pronunciations while ignoring the pronunciations of individual words in the phrase. See Pronunciation dictionaries.
Sample <meta> setting in a grammar file:
<meta name="swirec_multiword_replace" content="1"/>
