|Troubleshooting Speech Recognition / About Speech Recognition|
Partitioning a deployment into sites improves speech recognition, because each site has its own dynamic grammar, which is the largest component of the recognition space.
When a badge user speaks a command, Vocera attempts to process it by combining the dynamic grammar of a single site with several smaller grammars. Specifically, Vocera searches the following grammars while processing a badge user utterance:
Either of the following dynamic grammars:
The dynamic grammar of the caller's current site
The dynamic grammar of the site the caller explicitly connects to
The Global site grammar
The static grammar
The badge user's personal grammar
Vocera always includes the static grammar, the grammar of the Global site, and the badge user's personal grammar while processing voice commands. However, Vocera includes the dynamic grammar of only a single site, not the grammars of every site, while processing the command.
For example, suppose the Central Pacific Resort deployment has two sites—Carmel and Monterey—in addition to the Global site. If a badge user in Carmel issues the command, "Call Adda Turner", Vocera uses the following grammars to process it:
Contacting a badge user at a remote site is a two-step process. For example, suppose the Carmel user in the previous example wants to call Adda Turner, but Adda's home site is Monterey. The user needs to speak two commands to place this call:
Connect to Monterey.
Call Adda Turner.
Vocera searches the dynamic grammar of the Carmel site—the current site of the user placing the call—to process the first command. After receiving the "Connect to" command, however, Vocera searches the dynamic grammar of the Monterey site, as shown in the following illustration:
When a person places a telephone call to the hunt group number or the DID number of the Vocera system, the telephony Genie prompts the caller to say the name of the person or group, or enter an extension. Vocera processes the caller's response to the telephony Genie as follows:
If the site is not sharing a telephony server, Vocera searches the Global site grammar and the grammar of the telephony server's site.
If the site is sharing a telephony server and the caller spoke a response, Vocera searches the combined grammars of the Global site and any sites associated with the line that the call arrived on.
If the site is sharing a telephony server and the caller entered a touch-tone response, Vocera searches the combined databases of the Global site and every site that shares the telephony server.
See the Vocera Voice Server Telephony Configuration Guide and Working with Multiple Sites for additional information.